Strange Bedfellows   

A behind-the-scenes look at how some of the nation's top news organizations, normally fierce competitors, joined forces to recount disputed Florida ballots--while the Miami Herald conducted its own tally.

THE MEETING ON DECEMBER 14, 2000, resembled a gathering of the big Mafia families, only the attendees were the dons of Washington journalism. The mighty bureau chiefs of the Wall Street Journal, Los Angeles Times, Chicago Tribune and CNN filed into the New York Times' Washington bureau along with consiglieres from the Associated Press, Time and the Washington Post.
They were there to call a truce. They hoped to work together on a massive reporting project inspecting Florida ballots tossed out in the wildest presidential election in a century. Together they could cut costs and speak more authoritatively than if they pursued solo efforts that might end with conflicting results.
But such a project would require checking egos at the door. It would mean ceding control to the group. It would demand suppression of the very instincts that make each a respected journalist. For the next four or five months they would simply have to try to trust one another.
"Normally we don't look for opportunities to get in bed with the competition," Los Angeles Times Washington Bureau Chief Doyle McManus wrote in a December memo to his bosses, "but in this case, it might solve some problems."
Ultimately, the news dons formed an unprecedented coalition of media outlets better known for furiously competing with each other. The New York Times Co., the Tribune Co. (which owns the Los Angeles Times, the Chicago Tribune and other major newspapers), the Washington Post Co., CNN, the Wall Street Journal and the AP--aka the Big Six--joined forces with the Palm Beach Post and the St. Petersburg Times. Though invited, the television networks declined to participate.
Working together--one outlet/one vote regardless of size--they would compile raw data on the 180,000 disputed ballots. Members would fashion their own stories from the untabulated database. "There is no way the Wall Street Journal is going to turn its news judgment over to the New York Times," says Alan Murray, the Journal's Washington bureau chief.
How such a joint ballot review would work was painstakingly resolved through conference calls, sometimes with 20 unabashed journalists talking. "I was sure there were going to be a lot of knock-down, drag-out fights," says McManus. But it didn't happen once each agreed to the equivalent of a non-compete contract.
Washington journalists for such elite organizations pride themselves on their independence. They are used to running their own shows. Each savors nothing more than breaking a story that makes the others miserable. Somehow they'd have to learn to cooperate.
"As with all major projects, there are a million fork-in-the-road decisions," says Dan Keating, a database editor for the Washington Post. Keating and the New York Times' Ford Fessenden served as the coalition's point people in Florida. "We have to cost-weigh the benefit of doing things this way or the other," Keating says. "All these choices are tough calls for one project coordinator. Instead, we are trying to make them in teleconference calls with eight news organizations and more people than that on the phone."
The last time news outlets did anything remotely like this on a national scale was in 1976, after Arizona Republic reporter Don Bolles was murdered while working on a story exposing land fraud and organized crime. Thirty-eight journalists from 28 newspapers and television stations came to Arizona to finish what Bolles had started and put together a series that ran in such papers as the Miami Herald, Newsday and the Kansas City Star. The Washington Post and the New York Times declined to participate.
But 25 years later, an unwieldy collection including the Times and Post embarked on an equally consequential, although less perilous, project--combing through 180,000 rejected ballots cast in Florida's 67 counties and reporting their findings. They vowed not to declare a "winner" in the closely fought battle between President Bush and his unsuccessful opponent, Al Gore.
It was a project fraught with obstacles. Chief among them was reconciling the conflicting desires of the consortium's powerhouse members. The weekday-only Wall Street Journal insisted the others forgo a Sunday publishing date. Some wanted to count only undervotes--ballots that hadn't been marked clearly; others wouldn't participate unless overvotes (ballots that designated two or more candidates) were included. Some thought reporters should scour ballots; others said having journalists carry out that task would hurt credibility. Some wanted raw data; others wanted numbers tallied in a useable form. Some worried that, although partners, they might be big-footed by the Post or the New York Times, which assigned two people to work full time on the project.
"All of us are aware of the irony or paradox," says John Broder, the Times' Washington editor, "of working together on one of the more important historical stories of our time. We have agreed to give up any competitive advantage."
After three weeks haggling, their concerns assuaged, each signed a contract in January. Collectively they represent some of the wealthiest corporations in America. Yet a regional daily, the Miami Herald, was going to scoop them.

