|
|
|
US House Contest 13th District |
|
|
Written by Christine Jennings
|
|
Wednesday, 17 January 2007 |
|
IN THEUnited States House of RepresentativesCHRISTINE JENNINGS,Contestant,V.VERN BUCHANAN,Contestee.NOTICE OF CONTESTREGARDING THE ELECTION FORREPRESENTATIVE IN THE ONE HUNDRED TENTH CONGRESSFROM FLORIDA’S THIRTEENTH CONGRESSIONAL DISTRICTPursuant to the Federal Contested Elections Act,2 U.S.C. §§ 381-396KENDALL COFFEYCOFFEY & WRIGHT, LLP2665 South Bayshore Dr.PH-2, Grand Bay PlazaMiami, FL 33133(305) 857-9797MARK HERRONMESSER, CAPARELLO & SELF, P.A.2618 Centennial PlaceTallahassee, Florida 32308(850) 222-0720DONALD B. VERRILLI, JR.SAM HIRSCHBRIAN HAUCKJESSICA RING AMUNSONJENNER & BLOCK LLP601 Thirteenth Street, N.W.Washington, DC 20005(202) 639-6000December 20, 2006Introduction1. This is an action brought under the Federal Contested Elections Act, 2 U.S.C.§§ 381-396, to contest the Florida Elections Canvassing Commission’s November 20, 2006certification that Contestee Vern Buchanan received 369 more votes than Contestant ChristineJennings in the general election for Representative in Congress from Florida’s ThirteenthCongressional District. The vote totals in that certification are wrong because they do notinclude thousands of votes that were cast in Sarasota County but not counted due to the pervasivemalfunctioning of electronic voting machines. The number of uncounted votes in the County ismore than sufficient to change the result of the election. Indeed, statistical analysis based uponthe actual ballots cast in Sarasota County in the November 2006 general election demonstratesthat, had the votes lost to machine malfunction been included in the certification, ChristineJennings would have won the election by more than 3,000 votes.2. On November 20, 2006, the State of Florida’s Elections Canvassing Commissioncertified a total of 119,309 votes for Vern Buchanan and 118,940 for Christine Jennings. Thatcertification excluded the votes of thousands of Sarasota County voters who used the County’selectronic voting machines to vote in the election for the Thirteenth District seat and did not havetheir votes recorded. Indeed, the electronic voting machines in Sarasota County failed to recordvotes in this race for more than one out of every seven voters — nearly 15% of those who usedthe machines. There is no possibility that so many Sarasota County voters would havevoluntarily abstained from voting in this hotly contested, high-profile race, especially in anelection year when control of Congress was obviously at stake. Statistical analysis confirms thatcommon-sense conclusion. Even more strikingly, the eyewitness accounts of hundreds ofSarasota County voters, and the contemporaneous records of the Sarasota County Supervisor of3Elections, document that the electronic voting machines in Sarasota County systematically failedto record votes cast for candidates in the Thirteenth District congressional race — particularlyvotes cast for Contestant Jennings.3. By law, every polling place in Florida displays a “Voter’s Bill of Rights” statingthat “Each registered voter in this state has the right to: . . . Vote on a voting system that is inworking condition and that will allow votes to be accurately cast.” FLA. STAT. § 101.031(2). Inthe election challenged here, Sarasota County election officials failed to deliver on that promise.Indeed, the failure to count the votes of the thousands of Sarasota County voters who went to thepolls and cast votes in the Thirteenth District race is a miscarriage of the electoral process thatcan — and must — be remedied. These citizens should not forfeit their constitutional right tovote because the County’s paperless electronic voting machines malfunctioned. Yetdisenfranchisement is exactly what will happen unless the Florida Election CanvassingCommission’s certification is declared null and void.4. On behalf of herself and the thousands of her fellow Florida citizens facing suchdisenfranchisement, Contestant Christine Jennings therefore claims the right to thiscongressional seat and seeks appropriate relief under the Federal Contested Elections Act. It iscritically important that the United States House of Representatives — exercising itsconstitutional authority to “Judge . . . the Elections, Returns and Qualifications of its ownMembers,” U.S. CONST. art. I, § 5, cl. 1 — provide that relief promptly, by resolving that (a)there has been no valid election for the Representative in the One Hundred Tenth Congress fromFlorida’s Thirteenth Congressional District, (b) Contestee Vern Buchanan is not entitled to a seatas a Representative in the One Hundred Tenth Congress, and (c) the Governor of the State ofFlorida should be notified that the office is vacant, so that he can issue a Writ of Election to fill4the vacancy pursuant to Article I, Section 2, clause 4 of the United