Ballot reviews agree on this: It was close
© St. Petersburg Times,
George W. Bush will make his third trip to Florida as president when he visits the Everglades and Tampa on Monday, yet it remains unclear whether most Floridians intended for him to move into the White House.
Nearly seven months since the historic election, more than 15 newspaper-led reviews of ballots in all or parts of the state reflect the fragility of Bush's victory over Al Gore.
The Supreme Court stopped the official recount in December, handing Bush a razor-thin victory in Florida and the presidency. But the media recounts since then confirm that the Republican's 537-vote advantage out of nearly 6-million votes cast was, at best, a rough estimate.
Virtually every review of the ballots has produced different numbers.
In Miami-Dade County, a Palm Beach Post recount found a net gain of six votes for Bush. A Miami Herald recount in the same county found a net gain of 49 votes for Gore.
In many counties, including Pasco and Pinellas, the number of punch card ballots rejected by computers change when they are rerun through the machines.
And in the first statewide hand recount by the media, the Herald and USA Today concluded this spring that Bush or Gore could have won.
Bush won by up to 1,665 votes or Gore won by three, depending upon the standard used to count undervotes -- those ballots where a machine did not record a vote for president.
The newspapers came up with different numbers after looking at overvotes, which are ballots that were initially discounted by computers because more than one candidate was marked. The overvotes were examined to see whether a voter's intent could be determined.
Counting undervotes and overvotes where voter intent could be determined, the Herald and USA Today found Bush won by 407 votes when counting chad hanging by two corners on punch card ballots. That two-corner standard is one many counties were prepared to use during the official recount.
The Herald/USA Today review found Bush won by 152 votes when only clean punches on punch card ballots were counted. That is the standard the Bush camp sought but that only one state, Indiana, requires.
But Gore won by 332 votes when any dimple was counted as a vote, the most liberal standard that some counties were prepared to embrace. The Democrat won by 242 votes when dimples were counted as votes for president only when there were dimples in other races. That is the standard Palm Beach County initially embraced.
In any of those scenarios, the outcome was closer than the official result.
"There are people still walking around today who say, "I'm not sure if my vote counted or not,' " Florida Democratic Party Chairman Bob Poe said.
Countered Florida Republican Party Chairman Al Cardenas: "You're never going to have 100 percent accuracy. It doesn't happen in our 50 states, and it doesn't happen anywhere else in the world."
The ballot reviews and ongoing analysis of Florida's election problems have not undermined Bush's legitimacy in national polls, analysts said.
"I think people are interested in what these recount studies show, but I don't think they have much bearing on the legitimacy of Bush's presidency," said Andrew Kohut, director of the Pew Research Center for the People and the Press. "He is president. People accept him for better or worse."
Instead, the fallout from the closest presidential election in 124 years continues on other fronts.
The Florida Legislature approved sweeping reforms, including a ban on punch card ballot machines like those used in Pinellas, Pasco and Hillsborough counties. When the state's county election supervisors meet at Saddlebrook Resort in Pasco County this week, they will be greeted by salesmen eagerly hawking voting machines to replace the punch card systems.
State elections officials also will hold a workshop to discuss rules that would create the first state definition of voter intent and provide for uniform ballots that would be used throughout Florida.
In about a month, another media-sponsored recount will be published that is expected to be the last and most definitive effort to examine what happened when voters went to the polls Nov. 7.
A group of leading news organizations, including the St. Petersburg Times, has overseen a months-long effort to review about 180,000 ballots. The ballots include undervotes and overvotes in all 67 counties.
Besides the St. Petersburg Times, the group includes the New York Times, the Wall Street Journal, the Washington Post, the Associated Press, the Palm Beach Post, CNN-Time Magazine and Tribune Publishing.
Tribune Publishing, based in Chicago, owns the Orlando Sentinel, the South Florida Sun-Sentinel, the Chicago Tribune and the Los Angeles Times, among others.
The Washington Post Co. owns the Washington Post and Newsweek. The New York Times owns the Boston Globe and several newspapers in Florida, including the Sarasota Herald-Tribune, Ocala Star-Banner, the Ledger in Lakeland and the Gainesville Sun.
The newspaper consortium hired the National Opinion Research Center, a non-profit survey research firm affiliated with the University of Chicago, to review the ballots. NORC was charged with examining and recording the characteristics on each of the ballots, not with determining which marks should be counted as votes.