REVIEWING AMBIGUOUS BALLOTS appealed to many news organizations, but to none more than the Miami Herald. Here was a huge national story playing out in its backyard. In 1999, the paper won a Pulitzer Prize for uncovering voter fraud in Miami that overturned a mayoral election. The Herald decided to inspect only the undervotes in the state's 67 counties--those 60,000 ballots discarded because each lacked a clear mark for president.
The effort began around 10 p.m. on December 12, after the U.S. Supreme Court officially ended the election. Herald Assistant Managing Editor Mark Seibel turned to reporter Lila Arzua and asked her to get names and fax numbers for every election supervisor. Arzua, 24, had joined the paper only three months before. By 3 a.m., she'd found data on about 60 supervisors. The next morning, the list was finished. By noon, Seibel had drafted a letter and faxed it to each county.
"Why are we doing it? The truth is, it is crazy for people to go back and forth about what these ballots might have resulted in," Seibel said in January, the month he was promoted to managing editor. "It's important to know if it's obvious we put the wrong guy in the White House or the right man."
Other papers also took advantage of Florida's publicrecord laws, considered the most open in the country. Since ballots are public documents, anyone can inspect them. The Herald quickly won support from Tony Ridder, CEO of parent Knight Ridder. But Ridder had one stipulation: The Herald must hire an independent accountant. If reporters reviewed ballots, the tally would not be considered credible.
"Tony Ridder got directly involved," says Seibel, "and that helped us to move more quickly." ("It's been great," he says. "We get a bill from the accountant for $140,000 and corporate pays it.")
From the start, every news organization was acutely aware of the need to establish uniform standards for observing ballots. Florida's counties use widely differing standards for manual recounts; the lack of uniformity was a major factor in the U.S. Supreme Court's decision to stop the official counting.
Not only did anyone doing a tally need strict criteria, they needed to get to every county. But Florida is a big state. Logistically, it looked like a nightmare to get to each county, obtain approval and arrange for machines to segregate votes. In Palm Beach County alone, election officials would have to segregate undervotes from 462,350 total county votes and hold each of the 9,150 ballots up for inspection.
The Herald chose to focus on undervotes only because to do so reflected a real-world scenario that would have played out if the U.S. Supreme Court had not intervened. Says Herald Executive Editor Martin Baron, "At no time did any court order a manual recount of overvotes."
Although it promised to be costly, Knight Ridder didn't balk. (The final cost will be about $500,000.) Soon after, Knight Ridder papers would be trimming their budgets to keep profit margins up during an economic slump.
Finding an auditor was the first priority. The "Big Five" accounting firms declined, fearing any result might make them appear partisan and hurt business. Meanwhile, Arzua began scheduling counties. Although ballots are public records, they can be physically handled only by election officials under Florida law. Arzua had to make appointments for county employees to hold up ballots one at a time.
Broward County scheduled the first ballot viewing December 18--just four working days after the election ended. The Herald wasn't ready. "The truth is, the Thursday before the 18th, we were publicly committed to doing a review and we didn't have an accounting firm," says Seibel. "We were in a bit of a panic."
Finally, the Herald found a company willing to take on the task: BDO Seidman LLP, one of the country's largest accounting and consulting firms. Over the weekend, Seibel trained an accountant and his supervisor on the method for reviewing ballots. Seibel had created a form. A reviewer would mark down the ballot number, whether it was dimpled or had a hanging chad, if the vote was for Bush or Gore, if there was no vote, if other races were dimpled or clean. The counter could also add comments. Then data would be entered into a computer.
Monday morning, election workers began holding up 6,600 punch cards at the Voting Equipment Center in Fort Lauderdale for reporters from eight news organizations, a representative of BDO Seidman, Republican Party officials, two local voters and a representative of Judicial Watch, a conservative watchdog group. Says Seibel, "It was just terribly chaotic."