States Constitution andChapter 100 of the Florida Statutes.5. These remedies are exceptional, but they are by no means novel or precedentsetting.The House has never hesitated to grant the exact remedies requested by ContestantJennings when circumstances warrant such relief. In contested-election cases, the House hasfound the contestant to be entitled to the seat on 128 occasions. And the election has beenvoided, and the seat vacated, in another 66 cases. See, e.g., 78 CONG. REC. 1510-21 (1934)(agreeing to House Resolution 231, which provided that there had been no valid election, that thestate-certified winner was not entitled to a seat, and that the Speaker of the House should notifythe governor of the vacancy). Likewise, Florida law provides for analogous remedies. See, e.g.,FLA. STAT. § 102.1682(1) (providing for entry of a “judgment of ouster” against the contestee);Craig v. Wallace, 2 FLA. L. WEEKLY SUPP. 517a (2d Jud. Cir., Leon County 1994) (setting asideelection results and requiring a new election for state representative because irregularitiesprevented votes from being properly cast on three Votomatic machines).6. The current election result in Florida’s Thirteenth District cannot stand. Thevoters of the Thirteenth District — all of the voters, including those disenfranchised by machinefailure — should decide the outcome, and the proper remedy is therefore to hold a new electionin the district as promptly as possible. The resolution that Contestant Jennings requests here willensure that the will of the people of Florida’s Thirteenth Congressional District is respected, andwill restore the confidence of the electorate, which has been badly fractured by this machineinduceddebacle.5Grounds for Contesting the Election7. On November 7, 2006 (“Election Day”), the State of Florida conducted anelection for numerous offices, including Representatives in Congress. Early voting and votingby absentee ballot were permitted for this election (as for all elections in Florida).8. Both for early voting (from October 23 to November 5) and for Election Dayvoting (on November 7), Sarasota County made use of an electronic voting system, called the“iVotronic” touch-screen voting system, manufactured by Election Systems & Software, Inc.(“ES&S”), a privately held corporation. Sarasota County does not use the iVotronic electronicvoting system (or any other electronic voting machines) for absentee balloting. For absenteeballoting, Sarasota County uses paper ballots read by optical-scanning equipment.9. The first unofficial results reported on November 8, 2006 for the ThirteenthDistrict congressional race showed that in Sarasota County, there were 58,534 votes forBuchanan, 65,367 votes for Jennings, and 18,383 undervotes. The term “undervote” describes asituation in which a voter cast ballots for other candidates or ballot measures but did not registera vote for the particular office. See FLA. STAT. § 97.021(37).10. On November 13, 2006, the Elections Canvassing Commission ordered a machinerecount for the race pursuant to Section 102.141(6), Florida Statutes, because the difference invotes recorded for Buchanan and for Jennings was less than one half of one percent of the totalvotes recorded district-wide.11. On November 15, 2006, the Honorable Sue M. Cobb, Florida Secretary of State,released the results of the machine recount and ordered a mandatory manual recount pursuant toSection 102.166(1), Florida Statutes, because the difference in votes cast for Buchanan and forJennings was less than one fourth of one percent district-wide.612. As anticipated, neither the machine recount nor the manual recount altered thenumber of congressional undervotes recorded on the iVotronic system in Sarasota Countybecause merely “recounting” electronic ballots, unlike paper ballots (or absentee, overseas, orprovisional votes), is inevitably a meaningless exercise. The machine “recount” consists merelyof comparing the counters on the precinct tabulators with the overall election returns, and themanual “recount” consists simply of printing out the ballot-image reports from themalfunctioning iVotronic system and counting by hand the ballot images that recorded no choicefor the particular race in question. See FLA. STAT. §§ 102.141(6)(b), 102.166; FLA. ADMIN.CODE Rule 1S-2.031.13. By November 18, 2006, county canvassing boards in the three counties whollycontained in Florida’s Thirteenth Congressional District (DeSoto, Hardee, and Sarasota) and thetwo counties partly contained in the district (Charlotte and Manatee) had officially certified theirelection results and filed them with Florida’s Division of Elections. On November 20, 2006,Florida’s Elections Canvassing Commission, having compiled the official results from those fivecounties, certified the election returns and declared that Contestee Buchanan had been elected toCongress.
