about 10 mins. 57 secs., watch JFK Jr's. expression :) at the mentioning of Richard.
IN THE UNITED STATES DISTRICT COURT
FOR THE SOUTHERN DISTRICT OF OHIO
KING LINCOLN BRONZEVILLE :
NEIGHBORHOOD ASSOCIATION, et. al. :
: Case No. C2 06 745
v. : Judge Algenon L. Marbley
J. KENNETH BLACKWELL, et. al. :
DECLARATION OF RICHARD HAYES PHILLIPS
Pursuant to 28 U.S.C. § 1746, I, Richard Hayes Phillips, declare the following:
1. I have personal knowledge of the facts alleged herein.
2. I reside at 4 Fisher Street, Canton, New York. I hold a Ph.D. in geomorphology from the University of Oregon. Both my graduate and post-graduate work in groundwater hydrology required that I examine reams of statistical evidence in order to spot anomalous data. I have been recognized as an expert witness in Federal District Court in Albuquerque, New Mexico, and at Resource Conservation and Recovery Act (RCRA) hearings in Santa Fe, New Mexico, both in regard to potential groundwater contamination due to nuclear waste disposal. I have published five professional papers on the nuclear industry in Oklahoma, and have written or co-authored eleven professional papers on nuclear waste disposal in New Mexico for administrative proceedings in the United States and Mexico.
3. I also have a long-standing interest in history and politics. I hold a B.A. in political science from the State University of New York at Potsdam, and an M.A. in history from the University of Oklahoma. I have worked in five presidential campaigns, including three New Hampshire primaries, and I taught American history at three colleges and universities in New Mexico for three years.
4. After the 2004 election, I examined precinct canvass results for 9,343 of 11,323 precincts in the State of Ohio, identifying anomalous or irregular election results from the 2004 presidential election, utilizing the same analytical techniques which have enabled me, in the past, to detect unusually rapid groundwater flow or
groundwater contamination. In December 2004 I was retained as an expert witness in the Moss v. Bush lawsuit, for which I submitted 21 research papers to the Ohio Supreme Court, was deposed for 4 hours, and was not even questioned by opposing counsel. The Columbus Free Press has published more than 40 of my papers on the 2004 election on the worldwide web, subject to peer review and criticism. My research papers were a primary source for the report by Congressman John Conyers on the 2004 presidential election. I have been an invited speaker at national conferences on election reform in Nashville, Columbus, and Houston. With Louis Harris and others, I was a consultant to Rolling Stone magazine in preparation for a major article by Robert F. Kennedy, Jr. on the 2004 election; I made a trip to Ohio specifically to gather evidence in response to questions by Contributing Editor Tim Dickinson; my statistical analyses were the source of the data for the map of suspect counties; and my name is frequently referenced in the electronic footnotes. I was a primary source for an upcoming book on the 2004 election published by the New Press, which incorporates several of my research papers, in whole or in part. Under the auspices of the Columbus Institute for Contemporary Journalism and the Ohio Honest Elections Campaign, and with the assistance of a team of citizen volunteers, I have recently collected and examined photographs and photocopies of poll books, voter signature books, and tens of thousands of ballots from eleven counties in Ohio, and have found evidence of ballot tampering in all of them. My investigation is ongoing.
5. Shortly after the 2004 presidential election, I examined the unofficial county-by-county election results posted online by Ohio Secretary of State J. Kenneth Blackwell. Unable to investigate all 88 counties in Ohio due to time and resource limitations, I identified seven counties where the numbers were so irregular as to warrant further scrutiny. These were Butler, Clermont, Cuyahoga, Delaware, Hamilton, Miami, and Warren counties.
6. I also studied, early on, the effect of a shortage of voting machines in Franklin County. I meticulously compared election results with the number of registered voters per voting machine for each precinct in the City of Columbus. The median Kerry precinct had a 50.78% turnout, and the median Bush precinct had a 60.56% turnout. In my professional opinion, the lower turnout in the wards won by Kerry was due directly to a shortage of voting machines. I calculated what the citywide vote totals would have been if all wards had enjoyed a turnout of 60% or more. It is my conclusion that John Kerry’s margin of victory in the City of Columbus was wrongly reduced by 17,000 votes.
7. In Cuyahoga County, irregularities included precincts where hundreds of votes turned up in the columns of third-party candidates; precincts where Kerry votes appeared to have been shifted to Bush or to the column left void when Ralph Nader was disqualified from the ballot; and precincts with less than 30% voter turnout. In both Cuyahoga and Hamilton counties, there were many heavily Democratic precincts with very high percentages of undervotes and overvotes; I subsequently discovered the same pattern in Montgomery, Stark, Summit, and Trumbull counties.
8. In the other five counties (Butler, Clermont, Delaware, Miami, and Warren), I found the precinct-by-precinct canvass results to be highly improbable when compared to the 2000 presidential election (Bush vs. Gore), or the 2004 Chief Justice contest (Moyer vs. Connally), or the 2004 ballot initiative banning gay marriage (Issue One). Specifically, it is unlikely that Kerry received fewer votes than Gore, in light of the huge increase in voter turnout and the absence of Ralph Nader from the ballot. It is unlikely that Bush got all the new voters, all the Nader voters, and some of the Gore voters, or the equivalent. It is unlikely that Kerry received fewer votes than Ellen Connally, an underfunded, African-American, municipal judge from Cleveland running for Chief Justice against Thomas Moyer, the incumbent, in a down-ticket race that drew 1.2 million fewer votes statewide than the presidential race. One would expect Bush to receive more votes than Moyer, and Kerry to receive more votes than Connally. And it is unlikely that Bush received more votes than Issue One, which passed with 61% of the vote, compared to 51% for Bush, statewide. One would not expect thousands of gay marriage supporters to vote for Bush. Moreover, these voting patterns were anomalous, having occurred in relatively few counties in Ohio.
9. I have subsequently investigated reports of voter suppression, failure to count all the ballots cast, and possible alteration of the vote count in many other counties in Ohio. But the eleven counties listed above have drawn my repeated attention, and so I submitted precinct-specific public records requests on behalf of the Ohio Honest Elections Campaign, beginning April 3, 2006, to the Boards of Elections in all eleven counties.