C hoped to work with the Herald. The paper knew Florida, had already filed requests to examine the ballots and had hired an auditor. "It became very clear that the expense was too big for any one organization, and the odds that [parent company] Dow Jones was going to hand me a million dollars were nil," says Murray of the Wall Street Journal.
Some approached the Herald in mid-December. But a joint effort did not appeal to the Miami daily. "We were ready to go," says Baron. "We were fully capable of doing this by ourselves. So why do we need to be involved with a big, unwieldy committee?"
Seibel also didn't like the idea. "The idea of sitting down with eight big egos who would be a pain in the ass is not appealing," he says. But the sheer notion of collaboration is what chiefly bothered him. "It's hard enough to convince the American public that the media is not a conspiracy," Seibel says. "It's not like we are a bunch of poor companies. A half-million bucks wouldn't be much for any of them. So why do we have to do it together?"
Nor did Herald editors want to work with the competition. The Palm Beach Post and the Tribune Co., parent of South Florida's Sun-Sentinel, are consortium members. "Why give up an exclusive in one of the most competitive news markets in the country?" asks Baron.
New York Times Publisher Arthur Ochs Sulzberger Jr. and Washington Post Publisher Boisfeuillet Jones Jr. called Ridder in unsuccessful efforts to persuade him to join the consortium. On December 20, Times Managing Editor Bill Keller tried again. "One week into the project, Bill Keller called," says Baron. "Would we reconsider? Could they join us? Their particular interest was that we had an accounting firm and they didn't want it done by reporters. Neither did we. They were also interested in having one count and not having competition. He asked what conditions we required."
Baron replied by letter. "Basically we said, No. 1, we want to be in control of the project just like we have always been," Baron said in a March interview in his Miami office. "They thought we were so desperate for money we could compromise. The fact was, we wouldn't. They misjudged us."
Consortium members were outraged by the Herald's stance. "The Herald would only be part of our group if they could be first among equals," says Chicago Tribune Washington Bureau Chief James Warren.
To Baron, it was all about control. "Can you imagine if the Miami Herald said to the New York Times or Washington Post, if this happened in their area, 'We believe you should be partners of a consortium and a steering committee should run it and the Miami Herald must be on it,' " Baron says. "Now, they've spun this that the terrible Miami Herald has asked for outrageous conditions. We never asked for partners. We never initiated any phone calls."

SHORTLY BEFORE CHRISTMAS, talks between the consortium and the Herald broke down. Endless conference calls then ensued among consortium members. Occasionally, as many as 20 people participated. Sometimes no one knew who was talking. "That was weird," says Keating. "You are discussing this sophisticated project, and you don't know who is on the call."
Keating and Fessenden searched for an auditor and encountered the same resistance as the Herald had. "They all said no," Fessenden says. "Some explicitly said they could not afford the political exposure. They couldn't afford to piss off the Republicans. One I called and explained what we wanted to do. There was silence. 'Is this a joke?' he asked. 'You are really going to put the country through this?'... We even asked the New York Times' own auditors. They explained politely that was not the kind of thing they want to do."
By early January, "we were despairing two times a day," Fessenden recalls. "There was no Yellow Pages for this."
And then Kirk Wolter called. Wolter is head of the National Opinion Research Center (NORC), which is affiliated with the University of Chicago. The nonprofit, founded in 1941, conducts scientific survey research used to shape public policy. Wolter had heard about the project and wanted to participate. Keating and Fessenden were overjoyed. Says Keating, NORC is the "Cadillac in this field."
A contract was drawn up. No outsiders could attend meetings, which would be off the record. The Big Six would each pay 15 percent of the cost and the two Florida papers would cover 5 percent each. No data would be released until all counties were reviewed. They agreed on an embargo between data release and publishing dates, so each would have time to analyze results and do interviews.
But there were other issues. For example, should NORC ask coders about their political affiliations? "That ended up being quite passionate," McManus says. "There was a 'no' and 'yes' camp. The 'no' camp said we were better off not knowing. The 'yes' camp said we were in the business of knowing information."
The "no" camp lost. Coders could not have worked on or donated money to a political campaign in the last 10 years. And they had to pass an eye exam.
On January 9, an agreement was signed. The following day, USA Today announced plans to join the Herald's effort and share some of the costs. It would also examine the overvotes.