|
14.
|
The official results from the five counties were as follows:
|
|
|
Buchanan
|
Jennings
|
|
Charlotte:
|
4,460
|
4,277
|
|
DeSoto:
|
3,471
|
3,058
|
|
Hardee:
|
2,629
|
1,686
|
|
Manatee:
|
50,117
|
44,432
|
|
Sarasota:
|
58,632
|
65,487
|
|
____________________________________________________
|
|
TOTAL:
|
119,309
|
118,940
|
____________________________________________________ 715. As required by state law, the official returns from the five counties also reportedundervotes, which exhibited a sharply aberrant total for Sarasota County:
|
Undervote
|
|
Charlotte:
|
225
|
|
DeSoto:
|
142
|
|
Hardee:
|
265
|
|
Manatee:
|
2,324
|
|
Sarasota:
|
18,412
|
|
__________________________________
|
|
TOTAL:
|
21,368
|
16. Sarasota County, the one county carried by Jennings, accounted for barely half ofthe congressional candidates’ recorded votes district-wide, but fully 86% of the district’scongressional undervotes:
|
|
Buchanan
|
Jennings
|
Undervote
|
|
Sarasota County:
|
58,632
|
65,487
|
18,412
|
|
The Four Other Counties:
|
60,677
|
53,453
|
2,956
|
|
___________________________________________________________
|
|
TOTAL:
|
119,309
|
118,940
|
21,368
|
17. The undervote total for the congressional race in Sarasota County is extremelyabnormal in numerous respects, including the following:a. A total of 88,927 ballots were cast in this race on Election Day in SarasotaCounty on the electronic voting machines. Jennings received 39,930 votes and Buchananreceived 36,619 votes. There were 12,378 undervotes. The undervote rate on Election Day inSarasota County was therefore an extraordinary 13.9% of the ballots cast on the electronic votingmachines.b. A total of 30,832 ballots were cast during the early-voting process inSarasota County, on the same type of electronic voting machines. Jennings received 14,509votes, and Buchanan received 10,890 votes. There were 5,433 undervotes. The undervote rate8in the early-voting process in Sarasota County was therefore an extraordinary 17.6% of theballots cast. And the combined undervote percentage for early and Election Day voting on theelectronic voting machines was an equally extraordinary 14.9%.c. In vivid contrast, of the 22,613 votes cast in this race by paper absenteeballot in Sarasota County (which were recorded by optical-scanning devices, not by electronicvoting machines), Jennings received 10,981 votes, and Buchanan received 11,065 votes, andthere were just 566 undervotes recorded — an undervote rate of only 2.5%, which is consistentwith historical norms and expectations.d. In equally vivid contrast, the percentage of undervotes for the House ofRepresentatives race in other counties within the Thirteenth District did not remotely approachthe undervote rates for the electronic voting machines in Sarasota County. The undervote ratefor this race was 2.5% in Charlotte County, 2.1% in DeSoto County, 5.8% in Hardee County,and 2.4% in Manatee County. The combined undervote percentage for these four counties wasonly 2.5% — one-sixth the undervote percentage recorded in Sarasota County for votes cast onelectronic voting machines.e. In addition, the undervote percentage recorded in Sarasota County forother high-profile races is a small fraction of the 14.9% undervote rate on electronic votingmachines for the congressional race. For example, the undervote percentage recorded inSarasota County for the Governor’s race was 1.3% and the undervote percentage for the UnitedStates Senator’s race was 1.1%.f. Finally, the percentage of undervotes on electronic voting machines forthe congressional contest in Sarasota County in 2006 is almost seven times the rate of undervotes9for the Thirteenth District congressional race in 2002 (the last midterm election), which was2.2%.18. In 2001, Sarasota County became the first county in the State of Florida topurchase the iVotronic voting system. The system has been used since 2001 in at least 19separate primary, general, and local elections. In the 2006 election, Sarasota County voters wereasked whether to adopt a proposed county-charter amendment requiring that as of January 1,2008, all county voting systems provide a voter-verified paper ballot and that mandatoryindependent audits of election results be conducted in every election comparing hand counts tomachine counts. The county adopted the proposed charter amendment with the support of 55.4%of voters, indicating that voters themselves have lost confidence that the iVotronic system iscapable of correctly recording their votes. Significantly, the undervote rate for this proposedcharter amendment was only 6.2%.19. The statistical evidence alone indicates that the staggeringly large number ofundervotes in Sarasota County is due to the malfunctioning of the iVotronic electronic votingmachines. In fact, preliminary expert statistical analysis of the reported election resultsconcludes there is little doubt that the use of the iVotronic machines in Sarasota County causedthe extraordinarily high rate of undervotes