10. I now have photographs or photocopies of poll books, voter signature books, and/or tens of thousands of actual ballots (punch cards or optical scans), from precincts of my choosing in all eleven of these suspect counties. I have documentation of ballot tampering in all eleven counties, and I have observed lax security procedures that leave the forensic evidence vulnerable. At this time I can only summarize the evidence, or give examples, because evidence is still being analyzed, the findings are preliminary, and the investigation is ongoing.
11. In Butler County, a Board of Elections employee volunteered the information that "the punch cards have always worked well for us," and "the tabulators have always worked well in the past," but that on the morning of Election Day, November 2, 2004, a technician from ES&S (Election Systems & Software) came to the Board of Elections office and "reprogrammed all of our tabulators." It is my understanding that the technician, whose name I hope to ascertain, was uninvited. The employee stated that there were six tabulators (four were utilized, and two were on standby). Photocopies of ballots made by an intern revealed, in one precinct (Monroe City 4CA), 52 consecutive ballots for Bush near the top of the stack, and 212 consecutive ballots for Bush near the middle of the stack.
12. In another precinct in Butler County (Ross Township 4JB), there were runs of 41, 29 and 25 consecutive ballots for Bush, and never more than 4 in a row for Kerry. According to an e-mail from the Butler County Board of Elections, these were not among the 10 precincts, constituting 3% of the ballots cast, which were hand-counted in the "recount" of December 15, 2004. It is highly improbable that this is the order in which the voters lined up at the polls on Election Day. It is more likely that these ballots have been sorted by hand. There is no legitimate reason for these ballots to have been sorted because the counting was done on an electronic tabulator on a tight deadline (the unofficial results were certified at 11:00 P.M. on Election Night). In Ross Township 4JB, either Bush received the votes of 70% of the Ellen Connally supporters, 60% of the supporters of gay marriage (opponents of Issue One), and 50% of those who voted for Ellen Connally and against Issue One, or else not all of the ballots are legitimate.
13. The Butler County Board of Elections was able to produce the unused punch cards. They also produced, in the form of a packing slip, what appeared to be a complete inventory for each precinct. The punch cards were delivered to the Board of Elections in shrink-wrapped pads of 100 or more. For two precincts (Middletown 3BC and 3BI) we examined every unused punch card except those still shrink-wrapped. All the punch cards have stubs that are numbered sequentially. [ See EXHIBIT A and EXHIBIT B ] All were accounted for, and all were pristine. This only raises the question as to why the Boards of Elections in Clermont, Hamilton, Montgomery, Stark, Summit, Trumbull, and Warren counties cannot produce their unused ballots.
14. In Clermont County, eleven weeks of negotiations with the County Prosecutor’s office were required in order to gain access to the public records, and even then we were actively discouraged from completing our tasks in an expeditious manner. On each scheduled visit we were denied access to certain records and told to return another day. In the aftermath of the recount of December 2004, sworn affidavits were signed by three witnesses stating that they observed, on some ballots, white oval stickers over the Kerry-Edwards mark, and the Bush-Cheney oval filled in. The County Prosecutor’s office confirmed this in writing, stated further that there were fewer than 100 such ballots countywide (out of 89,822 ballots cast), and claimed that there had been an FBI investigation into the matter upon the request of Congressman John Conyers, but that the FBI report could not be located.
15. In Clermont County, in the first 4 precincts that we photographed, with 3589 ballots cast, we saw not one stickered ballot. Board of Elections personnel have confirmed that these 4 precincts were among the 97% of the ballots that were recounted by machine in December 2004, and that the observers who reported seeing stickered ballots were observing the 3% of the ballots that were hand-counted. The Board of Elections provided a statement, in writing, that 14 precincts were chosen for the hand count because they were the smallest in the county, with only 2696 ballots cast among them. Ohio Secretary of State J. Kenneth Blackwell’s directive at that time required that the hand-counted precincts be chosen at random. We have now photographed ballots from 20 precincts, including the 14 hand-counted precincts,
and have seen only one stickered presidential ballot, in Pierce Township H. [ See EXHIBIT C and EXHIBIT D ] In this case, there is a white oval sticker over the Kerry-Edwards mark, and the Bush-Cheney oval is filled in with a distinctively different pencil than was used on the rest of the ballot. Also in Pierce Township H we have digital photographs of 6 ballots rubber-stamped with the word "DUPLICATE." The Board of Elections has been unable to provide the original ballots or the stickered ballots from which these are said to have been duplicated, and thus the validity of the duplicate ballots remains unverified.
16. After an initial refusal by Clermont County Board of Elections personnel, we photographed the contents of a cardboard box labeled "Voided Ballots (that were duplicated) for duplication of 7000 Series 11-2-04 Election." Numerous Board of Elections personnel stated that these were absentee ballots from precincts in western Clermont County that were remade by the Board of Elections, but no one could remember why. The box contained 60 voided ballots for the presidential election, including 46 from Miami Township F1F (for which we hope to photograph the duplicate ballots), 5 from Miami Township B (for which I have learned from Ron Baiman, by personal communication, that there are at least 36 duplicate ballots, which he has videotaped, and which we hope to photograph), and 1 from Pierce Township H (for which there are 6 duplicate ballots, which we have photographed).
17. Clermont County Board of Elections personnel volunteered the information that absentee ballots, when returned, are stacked up "for a few days" beneath a telephone book in order to flatten out the folds that ensue from their having been sent through the mail. Only on our third visit, when I asked again about this procedure, did anyone say that these ballots were placed in a lock box. The security of the absentee ballots is an important question. In at least four precincts, there are too many absentee ballots in the stacks. According to Board of Elections personnel, absentee ballots and provisional ballots are distinguishable from regular ballots because they are rubber-stamped with a red "smiley face" or with the letters "EO" also in red. Even if one assumes that every absentee ballot mailed was returned by the voter, and that every provisional ballot was ruled valid, there are too many rubber-stamped ballots (32 in Miami J, 4 in Miami K, 11 in Batavia K, and 12 in Goshen N). It is my conclusion that the "EO" stamp does not signify absentee ballots, but when questioned, no one at the Board of Elections could offer an alternative explanation.
