Direct testimony: Presented to Election Assessment Hearing, Houston, Texas, July 29, 2005

I have investigated the Ohio election results, precinct by precinct, and have found three categories of problems: voter suppression, ballots cast but not counted, and alteration of the vote count.

In the City of Columbus, discriminatory allocation of voting machines led directly to lower turnout in Democratic precincts. Urban Democratic precincts had too few voting machines and long lines; suburban Republican precincts had enough voting machines and short lines; 122 voting machines were not provided to any polling station anywhere. As a result, voter turnout was 60% in Bush precincts, and 50% in Kerry precincts. This wrongly reduced Kerry’s margin of victory in Franklin County by about 17,000 votes.

In Lucas County, other means of voter suppression led directly to lower voter turnout in Democratic precincts. The 88 precincts with the lowest turnout were all in Toledo; all were won by John Kerry; and complaints were filed in 31 of these precincts. Among the complaints were: long-time residents removed from the voting rolls, broken voting machines, polling stations running out of ballots and turning people away, voters sent back and forth between polling places, and long lines not designated by precinct so that voters waited in the wrong line. One-third of provisional ballots were not counted, often because people voted at the wrong table in the right polling place.

Statewide, there were 35,000 provisional ballots and over 92,000 regular ballots that were not counted as votes for president. These uncounted ballots, most of them punch cards, are highly concentrated in precincts that voted overwhelmingly for John Kerry, by margins of 12 to 1 in Cleveland, 7 to 1 in Dayton, 5 to 1 in Cincinnati, 4.5 to 1 in Akron, 3 to 1 in Lorain County, 2.7 to 1 in Stark County, and 2.3 to 1 in Trumbull County. This cries out for an examination of the uncounted ballots and the machines that failed to count them.

In the State of Ohio, the order in which the candidates’ names appear on the ballot rotates from precinct to precinct. The sequence remains the same, but the starting point differs. If voters stood in the wrong line, and placed their punch cards in voting machines not intended for their precinct, their votes were shifted to candidates not of their choosing. In Cleveland, where Kerry had overwhelming support, there are many obvious examples where votes intended for Kerry were shifted to Bush, or to third-party candidates, or into the column left void when Ralph Nader was removed from the ballot. This artificially reduced voter turnout, and wrongly reduced Kerry’s margin of victory in Cuyahoga County by about 22,000 votes. Also there are complaints and sworn affidavits from voters in Cuyahoga County who were given punch cards with the hole for George W. Bush already punched out of them.

In Miami County, after 100% of the precincts had reported, more than 18,000 votes were added to the totals. The vote percentages for Bush and Kerry remained the same, and the final margin for Bush was 16,000 votes exactly, almost as if the tabulators were programmed to turn out that way. This resulted in voter turnout as high as 98.55% in one precinct, where a door-to-door canvass of voters identified more than enough non-voters to prove that the certified results are fraudulent.

In Mahoning County, the Board of Elections reported that 20 to 30 touch screen machines had to be recalibrated because votes were being counted for the wrong candidates. Voters had to scroll through as many as five times before their choice for president was registered. In some precincts, machines failed to record votes for Kerry and defaulted to no choice at all. In other precincts, touch screens were programmed to default to candidate Bush unless the voter successfully overrode the default choice. These voters either saw the touch screen hop from Kerry to Bush, or never saw a presidential choice appear on the screen, and yet were recorded as having voted for president.

Three counties in southwestern Ohio – Butler, Clermont and Warren – provided Bush with a combined plurality greater than his statewide margin of victory. These results, when examined at the precinct level, are almost impossible to explain. In Butler County, there were 12 precincts where, compared to 2000, voter registration increased by 34% to 178%, and entire townships where Kerry received fewer votes than Gore, despite large increases in voter turnout. In Clermont County there were 24 such precincts where Kerry received fewer votes than Gore. In Warren County there were six entire townships where voter registration increased by 28% to 79%. In all three counties, Kerry received fewer votes than Ellen Connally, a little-known, underfunded African-American municipal judge from Cleveland, running for Chief Justice. There must have been at least 13,500 voters who supported both Connally and Bush, or else the certified results are fraudulent. In these three counties, and in Delaware County as well, Bush received more votes than Issue One, the constitutional amendment prohibiting gay marriage. There must have been at least 10,500 supporters of gay marriage who voted for Bush, or else the certified results are fraudulent. A comparison of these counties with the voting patterns statewide indicates that as many as 50,000 votes may have been shifted from Kerry to Bush, thus affecting the margin by 100,000 votes.

These accounts of voter suppression, ballots cast but not counted, and alteration of the vote count, were sufficient to reverse the outcome of the presidential election. It is my professional opinion, having exhaustively examined the available evidence, that the 2004 presidential election was stolen.