in that county. The fact that undervote rates from therest of the district and from absentee voters in Sarasota County were so much lower than ratesfrom voters using the iVotronic machines in Sarasota County rules out the possibility that theextraordinarily high Sarasota County electronic undervote rate was caused by factors commonthroughout the district — such as voter abstention due to negative campaigning or dissatisfactionwith both candidates. Evidence that such alternative explanations were causing high undervoterates would have shown up throughout the district, not in a single county, and not just on one10type of voting machine in that county. Additionally, the fact that a higher undervote rate waspresent on identical electronic voting machines in two different modes of voting that occurred atdifferent times — early voting (from October 23 to November 5) and Election Day voting (onNovember 7)— creates an overwhelming likelihood that the problems pertain to the use of theseelectronic machines in Sarasota County.20. Compared to the malfunctions of the electronic voting machines, data availablethus far suggests that poor ballot design is a less likely explanation for the undervotes of thismagnitude. The most egregious examples of voter confusion caused by ballot design in otherraces have not yielded undervote percentages remotely as high as those present in the ThirteenthDistrict congressional race. For example, with the infamous “butterfly ballot” used in PalmBeach County, Florida in the 2000 presidential race, fewer than 1% of the voters erroneouslycast their ballots for the third-party candidate Pat Buchanan, and 4% of the voters erroneouslycast “overvotes” by selecting two or more candidates. If the ballot design here is in fact capableof disenfranchising nearly 15% of the electorate, that alone merits close attention from thisHouse.21. The poor ballot design’s most likely role here was as a contributing factor thathelped to trigger a software “bug” in the machines. Page 2 of the Sarasota County iVotronicballots contained both the two-candidate race for Representative in Congress and the sevencandidaterace for Governor — nine ballot lines in total. No other page of Sarasota County’selectronic ballot crammed so many candidates onto a single screen. Similarly, Page 3 ofCharlotte County’s iVotronic ballots contained both the seven-candidate gubernatorial field andthe two-candidate race for Attorney General. And, just as in Sarasota County, this designtriggered a bizarre undervote pattern, with a 0.8% undervote rate in the gubernatorial election11and an extraordinary 24.7% undervote rate in the Attorney General election that was on the samescreen. (In Sarasota County, the equivalent undervote figures for the two contests compressedonto one page were 1.3% and 14.9%.) In other counties around Florida, iVotronic ballot screensshowing nine candidates exhibited a similar pattern, with low undervote rates in the multicandidategubernatorial election and peculiarly high undervote rates in the two-candidateelection that shared the same screen.22. Some have theorized that this ballot design confused voters — especially“straight-ticket” voters who may have skimmed rapidly through the ballot, looking only forcandidates from one political party. If this is true, the magnitude of the undervote that thisproblem caused is still startling. But based on the data available to date, it appears a less likelyexplanation for the undervote than the machine malfunction itself. According to this theory, the“confused voter” (a) selected his party’s gubernatorial candidate while overlooking the othercontest on the same screen, (b) then moved quickly on to the next screen, and (c) at the end ofthe voting process, when faced with the “Summary Ballot” review screen, ignored the warningthat stated in red letters, “No Selection Made” for “U.S. Representative in Congress.” Only themost egregious ballot-design flaw would have so confused the intelligent voters of Sarasota andCharlotte Counties. And it is important to note that the pattern of low undervote rates in thegubernatorial contest and extraordinarily high undervote rates in the two-candidate contestdisplayed on the same screen held true regardless of whether the two-candidate contest appearedon the screen immediately above the gubernatorial field (as in Sarasota County) or immediatelybelow the gubernatorial field (as in Charlotte and other counties). That the undervote rate wasnot at least somewhat elevated for both offices suggests that this is (based on current data) a lesslikely explanation than the machine flaw; were ballot design the sole cause, one would not12expect the undervote rate to be elevated solely for the top-of-the-screen contest in SarasotaCounty and for the bottom-of-the-screen contest in the other counties. Indeed, implicit in the“confused voter” theory is the notion that Sarasota County voters read ballot screens from thebottom up while voters in Charlotte and the other counties