18. My records requests to Clermont County specifically asked "to view and photograph all the optical scan ballots" (emphasis added). To date we have not seen any of the unused ballots, which are necessary for a complete audit. All the ballots delivered by Dayton Legal Blank must be accounted for. There are no boxes labeled "spoiled" or "unused" on the shelves with the voted ballots. Ohio Public Records Law, as interpreted in Directive 2004-43 from J. Kenneth Blackwell, dated October 25, 2004, specifically states: "All used and unused ballots must be retained for at least 22 months." On August 24, 2006, in response to my most recent inquiry about the whereabouts of the unused ballots, I received this e-mail from Mike Keeley, Director, Clermont County Board of Elections: "As time permits, we are
still attempting to locate said un-used ballots. When they are located you, and numerous other requestors, will be so notified." My initial public records request to Clermont County was dated April 3, 2006. As of this writing, the Clermont County Board of Elections has had twenty weeks to locate the unused ballots.
19. In Delaware County we met with resistance when we asked to look at punch card ballots. Private police were summoned five times, and the County Prosecutor’s Office was summoned twice. Shortly before 4:30 P.M. on April 3, 2006, when we arrived at the Rutherford B. Hayes Administration Building after one week’s absence, an armed, uniformed policeman working for Global Security Services ran inside the building and into the Board of Elections office, locking the door behind him during business hours. According to the Global Security Services website, "GSS is comprised of former Secret Service and FBI agents, CIA operatives, military special operations personnel such as Navy SEALs." [ www.gts-gss.com ] We were shortly thereafter let into the Board of Elections office, whereupon we submitted two more public records requests in writing. According to Board of Elections telephone records, which we have obtained from the Delaware County Auditor, a fax was sent to 937-435-8352 at 4:33 P.M. on April 3 (duration 0.6 minutes). This is the fax number for Dayton Legal Blank, which supplies ballots to Delaware County. [ http://www.dlbinc.com/ ]
20. Dayton Legal Blank is owned by Election Systems and Software (ES&S), through two acquisitions: in November 1985, Business Records Corporation acquired Dayton Legal Blank; and in 1997, ES&S purchased Business Records Corporation. [ http://www.blackboxvoting.org/bbv_chapter-8.pdf ] In November 2004, ES&S was the voting machine vendor for Butler, Clermont, Cuyahoga, Delaware, Hamilton, Miami, Stark, and Trumbull counties, among other counties in Ohio.
[ http://verifiedvoting.org/verifier/map.php?&topic_string=5std&state=Ohio ]
21. On Friday, June 9, 2006 I learned from Kim Spangler, Director, Delaware County Board of Elections, that the Board manufactures its own voter signature books, and that the Board had already ordered its ballots for the May 2, 2006 primary prior to April 3, 2006, so these cannot be the reasons for the fax.
22. On two subsequent occasions we have been able to view ballots for precincts of my choosing in Delaware County, with both a Republican and a Democrat present to handle and photocopy the ballots for us. On the first occasion we examined the ballots for Orange K. There were 85 ballots, said to be absentee ballots, identified by precinct not by the printer, but by handwriting. We have no way of verifying whether these ballots were identified by precinct before or after the ballots were cast. The rest of the ballots appeared to be in random sequence. When I asked to see the unused ballots for Orange K, Jeff Burkam, a Board of Elections member, told me that he "cannot produce them." He further stated: "I don’t know if we still have them two years later." Kim Spangler has since confirmed by e-mail on August 23, 2006, and by telephone on August 28, 2006, that the Delaware County Board of Elections does still have the unused ballots for the November 2, 2004 election.
23. On the second occasion we examined the ballots for Genoa I, having amended our
records request the previous afternoon to include this precinct without notice. Board of Elections employees read out loud the numbers punched for president, and we kept running tallies of how many consecutive ballots were punched for each candidate. The ballots for Genoa I were bound in three separate bundles. In the second bundle there were 75 consecutive ballots for Kerry, followed by 274 consecutive ballots for Bush, and in the third bundle there were 359 consecutive ballots for Bush. [ See EXHIBIT E and EXHIBIT F ] It is highly improbable that this is the order in which the voters lined up at the polls on Election Day. It is more likely that these ballots have been sorted by hand. Genoa I was not one of the 4 precincts, constituting 3% of the ballots cast, which were hand-counted in the "recount" of December 15, 2004. We have obtained the minutes from the Board of Elections specifying that the hand-counted precincts were Harlem A, Genoa C, Troy B, and Porter B.
24. There are indications that the election results in Delaware County may have been programmed in advance. (The technician at the recount in Delaware County in December, 2004 is identified in the minutes of the Board of Elections as "Sam Hogsett, Technical (sic) for Election Systems and Software." Before witnesses, his address was given as Crown City, Ohio). There were three contests for seats on the Ohio Supreme Court. The losing Democrats, all of whom ran competitive races, were Ellen Connally, Nancy Fuerst, and William O’Neill. In the official precinct canvass results, if one numbers all 123 precincts in sequence, O’Neill ran behind Connally and Fuerst in nearly all the odd-numbered precincts, and ahead of Connally and Fuerst in nearly all the even-numbered precincts. Only 11 of 123 precincts deviated from this pattern. It is highly unlikely that this odd-even pattern reflects the true voting pattern in Delaware County. It cannot be wholly attributable to ballot rotation. William O’Neill has confirmed by telephone that he did not campaign only in the even-numbered precincts. By comparison, O’Neill ran ahead of Connally in only 8 of 289 precincts in Butler County.
