To start, voting analyses of selected Florida and Ohio precincts conducted by the University of Pennsylvania's Steven Freeman and independent investigator Faun Otter have revealed surprisingly high percentages for Bush. Those skeptical about the results further suggest spoiled ballots and provisional votes, which may have a disproportionate impact on the results in the areas with high concentrations of minority voters, could have made the difference.
The earliest exit poll data released on Nov. 2 indicated Kerry -- who had run narrowly behind Bush but within the margin of error for most of the race -- was rolling to victory and carrying many of the battleground states, including Florida and Ohio, by higher than expected margins. These same polls also suggested the Republicans were ahead in most of the tight U.S. Senate races.
By the end of the night, however, the predictions in the presidential exit were wrong while the Senate projections were largely correct.
Exit polling by Edison Media Research and Mitofsky International, which created the National Election Poll for ABC, AP, CBS, CNN, Fox, and NBC, had shown Kerry leading by 3 percentage points in Florida and by 4 points in Ohio. Kerry lost Florida by 5.2 percent, with Bush running ahead of his 2000 performance in 58 of the state's 67 counties. In Ohio, the margin was 2.5 percent.
Florida's 8.2-percent spread -- between the early exits and the results -- is more than double the standard error rate. In Ohio, the spread is 6.5 percent.
In Baker County, Fla. located near the city of Jacksonville and just across the border from Georgia, there are 12,887 registered voters: 69.3 percent are Democrats, 24.3 percent are Republicans. Yet 2,180 of county residents voted for Kerry while 7,738 voted for Bush -- the opposite of what some election critics say was the typically pattern elsewhere in the United States.
In Florida's Dixie County, located on the Gulf Coast between Tallahassee and Tampa, 77.5 percent of the 4,988 registered voters are Democrats, 15 percent are Republicans. On Election Day, Bush carried the county with 4,433 votes vs. 1,959 for Kerry.
Nationally, few outlets have pursued the story of what happened in Baker and Dixie, why and whether it actually indicates a problem with the counting of the ballots. Most of the coverage of the alleged irregularities has focused on why the exit polls were so far off. Skeptics dismiss them as flawed or somehow favoring Kerry and say that, though they may have influenced the narrative of election coverage, they couldn't affect the outcome.
To explain the difference, some unconvincing theories have been floated including the one offered by the architects of the sampling system used for exit polling. They say Kerry voters were simply more willing to answer the questions. It's called the "chattiness thesis" and it sounds like a weak excuse -- but so was the pollsters' earlier claims that the numbers were right, the media just read them wrong. In an article for Tom Paine.com, a liberal Internet publication, Greg Palast, an author and frequent critic of the 2000 election returns in Florida, goes farther.
"Although the exit polls show that most voters in Ohio punched cards for Kerry-Edwards, thousands of these votes were simply not recorded," he writes. Palast says he thinks the election was decided by "spoilage," the small part of the vote that is voided and thrown away.
In Ohio, as in Florida four years ago, a large number of spoiled votes were cast on punch cards, 54 percent of which were cast by black voters, according to statisticians investigating the issue for Verified Voting, a group formed by a Stanford University professor to assess electronic voting. Verified Voting has collected 31,000 reports of alleged election abnormalities.
Other factors also could have affected the vote count, including last-minute legal challenges filed in several states, both by Democrats trying to block Ralph Nader from appearing on state ballots and Republicans concerned about lax registration rules. Long lines at precincts in the evening and the large number of total provisional ballots cast across the United States also may have influenced the outcome somewhat.
Taken together, such factors could significantly change the vote in some areas, bringing the count more into line with the exit poll results.
Were the election results manipulated in some way? At the moment, the question invokes the same kind of polarizations generated by the election choice itself; a much more thorough analysis is needed -- and will not be quick in the offing -- before the Internet chatter can taken seriously, even though some will always believe it did in fact occur.
Even if the thesis can eventually be demonstrated to be accurate, that some form of manipulation did occur, the technology involved is so complex that those responsible will likely escape the consequences.
(Greg Guma edits the Vermont Guardian, a statewide weekly, and Toward Freedom. He can be reached at firstname.lastname@example.org )
(United Press International's "Outside View" commentaries are written by outside contributors who specialize in a variety of issues. The views expressed do not necessarily reflect those of United Press International. In the interests of creating an open forum, original submissions are invited.)