read ballot screens from the top down,so that, in either case, the “confused” voters spotted their preferred gubernatorial candidate first,and then quickly moved on to the next screen, inadvertently bypassing the other contest (the topof-the-screen congressional race in Sarasota, the bottom-of-the-screen Attorney General’s raceelsewhere). While no empirical support has been found for this notion, evidence that it couldhave caused such unusually high undervotes would be extremely troubling. Regardless, thepattern of undervotes, combined with the actual ballot designs used in the various Floridacounties, suggests that an error or irregularity in the machines’ source code tends to convert intoundervotes some of the votes actually cast for candidates in the two-candidate field, wheneverthat field shares a screen with too many other ballot lines. Like any computer-based system, anelectronic voting touch-screen is most likely to malfunction when filled to capacity.23. As powerful as all the statistical evidence is, it is far from the only indication thatthousands of votes in Sarasota County simply were left out of the certified election results for thecongressional race because of the failure of paperless electronic voting machines. A variety ofcontemporaneous sources document widespread problems with the iVotronic electronic votingsystem in Sarasota County. These documents, including both the statements of voters andcontemporaneous records maintained by the Sarasota County Supervisor of Elections, identify aconsistent pattern of voter difficulty in having votes recorded in the House of Representativesrace — and not in other races on the ballot.1324. Contestant has obtained affidavits memorializing the eyewitness accounts ofhundreds of Sarasota County voters attesting to their difficulties attempting to cast a vote forChristine Jennings in early voting and on Election Day on the iVotronic electronic voting systemin Sarasota County. The following statements are representative of the memorialized eyewitnessaccounts of these hundreds of voters: “I went through the ballot making my selections on the IVotronics touchscreen voting machine and took my time making sure that I voted in everyrace. I am certain that I cast a vote for Christine Jennings. When Ireviewed the ballot at the end of the voting process, I noted that the racefor the 13th congressional district . . . indicated that I had made noselection. I double-touched the 13th Congressional District race and againcast my vote for Christine Jennings. . . . I have more than 15 yearsexperience in selling computer systems, five of those years are in sellingtouch screen systems. Based on my experience, I believe there was asoftware bug in the voting machine software causing the software not toregister the touch.” “I took a sample ballot, which I had previously filled out and my intentionto vote in every race. I believed that I voted for Christine Jennings but Icame to the review screen it said I had not cast a vote in the Congressionalrace. . . . I used the back arrow and it took me back to Congressional raceand I recorded a vote for Christine Jennings.” “When my husband and I voted on the iVotronics touch screen votingmachines, I was told by a poll worker to be sure and check the District 1314Congressional race because several voters, even at that early hour, hadcomplained that they had voted for Christine Jennings, but the summarypage did not reflect their votes for Christine Jennings.” “When I voted on the iVotronics touch screen voting machine I touchedthe screen for Christine Jennings and it showed I voted for ChristineJennings. But when I reviewed the summary page at the end of the ballot,it did not show a vote for Christine Jennings or anyone else.” “There was no warning or mention of any problems however, I was awarethere may be a problem with the Congressional vote based on variousmedia reports. I went through the ballot and specifically remember votingfor Christine Jennings. When I arrived at the review screen, there was nocandidate selected for the Congressional vote. I called a poll worker overand explained the situation and she told me that I did not ‘press hardenough’ when selecting the vote and I then returned to the vote screen andrecast my ballot, I then confirmed it on the review screen.” “When I voted on the touch screen voting machine I touched the screenvoting for Christine Jennings and when I reached page 15, the summarypage, it indicated that I had not voted for Jennings. I immediately calledthis to the attention of a poll worker who showed me how to go back andvote for Jennings. I followed her instructions and again voted forJennings. It did appear on the summary screen this time and I hope wasduly registered.”15 “When I voted on the iVotronics touch screen voting machine I touchedscreen and voted for Christine Jennings for U.S. Congress Florida District13. When I reviewed my ballot before hitting the red button and actuallyvoting, I saw the review screen