25. The chain of custody on November 2, 2004 could have allowed for ballot tampering in Delaware County. According to Board of Elections personnel, whichever party wins the precinct in the gubernatorial election gets to choose the precinct judge for the next four years. The precinct judges, unaccompanied, bring the ballots to the Board of Elections after the polls close on Election Night. In Delaware County, almost all of these would be Republicans, as Bob Taft, the Republican, carried Delaware County with 71.9% of the vote to 23.5% for Timothy Hagan, the Democrat, winning 121 of 123 precincts. On November 2, 2004, the ballots were driven to an underground parking area, where they were unloaded by teams of teenaged volunteers, including Boy Scouts, and carried into the Board of Elections building where, again according to Board of Elections personnel, a "mentally retarded man" scraped the chads from the ballots before they were run through the tabulator. For the record, even now, 20 months after the election, we are not allowed to touch the ballots in Delaware County. They must be handled by county
employees, one Republican, one Democrat. Moreover, there were 79,691 ballots cast on Election Day and, according to the minutes of the Board of Elections, the unofficial results were certified at 12:40 AM, only 5 hours and 10 minutes after the polls closed. The "mentally retarded man" would have had to examine, and clean as needed, four or five ballots per second, which calls the whole story into question.
26. The Delaware County Board of Elections has not been able to account for the 100,676 voters officially reported to have been registered on November 2, 2004. The number of names of registered voters provided on spreadsheets by the Board of Elections is short by 6,234. There were no more than 94,442 names on the voter rolls on November 2, 2004. The Board of Elections has confirmed in writing that 7,184 voters were purged from the rolls on April 8, 2005, and Kim Spangler stated on tape that these voters could have been purged before the November 2, 2004 election. Of the 87,258 active voters whose names should have been on the rolls, 81,175 (93.03%) are officially reported to have voted, a 45.06% increase in ballots cast compared to the 55,959 who voted in the 2000 presidential election. This calls the official number of ballots cast into question.
27. In Miami County, the voter turnout percentages in the official results have always been suspect. I am on record before the Ohio Supreme Court as saying "There simply was not a 98.55% turnout in Concord South West precinct or anywhere else in Ohio. Nor was there a 94.27% turnout in Concord South precinct. … Nor do I believe the reported turnout of 56.55% in Concord South East Precinct, which is no doubt contiguous." On Monday, June 19, 2006, Steve Quillen, Director, Miami County Board of Elections, handed us a printout of new turnout data for each of 82 precincts, calling it the "freely amended results." Quillen’s printout, drawn from a computer database of voter histories, indicates that the turnout was 82.1% in Concord South West, 79.5% in Concord South, and 83.3% in Concord South East. Quillen explained that on Election Day, some precincts with heavy voter turnout were at risk of running out of ballots. Volunteer Boy Scouts were dispatched to deliver more ballots to these precincts, but delivered the wrong ones. The explanation is plausible on its face, because optical scan ballots are marked along the left-hand margin with a bar code that identifies the precinct, and the candidates’ names are listed on the ballot. If a voter casts a ballot for the wrong precinct, the tabulator should count the vote for the correct candidates, but attribute the votes to the wrong precinct.
28. If the Boy Scout explanation is true, then some voters in a precinct with low reported turnout must have filled out ballots intended for other precincts, which then would have high reported turnout. I expected to find ballots for Concord South West (with an official turnout of 98.55%) and Concord South (with an official turnout of 94.27%) mixed in with the ballots for Concord South East (with an official turnout of 56.55%). That is not what I found. It was the stack for Concord South West, where all but ten registered voters (679 of 689) reportedly voted, that contained ballots from the wrong precincts, 200 in all (127 from Concord South, and 73 from Concord South East). These extraneous ballots could not have been
cast in Concord South West, because there were not enough voters to cast them. According to the voter signature book there were 692 registered voters in Concord South West, and no more than 556 of them voted. On pages 20 and 21 alone, there are 13 registered voters listed who did not vote. [ See EXHIBIT G ] If ballots for the wrong precincts were cast in Concord South East, as the Boy Scout explanation would require, then ballots for all three precincts, more than 300 ballots in all, must have been moved to the Concord South West stack sometime after the ballots were cast. Altogether, I found 1884 ballots, including absentee ballots and "remakes," for these three precincts combined. According to the poll books, there were 2003 ballots cast in these three precincts, so there are 119 ballots unaccounted for, 111 of them being absentee ballots.
29. It is possible that the Boy Scout explanation is a cover story, and that an error was made when the evidence was doctored. This would have been easy to do, because the ballots are stored in a completely unsecured location. They are stored in the Power House across the street from the County Court House. Only one key is required to unlock the door, and numerous persons, including maintenance personnel, have the key. On Monday, June 26, 2006, a custodian approached us and asked if we had the key, or if we needed Steve to lock up for us. On Tuesday, June 27, 2006, a key with a tag labeled "Power House" was left unattended all day long in plain sight on a chair where the records are stored. We photographed the key in place, but did not touch it. At the end of the day there was no one to lock the door, so I left a colleague to guard the entrance to the building while I searched for a county employee to lock the door, and we did not leave the premises until the Power House was securely locked.
30. At the recount of December 16, 2004, three precincts containing more than 3% of the ballots cast in Miami County were hand counted. If the hand count had failed to match the tabulator count, a full hand count of the entire county would have been required. I found that the ballots for the hand-counted precincts (Piqua 2-A, Troy 3-E, and Bradford) were sorted according to the candidate. (For example, the ballots for Bradford were stacked as follows: 116 for Kerry, 119 for Bush, 2 for Peroutka, 4 Undervotes, 79 for Kerry, and 120 for Bush). Our counting of ballots for all three precincts was recorded on videotape. It is highly improbable that this is the order in which the voters lined up at the polls on Election Day, and that these sequences can only be attributed to ballot tampering, especially in light of the fact that the Green Party website, on which observations of the witnesses to the recount are posted, contains this statement concerning Miami County: "Note of irregularity: ballots seemed to be already sorted by candidate – how would this occur?" [ http://www.iwantmyvote.com/recount/ohio_reports/counties/miami.php ]
31. The Green Party website also states, referring to Miami County, that "a significant number of spoiled ballots (about 500) that could not be read by the optical scan machine had to be remade by B.O.E. staff with the intent of the voter being replicated to the best of the staff member's ability. It should be noted that this is a subjective process and that in the small sample we reviewed, there were
questionable choices made previously during the initial counting by B.O.E. staff members in their attempt to transfer the voter's intent to a new ballot." We found these "remakes" in a box labeled "12-16-04." [ See EXHIBIT H ] There were 570 remakes in all – 291 for Bush, 200 for Kerry, and 79 undervotes, or 13.9%. If these were remakes of ballots cast on Election Day, then the remaining 49,744 ballots in the unofficial count included only 248 undervotes, or 0.50%. This disparity is too great to be attributable to random chance. It is more likely due to ballot tampering, possibly by "remaking" ballots on which a presidential choice had been erased or covered over.