NOW THE SWORDS WERE drawn; each side claimed the other was behaving badly.
"Once the consortium realized that Miami was going to beat them on the undervotes, the whole philosophy behind the consortium became that they were going to do it right and do it complete," says USA Today Projects Editor Doug Pardue. "Unfortunately, that implied that Miami is not doing it right. I've seen some of the comments, and I think they do a disservice to our business."
Even before it started, the consortium conceded it would finish second. NORC didn't start reviewing ballots until February 5. By then, the Herald had looked at 57 counties. "I still believe our product will be superior because of a more careful and more thorough methodology," the Times' Broder says. "The Herald's stories will get tremendous bounce, I'm sure, for being first. But I think in the fullness of time the consortium's results will be seen as more authoritative."
The consortium members claim they're not in it for the scoop. "This is more of a public-service project," McManus says. "Otherwise we would not be willing to cede control. That's the difference between the Herald and us. Marty said this is a competitive news story. The rest of us said, 'Maybe it isn't.' "
But Seibel and Baron doubt the consortium's sincerity. On February 26, the Herald ran a story indicating that if Gore had gotten the four-county manual recount he'd requested in court, he would not have won. The story made headlines. And it also garnered much criticism by readers who claimed consortium members such as the AP and the New York Times intentionally ignored or downplayed the story.
The Post ran the story on A4. Dozens of readers complained the article should have gone on the front page. "These readers saw the Post decision as reflecting political bias and playing down the efforts of rival newspapers," wrote Post Ombudsman Michael Getler. Post editor Leonard Downie Jr. told Getler the paper "would not even consider" putting media recount results on its front page if the Post had not participated. (In April, the Post played a story on the Herald's final findings on page seven.)
But the methodology used for the recount is the true battleground. Three NORC-trained coders would study the undervotes, while the Herald utilized two--an accountant and a Herald reporter, deferring to the accountant when disagreements arose.
"We think there's going to be a huge quality difference between ours and the Herald's, even if the results might be the same," says Keating, who worked at the Miami paper for 11 years. "Ours is designed from the start to have statistical reliability. There are other methodological things we are doing that are taking more time."
Baron begs to differ. "Basically we are doing the same thing the others are," he says. "They are making this vaunted claim that three people are looking at everything and now they have just one person, like we do, looking at the overvotes." Indeed, the consortium decided in March that only one person needed to look at overvotes. It's far easier to see if an overvote was marked more than once, while an undervote is more subjective.
Baron and Seibel are taken aback that the consortium first championed a joint project, then turned around to criticize the Herald. Two counts are better than one, Baron argues.
"I think we saw with Voter News Service that having all media rely on one source of information is a route to potential disaster," Baron says (see "How They Blew It," January/February). "When did 'competition' in the news field become a bad word?"

IN JANUARY, IT APPEARED the Herald would be able to meet its goal of publishing its recount findings in March. By the end of the month, only two counties, Duval and Holmes, remained to be tabulated. Tiny Holmes' 139 disputed votes were scheduled for inspection on March 12. Duval, one of 26 Florida counties that use punch cards, initially agreed to machine-sort 291,000 ballots. Then it changed its mind and decided to charge the Herald $75,600. So on January 31, the Herald sued.
On February 27, a victory. Circuit Judge L. Page Haddock, of Jacksonville, ordered election workers to start manually sorting undervotes, to protect them from further damage.
Seibel negotiated a time for two reporters, a photographer and four accountants to start work: at 7 a.m. on Monday, March 5. He began to feel optimistic until his phone rang on the afternoon of March 2. There was a new wrinkle: The Herald planned to examine only the 5,000 undervotes; now the consortium wanted to look at 22,000 overvotes as well.
"It would be just as easy to make three piles as one," Keating says. But the Herald objected. If the county had to separate both sets of votes, a four-day job would turn into an effort that would take three times as long, Baron says. "We saw this as a potential delaying tactic."
Alliance members saw it as the Herald trying to thwart them. "Why do this if you are Marty, unless you are protecting your franchise?" said Fessenden when he heard of the Herald's objections. "The Herald wants to make sure they are first."
The consortium and the Herald shared the same law firm, Holland & Knight. That worked if seeking access was the goal. But now there was a conflict, and both would need to find new attorneys.
"It's clear to me all along they've tried to catch up with us," Baron said at the time. "Their latest action in Duval County would have the effect of delaying us. Whether that's what they intended or not, I don't know, but I'm not willing to give them the benefit of the doubt."
Monday morning, Judge Haddock held a hearing on the conflicting requests. Seibel and the Herald's attorney, Rutledge Liles, appeared in court. No one showed up for the consortium. The judge ruled for the Herald, saying it would be "prejudicial" to the newspaper to alter the sorting process. The consortium would have to wait. Keating was shocked.
"I believe in competition; that's not competition," says Keating. "To me competition is doing good journalism. That hearkens back to the old-fashioned thing wire-service reporters did when someone died. They'd go get a picture from the family and make sure they got all the other pictures in the house so the competition couldn't get one."

BY APRIL, THE CONSORTIUM members weren't just unhappy with the Herald. They had their own internal strife as well.
The Wall Street Journal and some other news organizations wanted help from NORC in working with the data. The Post and the New York Times didn't, since each has deep database teams eager to tackle raw data. Nor could members agree on how much time--a few days, a week or two--should be allowed between receiving the data and publishing stories. Many feared the data would be leaked or that "someone like Matt Drudge might catch wind of it and post them before we can," says the Wall Street Journal's Phil Kuntz.
Some worried that there might be eight different results, and they debated using certain standards for analyzing data. "The real competition," says Keating, "is in putting this stuff in context, not whether it adds up to 100 or 120."
In the final analysis, it will be up to readers and viewers to judge the validity of the dueling recount efforts. "Each study is going to rise and fall on its own merits," says Sean Holton, projects editor for the Orlando Sentinel, owned by consortium member Tribune Co. "As far as the public being confused, they might be."