did not show a vote for Christine Jennings.I was afraid I would lose my other votes if I tried to go back and correctthe problem, so I then went ahead and cast my ballot without confirmingthat the machine had registered my vote for Christine Jennings.” “I attempted to vote for Christine Jennings in the District 13 race andexperienced the following difficulties: I was well-aware of the difficultiesin the early voting in District 13 race and so I carefully voted in eachelection on the ballot, including that race. When I got to the review page,my vote for Christine Jennings was not reflected. I called out to a pollworker to alert them that my vote in the District 13 race had not beenrecorded. The poll worker who came to assist me informed me that thesame thing had happened to her when she had voted earlier. She guidedme back to the District 13 page and I pressed the touch screen again toreflect my vote for Christine Jennings. The poll worker then guided meback to the review page where my vote in the District 13 race wasreflected and I then pressed the vote button.” “When I voted on the iVotronics touch screen voting machine, I wentthrough the ballot to vote. I was being careful because I seemed to have topress hard for my votes to register. In addition, I knew to be carefulbecause my wife had been to vote previously and had overheard some16women who had a problem voting discussing their problems with themachines. They were different machines. A neighbor also told me thatshe had encountered six different people who had a problem with thevoting machines. When the review sheet came up it said that I had notvoted in the Congressional race even though I knew I had voted forChristine Jennings. I went back and registered my vote again and thistime it indicated that I had voted for Ms. Jennings on the review screen.” “When I voted with the stylus on the iVotronics touch screen votingmachine, I am absolutely sure the box for Christine Jennings showed theX. On the Review screen, however, Christine Jennings’ name showed butthe box beside her name was blank. I clicked on the review ballot andcorrected my vote and it then showed an X beside her name. After that, Iregistered my vote with the Red button at the top of the screen. Aftervoting, I asked my husband if anything unusual happened when he voted(on a different machine). He told me that when he reviewed his ballot, thebox by Christine Jennings’ name was blank and he had to correct it. Atthat time, I reported this to a poll worker named Charlie, who said hewould report it.” “I had heard prior to going to the poll that there were problems with thevoting machines. When I went to vote, the poll worker also warned methat there had been problems with the machine registering theCongressional race. When I voted on the iVotronics touch screen votingmachine, I voted for Christine Jennings. The screen indicated I had voted.17Yet when I got to the end, the review page indicated that I had not voted inthe Congressional race. I went back and voted for Ms. Jennings. Thistime my vote did register on the voting page.” “When I voted on the iVotronics machine I was being very methodical.When I voted in the Buchanan-Jennings race, I specifically voted forChristine Jennings and checked to make sure that the box was checkedbefore I went to the next page. When I got to the review screen itreflected no vote was cast for the Congressional race, but both candidates’names were shown. All of my other selections were properly recorded. Itouched where it said no vote had been cast and it took me back to theBuchanan-Jennings race. I then re-voted for Christine Jennings andcarefully rechecked the review page three times. I then pushed the votebutton. No report was made to the poll worker. Prior to voting, the pollworker recommended that I check the review page before casting my finalballot. I am a registered Republican and I believe these machines faileddemocracy.” “I voted on the iVotronics machine I took my time to be sure I did notmake any errors. When I voted in the Buchanan-Jennings race, Ispecifically voted for Christine Jennings and checked to make sure the boxwas checked before I went to the next page. When I got to the reviewscreen it reflected no vote was cast for the Congressional race. All of myother selections were properly recorded. I touched where it said no vote18had been cast and it took me back to the Buchanan-Jennings race. I thenre-voted for Christine Jennings and I then pushed the vote button.” “When I voted on the iVotronics touch screen voting machine I touchedthe screen for Christine Jennings and it showed I voted for ChristineJennings. But when I reviewed the summary page at the end of the ballot,it not only failed to show a vote for Christine Jennings, but the only nameto appear on the review page was Christine Jennings, next to a blank boxindicating no vote had been cast. I called a poll worker over andexplained what had happened and the poll worker pulled back the page forthe Congressional