32. On August 23, 2006, we hand-counted all the absentee ballots found in the Power House at Miami County. The absentee ballots for three precincts had been removed from the box in preparation for the recount of December 16, 2004, and there is no way now to distinguish them from the regular ballots for those precincts. For the other 79 precincts in Miami County, there were 2328 absentee ballots for Bush and Kerry. According to the official precinct canvass results provided by Mr. Quillen, there were 4656 absentee ballots counted for Bush and Kerry in these 79 precincts – exactly twice the number of absentee ballots in the box. Mr. Quillen has suggested that there must be another box of absentee ballots, but three researchers and Mr. Quillen himself have been unable to find a second box. It is highly improbable that a second box of absentee ballots exists with exactly the same number of ballots as the first box. It is far more likely that the absentee ballots were counted twice.
33. Examination of the voter signature books in Miami County reveals that, in the official results, the number of ballots cast is incorrect for all 82 precincts. In 44 precincts, 408 more ballots were reported than can be accounted for in the signature books. The discrepancy is ten or more in 12 precincts (61 in Troy 3-G, 38 in Concord South, 36 in Concord South West, 35 in Concord North, 26 in Spring Creek, 25 in Tipp City G, 17 in Piqua 2-A, 13 in Troy 3-F, 11 in Troy 1-E, 11 in Staunton North, 10 in Troy 4-C, and 10 in Pleasant Hill). In 33 precincts, 503 fewer ballots were reported than the number of signatures in the books. The discrepancy is 10 or more in 7 precincts (258 in Concord South East, 58 in Troy 4-F, 48 in Piqua 1-B/D, 22 in West Milton B, 15 in Laura West, 11 in Piqua 3-A, and 10 in Bethel South East). In 5 precincts the numbers match up, but only because the discrepancy in the absentee ballots is equal to the discrepancy in the regular and provisional ballots. On average, the discrepancy is 11.11 votes per precinct (911 votes in 82 precincts).
34. I asked Mr. Quillen for the name of the ES&S programmer for Miami County. He remembered only that his name is Sam, and that he lives in West Virginia. Upon information and belief, that would be Sam Hogsett, now or formerly of Crown City, Ohio, who also programmed the ES&S tabulators in Delaware County, Ohio for the 2004 general election and recount, and who has published letters to the editor in a West Virginia newspaper at www.herald-dispatch.com
35. In Warren County, according to newspaper reports, a "homeland security" alert prompted a lockdown of the county administrative building on Election Night and prohibited independent observers from watching the vote count. Our first records collection expedition, in which I did not participate, yielded photographs of poll books, voter signature books, and the front side of the punch card ballots for 12 precincts. However, our photographers were never shown the back sides of the ballots. As the front sides of the ballots are not identified by precinct, this led me to wonder if all the ballots in each of the stacks were actually cast in the specified precincts. I counted the ballots for 4 of these precincts and found that Bush got 57.7% (486 of 842) Ellen Connally voters, and 50.2% (472 of 940) gay marriage supporters. As the hand counts closely matched the official results, I began to wonder if punch cards had been shifted from precinct to precinct in order to manipulate the presidential vote count. In Ohio, the order in which the candidates’ names appear on the ballot (or, in the case of punch cards, on the voting machines) rotates from precinct to precinct. For example, a ballot punched for Kerry in Precinct Number 149 would be counted as a vote for Bush if it were transferred to the stack of cards for Precinct Number 148.
36. In our second records collection expedition to Warren County we photographed poll books, voter signature books, and the front and back sides of the punch card ballots for 9 precincts. As in Delaware County, on some of the punch cards the precinct was identified not in print, but by handwriting. Unlike Delaware County, however, all such cards that I observed in Warren County were identified in the same handwriting and with the same writing implement. This is not indicative of genuine absentee ballots mailed out to voters by different clerks on different days, using whatever writing implement was handy. It is noteworthy that none of the Board of Elections employees present would state that these were, in fact, absentee ballots.
37. If all the ballots in Warren County are in fact legitimate, then the Board of Elections should be able to account for all the unused ballots. The sum total of regular ballots cast, soiled and defaced ballots, valid provisional ballots, rejected provisional ballots, and unused ballots, should equal the number of ballots ordered from and delivered by Dayton Legal Blank. When I asked to see the unused ballots for these precincts, a Board of Elections employee stated: "I don’t know if we still have those. We’re not required to keep them." When I asked to see the pink header cards, which contain punch codes unique to each precinct so that the tabulator knows which precinct is being counted, I was told that they do not have them. In Butler, Cuyahoga, and Summit counties, by contrast, the header cards were plainly displayed on the tops of the stacks of cards for each individual precinct, and the punch code was different on each header card.
38. In Cuyahoga County we were granted unfettered access to the poll books and the voter signature books, all of which, according to Betty Edwards of the Candidate and Voter Services Department at the Cuyahoga County Board of Elections, are being scanned in order to preserve them for the record. The Department conducted a countywide audit and found provisional ballots unaccounted for in many of
Cuyahoga County’s 1436 precincts. The number of "good" provisonals plus "bad" provisionals, that is, those ruled valid plus those ruled invalid, should equal the total number of provisional ballots issued. They do not. I compared the number of "good" plus "bad" provisional ballots in the audit records with the number of provisional voters in the signature books and found that there are, for example, 23 provisional ballots unaccounted for in Cleveland 5F, and 15 provisional ballots unaccounted for in Cleveland 9P.