"This operation in many ways is like any survey or census," said Kirk Wolter, a senior vice president at NORC who is overseeing the Florida project. "It is collecting information."
The consortium's recount effort has been delayed by disputes with several counties over how to separate the undervotes and overvotes. The project is different from the Herald/USA Today review in several ways:
The consortium is reviewing considerably more ballots than the Herald/USA Today effort, which likely will produce different vote totals.
Unlike the Herald, the consortium insisted that counties resegregate their undervotes and overvotes by machine to try to mirror the totals the counties reported after the first statewide machine recount on Nov. 8. Virtually every county is expected to comply with the request.
The newspapers hope that the number of overvotes and undervotes NORC reviews will be within 1 percent of the totals reported by each county after the first mechanical recount.
Mark Seibel, the Herald's managing editor for news, said he expects the consortium's numbers to be different from the Herald/USA Today study. But he predicted the percentage of votes for Bush and for Gore will remain the same. He downplayed the significance of specific numbers, suggesting that the trend of the reviewed ballots -- for Bush or Gore -- is more important.
The Herald did not ask counties to resegregate ballots by computer because it wanted to duplicate what would have happened if the recount had not been stopped by the courts, Seibel said.
The Herald/USA Today review used different procedures than the consortium to review the ballots that are considered undervotes.
The Herald's recount numbers are based on reviews of overvotes by one accountant from BDO Seidman LLP. A Herald reporter also reviewed each ballot to check for variations, but any disagreements were automatically resolved in favor of the accountant.
The consortium relied on 20 teams of three coders each who were trained by NORC and whose work was monitored by a supervisor. Each of the three coders independently reviewed each ballot.
To review overvotes, the Herald/USA Today review relied on official computer tapes to electronically tally overvotes on 56,225 ballots in eight large counties instead of examining each ballot by hand. That number accounts for just more than half of the 111,261 overvotes the newspapers reviewed statewide.
Seibel said the Herald relied on computer records because more than 9 of 10 ballots classified as overvotes had no marks on them that could help anyone determine the voter's preference. He said reporters from the Herald and seven other newspapers looked at every ballot in every county that included a mark.
For the consortium, one NORC coder reviewed each overvote ballot.
"The proof will be in the results," Seibel said. "The consortium's methodology was more complex than ours. Whether it yields better results, I guess we'll have to wait and see the results."
Predictably, Democrats and Republicans offer different assessments of the Herald's statewide ballot review.
Poe said the combined results from the undervotes and overvotes indicate that more Floridians intended to vote for Gore than for Bush.
"A combination of things happened," he said. "The system failed and some voters made mistakes. The combination of those two was fatal for Gore."
But the recount ordered by the Florida Supreme Court focused on undervotes. Cardenas said the Herald's ballot review confirmed that Bush received more votes than Gore under the most commonly used standards to determine voter intent on punch card ballots.
"We surmise as the press did that there were imperfections in the process in Florida -- a process, incidentally, that had been created by a Democratic legislature and governor," he said. "But there were imperfections, and this governor and this legislature set out to correct them."
Aside from voter error and unreliable voting machines, other factors influenced the election's outcome.
For example, the Palm Beach Post and the Washington Post recently have recounted the botched effort by a private contractor hired by the state to compile a list of convicted felons. County election supervisors were supposed to use the list to remove felons from the voter rolls.
As a result of the errors, the Washington Post reported, at least 2,000 felons whose voting rights had been restored in other states were removed from Florida voting rolls. But those people presumably should have been allowed to vote, and black residents were disproportionately affected.
That situation cost Gore more votes than Bush, because exit polls on Election Day show the Democrat won nine of every 10 votes cast by blacks.
But the mistakes ran both ways.
More than 5,600 Floridians who voted Nov. 7 had names that were a match with those on the statewide list of suspected felons, the Palm Beach Post reported. Given the probable racial makeup of that group, those errors cost Bush more votes than Gore.
Other election problems were well-chronicled. Republicans and Democrats continue to debate how absentee ballots from overseas were counted. Some Florida Panhandle residents might have skipped voting because television networks projected Gore as the winner of the state before polls closed.
Cardenas said all these issues should be considered when debating the outcome of any media recount of the ballots.
"Add those factors," he said, "and it is beyond doubt we would have won using any standard."
© 2006 • All Rights Reserved • Tampa Bay Times
490 First Avenue South St. Petersburg, FL 33701 727-893-8111