race. I revoted for Christine Jennings, and my voteappeared to register in my second review of the summary screen.” “When I voted on the touch screen voting machine I encountered twoproblems with the machine. First, after I had voted for Christine Jenningson the top of the second screen, when I pushed my selection for Jim Davisfor Florida Governor next, the ‘X’ on the computer screen came upindicating that I had voted for Charlie Crist. I called a poll worker,advised her of the problem and she showed me how to change my vote toJim Davis. I then proceeded to vote on every race I saw on the ballot.When I got to the review screen, it showed Christine Jennings name, butunlike all the other names and races on the review screen, there was no Xin the box next to Christine Jennings’ name. I am certain that I hadinitially cast a vote for Christine Jennings as my two main purposes invoting were to vote for Christine Jennings for Congress and Jim Davis for19Florida Governor. I again called a poll worker who told me to hold myfinger down on the box next to Christine Jennings name on the reviewscreen until the X came up. I did so and then pushed the ‘Vote’ button.” “When I arrived at the polls I was warned by a poll worker that somevotes from ‘page 2’ were not being registered. I waited on line for 45minutes to vote and when I returned home, informed my wife of what Ihad been warned.” “I had heard earlier media reports and was aware that there were someproblems with the machines. When I arrived, I specifically asked if therehad been problems and I was told no issue or problems had arisen. I votedfor Christine Jennings on a touch screen and when I arrived at the reviewpage the Congressional vote was left blank. I called a poll worker over atthat time and she showed me how to move back and I re-cast my vote forChristine Jennings. On the final review page, I confirmed my vote wascast. I approached a poll worker to complain about the situation and filledout a complaint card.”25. Poll watchers also reported their observations of widespread occurrences of votersbeing unable to get their votes in the congressional race properly recorded by the iVotronicelectronic voting machines. One poll watcher reported as follows: “There were seveniVotronics touch screen voting machines at the precinct where I was watching the voters. Twoof the iVotronics touch screen voting machines stopped working while I was watching thevoters. After an hour or so, one was repaired and put back into service. The other was put back20into use without repair except that the poll workers instructed voters to hold their finger on thetouch screen for more time, rather than just touch [the] screen to get the vote to register. I heardseveral voters tell poll workers the iVotronics touch screen voting machine was not recordingtheir vote.”26. Contemporaneous official “Incident Report Forms” of the Sarasota CountySupervisor of Elections likewise document widespread occurrences of voters having greatdifficulty in getting the iVotronic electronic voting machines to record their votes in theThirteenth District race. Numerous such forms noted that iVotronic electronic voting machineswere “not recording votes.” One report from a particular precinct noted that a “voter voted onscreen — didn’t show up on review . . . asked poll worker for help . . . [c]ancelled ballot andmoved to another machine,” and went on to observe “more than one [voter] with trouble onmachine.” Another incident report observed that “[e]very other voter is complaining about theChristine Jennings contest not coming up.” Indeed, these incident reports document multipleinstances of frustrated voters telling election officials at the polling places that “votingmachine[s] would not let her vote for Jennings.”27. Other contemporaneous official forms maintained by the Sarasota CountySupervisor of Elections similarly document that iVotronic electronic voting machines used in thecounty were not recording the votes that voters had cast. Machines were taken out of service onElection Day because they were “slow to respond to touch” or “required a hard/extended touchbefore [a] vote was recognized,” or because they were “not recording some votes [and] thetouchscreen was not working properly — hard to record vote, needed to push hard and juggle torecord vote,” or because they were “not accepting votes.” Technical-support personnel reportedreceiving “several complaints that voters make selections that do not appear on the summary21screen” and that “the selection has to be highlighted . . . two or three times before the summarypage reflected the suggestions.” Other reports indicate that “voters reported making a selectionbut the selection did not appear on the review screen,” requiring further corrective action by thevoter, and that particular machines “miss[] selections on some pages.” One report by a SarasotaCounty technical-support person indicated that a particular electronic voting machine “will notregister votes no matter