39. Included with the voter signature books are "Pink Memo Sheets" on which presiding judges and poll workers write their observations. The "Pink Memo Sheet" for Cleveland 8G, at Cory United Methodist Church, contains the following statements in the handwriting of the Presiding Judge: [ See EXHIBIT I ]
11:00 INSPECTOR OVERRIDE. VOTE AT ANY VACANT BOOTH.
4:00 INSPECTOR CAME BACK. INSTRUCTED VOTER TO VOTE AT PRECINCT.
There were three precincts at this polling station, all with different ballot rotations. Because Ohio law requires that the order in which the candidates’ names appear on the voting machines must rotate from precinct to precinct, placing a punch card in the wrong machine will result in votes being shifted to candidates not of the voter’s choosing. For five hours at Cory United Methodist Church, voters were allowed to vote at any vacant booth, regardless of precinct. In Precinct 8I, 29 Kerry votes shifted to Nader, and 28 Kerry votes shifted to Peroutka, because Kerry voters went to machines intended for Precincts 8H and 8G, respectively. In Precinct 8G, 51 Kerry votes shifted to Badnarik because voters went to machines intended for Precinct 8I.
40. On June 8, 2006, I received an e-mail from the Board of Elections stating that Gary F. Barna, Executive Assistant to Board of Elections Director Michael Vu, had boxed up all the 2004 punch card ballots and was not sure where he put them (there were 687,255 ballots cast in Cuyahoga County). At 2:00 P.M. on Friday, June 16, 2006, at which time my records request had been pending for 10 weeks and 4 days, we showed up at the Cuyahoga County Board of Elections with two video cameras rolling. Mr. Barna let us into the storage room in the basement, where our people, not Board of Elections personnel, found the ballots, not in any of the locked cages, but in a more freely accessible location within the storage room. (On a subsequent visit, the door to the storage room was found to be unlocked). The ballots were stacked up on pallets, shrink wrapped all together, against the wall, hidden behind other elections materials. We tore off the shrink wrap, rearranged the boxes with the labels facing outward, set aside the boxes that contain the ballots for the precincts I had requested, and the boxes that contain the absentee ballots, none of which had ever been tallied at the precinct level, having only been tallied by Congressional, State Senate, and State Assembly district. We photographed the boxes in place, as we left them.
41. I returned to view and photograph the ballots on July 17-20 and July 31-August 1, 2006, being especially interested in regular ballots for which the tabulators recorded no vote for president. Among the 782 such ballots observed, I found only 136 genuine "undervotes," on which there were punches for other offices or issues, but no punch for president. I found 56 absolutely blank ballots with no vote for anything at all, some of which must have been substitutions for ballots cast by actual voters, as poll workers counted fewer voted ballots than voter signatures in the poll books, and Cuyahoga County auditors found too few unused ballots remaining. I found 177 ballots containing punches for "Candidate Disqualified," the column created when Ralph Nader was removed from the ballot. These votes were intended for other candidates, but were shifted to the Nader column in multiple-precinct polling stations when voters were allowed to place their punch cards in machines intended for precincts other than their own. Because the tabulators were programmed to count the Nader column as zero, the only way to prove what happened was to examine the actual ballots. In one precinct alone, Cleveland 6M, I found 81 ballots on which a punch intended for Kerry had been shifted to "Candidate Disqualified." [ See EXHIBIT J and EXHIBIT K ] In Cleveland 6L, at the same polling station, punches intended for Kerry were shifted to Bush, who received 82 (20.1%) of the regular ballots, but only 2 (4.8%) of the absentee ballots.
42. I observed 362 ballots in Cuyahoga County which contained two or more punches for president, thus creating an "overvote." This is consistent with three complaints posted on www.voteprotect.org in which voters in Cuyahoga County said that they were given punch cards with the hole for a presidential candidate already punched out of them. No more than 12 of these 362 ballots observed were pre-punched for Bush only. Most of the extra punches were for third-party candidates. A punch for any two candidates for the same office will spoil the ballot. Analysis of 213 ballots with multiple punches for president in 13 Cleveland precincts indicates that, at a minimum, Kerry lost 172 votes, and Bush lost 7 votes.
43. Pre-punching of ballots in Cuyahoga County is consistent with an established pattern of voter disenfranchisement in heavily Democratic inner-city precincts. On behalf of Rolling Stone magazine, and with the assistance of Troy Seman, I found that 168,169 voters, or 19.44% of the electorate, were purged from the voter rolls in Cuyahoga County on July 6, 2001 and January 5, 2002. In Cleveland alone, 63,721 voters were purged, or 24.93% of the electorate. Kerry won Cleveland with 83.36% of the vote. His margin of victory was 113,145 votes. For every six persons unable to vote for having been purged from the rolls, four votes were shaved from Kerry’s margin of victory. The percent of voters purged should correlate inversely with the percent turnout in preceding elections. The best indicator is the presidential election, because it always draws the highest turnout. Our statistical analysis indicates that the purge ratio in Cleveland was unrelated to the percent turnout in 2000, but was strongly related to the percent of the vote won by Al Gore in the 2000 election. [ See EXHIBIT L ]
44. Similarly, according to Robert F. Kennedy, Jr. in Rolling Stone, the Republican Party sought to invalidate the registrations of 35,427 minority voters in Ohio, targeted by zip code. My own statistical analysis of the list of 930 challenged voters in Lucas County substantiates this statement. 370 (39.8%) were in 4 wards in Toledo with only 7.8% of the voters, where Kerry got 86% of the vote.
45. Similarly, in Hamilton County, 3179 of 3342 (95.12%) of rejected provisional ballots (excluding those for which the location was not filled out) were from the City of Cincinnati, with only 38.55% of the voters, where Kerry got 66.31% of the vote. Only 163 (4.88%) of the rejected provisional ballots were from the suburbs with 61.45% of the voters, where Bush got 64.24% of the vote.