how hard you press screen.”28. Significantly, the records of the Sarasota County Supervisor of Electionsdocument that election officials were on clear notice, as a result of the extreme difficulties manyvoters encountered during the early-voting phase, that the iVotronic electronic voting machineswere malfunctioning with respect to the Thirteenth District congressional race. Nevertheless, theCounty election officials do not appear to have taken any steps to correct the serious machineproblems in advance of Election Day.29. The eyewitness accounts of the voters, poll watchers, and election workersdocumented here, as well as hundreds of others like them, attest to pervasive difficulties in therecording of votes in the Thirteenth District congressional race. Although many voters believedthat they were able eventually to overcome the machine difficulties and cast a recorded vote forContestant Christine Jennings, the problems the iVotronic system exhibited in recording thevotes of these and thousands of other voters provide substantial grounds for doubting whether thevotes were in fact counted. The information voters see on the touch-screen of an electronicvoting machine when they cast their votes is stored in the machine’s temporary, volatilecomputer memory. A permanent record of a vote is made only when — upon pressing the red“Vote” button above the screen — the voter’s recorded preference is transferred from thecomputer’s temporary volatile memory to its permanent nonvolatile memory. If, as the statistical22evidence suggests is overwhelmingly likely, a software “bug” or other malfunction disrupts orprevents the transfer of the recorded vote from temporary to permanent memory, the voter maywell see a vote cast for Jennings on his or her review screen even though no permanent record ofthe vote is ever recorded.30. The probability of machine error finds support in a statistical analysis conductedby Professor Charles Stewart III, the chair of the Political Science Department at theMassachusetts Institute of Technology (MIT). Professor Stewart’s analysis indicates that thedate when an iVotronic machine was “cleared and tested” by Sarasota County election workersor their contractors (as reflected by “Event Code 01” in the machine’s audit log) correlatesstrongly with the machine’s undervote rate: The machines prepared in the final days before thedeadline for completing all such preparations exhibited the highest congressional undervoterates. And another strong correlation exists between the number of machines “cleared andtested” on a given date and the undervote rate: As the County’s staff or consultants got busier,clearing and testing more machines on a single day, the congressional undervote rate climbed.31. The following graph demonstrates these facts. It shows the undervote rates forthe iVotronic machines that were prepared on each date leading up to the election. The area ofeach data “bubble” is proportional to the number of machines prepared that day, so a large circleindicates a busy day of machine preparation. Dark bubbles are days when Election Daymachines were primarily prepared; light bubbles are days when early-voting machines wereprimarily prepared.23As this graph shows, the congressional undervote rates were below 7% for machines setup on only three days — September 19, October 2, and October 5, 2006. On each of those threedays, the Sarasota County election workers cleared and tested only one machine. By contrast,the County’s busiest day of machine preparation — October 17, 2006— involved setting up 158machines, and on Election Day those 158 machines generated a congressional undervote rate ofnearly 21%.Figure 1. Undervote rate in the Thirteenth Congressional District race among machinesprepared on the same day. (The area of bubbles is proportional to the number ofmachines prepared on that day.)Clear and test date9/199/259/269/279/289/2910/210/310/410/510/610/910/1010/1110/1210/1310/1610/1710/1810/1910/2010/2310/2410/2510/2611/50.0%5.0%10.0%15.0%20.0%CD13 undervote rateElection dayElection day machineEarly voting machineNote: One early voting machine was prepared on 10/24; two were prepared on 10/25.2432. Professor Stewart also has analyzed the effect of machine-induced failure on theoutcome of the election for the Thirteenth District congressional seat. Based on his study ofpatterns in the undervote rates for other statewide or countywide races in Sarasota County,Professor Stewart estimated that the number of “excess” undervotes caused by the use of theiVotronic machines in Sarasota County was approximately 14,000. This is a conservativeestimate, as it suggests that slightly more than 3% of the Sarasota County voters intended not tovote for either congressional candidate, which is more than double the actual undervote ratefound in the November 2006 race for U.S. Senator or Governor. Professor Stewart’s estimatewas corroborated by the expert for the iVotronic machines’ manufacturer, who wrote that he was | |
Local News