46. Analysis of precinct results in urban areas across the State of Ohio indicates that the uncounted ballots (for which the tabulators recorded no vote for president) come disproportionately from precincts that voted overwhelmingly for Kerry. Precinct analysis also shows that among the precincts with the highest percentages of uncounted ballots, most (for example, 29 of 30 in Hamilton County, 13 of 18 in Montgomery County, 18 of 27 in Stark County, and 16 of 19 in Trumbull County) were of the "second rotation," where Bush was in the first position and Kerry was in the second position on the punch cards, and the third party candidates, Peroutka and Badnarik, were in the fourth and fifth position, respectively. This would have allowed the pre-punching of ballots to have escaped detection by the voters, as they would have been unlikely to scroll down the presidential column with their eyes after having punched the first position for Bush or the second position for Kerry. The candidates appeared in this order in only one-fifth of the precincts. The pattern of uncounted ballots coming predominantly from "second rotation" precincts cannot have occurred by random chance. It is more likely due to ballot tampering. This discovery prompted me to submit additional public records requests to Hamilton, Montgomery, Stark, Summit, and Trumbull counties.
47. In Hamilton County, 29 of the 30 precincts with the highest percentage of ballots not counted as votes for president (1,007 of 11,070 ballots cast, or 9.10%) had the second rotation. Of these, 28 precincts were won overwhelmingly by John Kerry, by a margin of 8340 to 844, nearly 10 to 1. Among the regular ballots in these 28 precincts, there were 873 not counted as votes for president. Of these, 677 (77.5%) were punched for Kerry, not for Bush, and were also punched for third-party candidates. In Cincinnati 24O, where there was one undervote, all 23 overvotes were double-punched for Kerry and either Badnarik or Peroutka, who received between them only 20 votes in all 28 of these precincts combined. [ See EXHIBIT M ] It is not credible that 677 of 697 Badnarik and Peroutka supporters also voted for Kerry, nor is it credible that voters are more prone to error in every fifth precinct, those with the second ballot rotation.
48. On Monday, August 21, I requested the unused ballots for the precincts we had examined in Hamilton County, in order to see if some of these ballots had been punched in advance for third-party candidates. On Tuesday morning, August 22,
John Williams, Director, Hamilton County Board of Elections, told me that they do not have the unused ballots. He seemed surprised to find them missing. I asked Mr. Williams about Directive 2004-43 from J. Kenneth Blackwell, which states: "All used and unused ballots must be retained for at least 22 months." Mr. Williams confirmed that Directive 2004-43 has not been superseded, and is still in effect. That night I sent the following e-mail:
49. "This is to confirm your advice to me that despite Directive 2004-43, dated October 25, 2004, from Ohio Secretary of State J. Kenneth Blackwell, you do not have and therefore will not produce the unused ballots from the November 2, 2004 election. If this is incorrect, or if you have subsequently found these materials, please advise me immediately so that I can review them in accordance with my original public records request under Ohio Revised Code 149.43."
50. On Wednesday, August 23, I received an e-mail response from John Williams, Director, stating: "That is correct." I also received on August 23 an e-mail response from Pamela M. Swafford, Deputy Director, stating: "We do not have the unused ballots from 2004."
51. The absentee ballots in Hamilton County have a three-punch array at the bottom of the punch card that is different from the triangular array at the bottom of the regular ballots. [ See EXHIBIT N ] Dennis Predmore, who has worked at the Hamilton County Board of Elections for 31 years, called it a "security code." This allows the tabulators to differentiate between regular ballots and absentee ballots, and would make it more difficult to count absentee ballots twice. Mr. Predmore told me that he personally instructed Dayton Legal Blank to punch this "security code" into the absentee ballots. This only raises the question as to why no other county whose ballots I have examined, whether punch cards or optical scan ballots, has designed a similar security code.
52. In Summit County, 14 precincts had 35 or more ballots not counted as votes for president (653 of 8,633 ballots cast, or 7.56%). All were in Akron. All were won overwhelmingly by John Kerry, by a margin of 7228 to 723, or 10 to 1. Of these 14 precincts, 13 were in Wards 3, 4, or 5. Six of the 14 precincts had the second rotation, and the pattern was exactly the same as in Hamilton County. Among the regular ballots in these six precincts, there were 276 ballots not counted as votes for president. Of these, 214 (77.5%) were punched for Kerry, not for Bush, and were also punched for third-party candidates. [ See EXHIBIT O ] Seven of the 14 precincts had the fifth rotation, where Bush was in the third position and Kerry was in the fourth. In one of these precincts, Akron 7J, there were 79 uncounted ballots, 67 of which had punches intended for Kerry that had been shifted to Nader. In the other six precincts with the fifth rotation there were 227 regular ballots, 114 (50.2%) of which were punched for Kerry, not for Bush, but were also punched for third-party candidates. Also in these six precincts, 24 votes were shifted to Nader. Another precinct, Akron 3L, had the first rotation, with Bush in the second position, and Kerry in the third position. There were 30 uncounted regular ballots, 15 of
which were backwards, upside down, misaligned, or completely blank; another 8 were punched for Kerry, not for Bush, and for third-party candidates.
53. In Trumbull County, there were 12 precincts in Warren City where more than 5% of the ballots cast were not counted as votes for president (272 of 4155 ballots cast, or 6.55%). All were won overwhelmingly by John Kerry, by a margin of 3310 to 562, nearly 6 to 1. Nine of these 12 precincts had the second rotation. Among the regular ballots in these nine precincts, there were 174 ballots not counted as votes for president. Of these, 125 (71.8%) were punched for Kerry, not for Bush, and were also punched for third-party candidates. [ See EXHIBIT P ] In the three precincts with other rotations, there were 48 uncounted regular ballots, of which 19 were double-punched for Kerry, not for Bush, and for third-party candidates, and another 5 were punched for Nader only.
54. In Stark County, there were 11 precincts, 9 of them in Canton City, where more than 5% of the ballots cast were not counted as votes for president (319 of 4741 ballots cast, or 6.73%). All were won overwhelmingly by John Kerry, by a margin of 3620 to 771, or 4.7 to 1. Seven of these 11 precincts had the second rotation. Among the regular ballots in these seven precincts, there were 168 ballots not counted as votes for president. Of these, 92 (54.8%) were punched for Kerry, not for Bush, and were also punched for third-party candidates. [ See EXHIBIT Q ] In the four precincts with other rotations, there were 98 uncounted regular ballots, of which 47 were double-punched for Kerry, not for Bush, and for third-party candidates.
55. In Montgomery County there were 29 precincts, 26 of them in Dayton, where 4.5% or more of the ballots cast were not counted as votes for president (600 of 10,403, or 5.77%). Kerry won all 29 precincts, 27 of them overwhelmingly, by a margin of 8412 to 648, or 13 to 1. Sixteen of these 29 precincts had 20 or more uncounted ballots, and all 16 had the second rotation. I photographed the uncounted regular ballots from 23 precincts altogether, 20 of them in Dayton, all of them with the second rotation. In these 23 precincts, there were 505 ballots not counted as votes for president. Of these, 276 (54.7%) were punched for Kerry, not for Bush, and were also punched for third-party candidates. [ See EXHIBIT R ] There were also 54 "dimpled chads," 45 of them for Kerry, and 2 "hanging chads" for Kerry. This is by far the highest incidence of any county whose records I have examined. In Cuyahoga County, by comparison, I found only 8 "dimpled chads" among 782 uncounted regular ballots. A dimpled chad results when the punching implement, or "stylus," fails to make a clean punch and leaves an indentation on the punch card. The "dimpled chads" in Montgomery County were found almost invariably on ballots with clean punches in the other columns, a clear indication that on some voting machines it was more difficult to punch a hole for president than for any other contest on the ballot.
56. Altogether I have photographed 3260 punch card ballots, in heavily Democratic urban areas in six counties, on which no vote was recorded for president. Of these,
1934 (59.33%) were punched for Kerry and one or more third-party candidates. Another 273 (8.37%) were punched for Nader only, nearly all of them due to voters using the wrong voting machines in multiple-precinct polling places. This is enough to establish a pattern indicating a deliberate scheme to disenfranchise voters in heavily Democratic urban areas by causing them to cast overvotes and spoil their own ballots, or to vote for candidates not of their choosing. I have examined only a partial data set, due to time constraints. According to official results there were 38,917 uncounted ballots in these six counties (13,478 in Cuyahoga, 9,038 in Hamilton, 5,062 in Montgomery, 3,322 in Stark, 5,415 in Summit, 2,602 in Trumbull). I have examined only 8.38% of these ballots, one out of twelve. That is why the ballots must be preserved. We need to learn the full extent of voter disenfranchisement.
57. If these double-punched ballots were pre-punched for third-party candidates in heavily Democratic precincts, then the unused ballots should have told the story. We should have been able to examine them for tell-tale punches. Unfortunately, Boards of Elections personnel in Summit, Trumbull, Stark, and Montgomery counties have stated to me that their unused ballots have been destroyed. I sent to the Director and Deputy Director of all four of these Boards of Elections the same e-mail message to which John Williams and Pamela M. Swafford of Hamilton County promptly responded. As of this writing, I have not heard back from any of them.
58. It is true that punch card ballots are no longer in use. However, the voting system now in place in Ohio has its own vulnerabilities that need to be addressed, specifically in regard to double-counting of absentee ballots.
59. Miami County is not the only county in Ohio where indications exist of double-counting of absentee ballots in the November 2, 2004 election. In Sandusky County, according to Barb Tuckerman, Director, Sandusky County Board of Elections, about 2,600 ballots from nine precincts were counted twice when they were mistakenly placed alongside a pile of uncounted ballots. The numbers have since been rescinded. Ms. Tuckerman said that the problem was discovered when one precinct in Clyde showed a voter turnout of 131%. This explanation does not withstand scrutiny, because Clyde has only six precincts, and the turnout in all of them was higher than 65.5%, ranging from 69.43% to 77.59%. Upon my request, the Sandusky County Board of Elections has confirmed, by fax, that there were 2,604 absentee ballots.
60. The Diebold tabulators currently utilized in at least 41 Ohio counties are programmed to report two sets of numbers for ballots cast. One, called "times counted," appears to equal the number of ballots cast. The other, called "cards cast," is equal to "times counted" plus the number of absentee ballots, which thus are counted twice. On the official website of Ohio Secretary of State J. Kenneth Blackwell, either number is posted, on an ad hoc basis, as "total voting." On November 8, 2005, absentee ballots in at least 12 counties were counted twice and
posted by Blackwell as "total voting," thus creating the appearance of large numbers of undervotes. On May 2, 2006, absentee ballots for Pike County were counted twice and posted by Blackwell as "total voting," even though the Pike County Board of Elections reported no such number. This ongoing system of discretionary or arbitrary use of alternative sets of numbers for absentee ballots cast presents an opportunity for double-counting of absentee ballots in selected counties for partisan advantage. Absentee ballots can be counted all at once, county wide, because the ballots themselves are identified by precinct according to a bar code recognized by the optical scanners. Absentee ballots can be counted twice, county wide, without having the number of ballots cast appear to exceed the number of registered voters in any precinct and, if the number for "cards cast" is selectively reported, without the number of votes counted appearing to exceed the number of ballots cast in any precinct. The Court should order implementation of a security bar code system by which optical scanners would distinguish between regular and absentee ballots, as the Hamilton County Board of Elections has done and has had in place for its punch card ballots over the past 30 years, and a uniform system of accounting for the number of absentee ballots cast in Ohio elections.
61. Having reviewed a substantial amount of forensic evidence, it is my conclusion that there is direct evidence of ballot tampering in each of the eleven counties whose public records I have examined, and that there is a compelling need to protect the evidence from destruction, presently scheduled to take place on or shortly after September 2, 2006. The Court should order all Boards of Elections to continue to preserve and protect all ballots, poll books, voter signature books, and associated records from the November 2, 2004 election until a suitable repository is found for their permanent preservation.
I declare under penalty of perjury that the foregoing is true and correct.
Executed this __first__ day of September, 2006.
Richard Hayes Phillips