Lesson Learned   

A behind-the-scenes look at election night coverage

CBS anchor Dan Rather, who had experienced a rough few months, set the tone for election night coverage at 8:22 p.m. with a memorable offering from his stable of trademark axioms. "Beware of certitude," Rather warned viewers.

Early and misleading exit polls had seeped onto Web sites such as the Drudge Report and Slate and ricocheted through the blogosphere. Left-leaning bloggers were dishing merrily about Kerry's early leads in the pivotal states of Florida and Ohio. Journalists privately pored over those numbers. But the networks, memories of the 2000 election night meltdown all too vivid, avoided them.

"We never reported any of those exit polls here on CNN," anchor Wolf Blitzer piously told President Bush's senior adviser Karen Hughes after midnight, then asked her about the mood at the White House when officials saw the early numbers.

The campaign season had been punctuated by major media missteps (the CBS National Guard story), by reckless replays (the Dean Scream) and by disproportionate attention to unproven allegations (the cable networks' handling of the Swift Boat Veterans for Truth). During the summer, the media fixated relentlessly and depressingly on the candidates' military records from three decades past.

But the media also provided serious issues coverage, careful fact-checking and meticulous investigations throughout the campaign (see "Campaign Trail Veterans for Truth"). The Los Angeles Times, the New York Times and other publications helped debunk the Swift Boaters' claims. The Baltimore Sun's Susan Baer wrote a prescient page-one article in October about same-sex marriage as the "new front in culture wars." The L.A. Times' Ronald Brownstein contributed a late- October story about the Kerry campaign, in a "virtually unprecedented move for a Democrat," all but surrendering the South.

On election night, the media mostly veered toward these loftier instincts. The networks turned suddenly solemn, their coverage restrained. In contrast to the vertigo-inducing debacle of four years ago--when the networks first awarded Florida to Vice President Al Gore, then to George W. Bush, then retracted it again--TV journalists in 2004 heeded Rather's warning. The usual race to scoop competitors became an unlikely contest in most ostentatious show of responsibility.

Such sober coverage doesn't make for riveting television. On November 2, "caution" was the shibboleth of election night analysis. Networks awarded key states to candidates very, very slowly.

New York Times writer Alessandra Stanley characterized their strained efforts to avoid the 2000 blunders as "almost painful," adding it "was the wise, responsible thing to do (quite literally, politically correct), but it left the anchors without much to say."

"Unfortunately," the San Francisco Chronicle's Tim Goodman complained in a column, "what we witnessed was a night packed with eggshell walking, overwhelming caution and relatively good manners. Exciting and fun? No. Informative. Not really. A humongous time killer? Check."

While television coverage was unusually tentative throughout the evening, political analysts and media critics give the networks solid marks for their performance.

"They did a good job of holding back," says Sherry Bebitch Jeffe, senior scholar at the School of Policy, Planning and Development at the University of Southern California, who spent election night on the set of KNBC, the NBC-owned station in Los Angeles. She praises NBC and CNN for their "quite deliberative and quite cautious" coverage and notes the afternoon drumbeat of early exit poll leaks signaled a momentum--the coveted Big Mo--for Kerry that proved false. "People weren't hyperventilating all over the place too early," she says. "They didn't buy in too deeply to the 'Faux Mo.' "

William Powers, a National Journal columnist who covers media and politics, concedes that "from a purely entertainment point of view, there was less drama," but he doesn't think entertainment should be the "primary function of what these mass outlets do on Election Day." Powers senses an awareness among the networks that "they're competing to get as close to the truth as they can because otherwise there's going to be a price to pay in reputation." The networks once again would be held up for popular-culture pillory through vehicles such as Michael Moore's films.

Jack Pitney, a professor of government at Claremont McKenna College in California, admits "a lot of us political junkies would have enjoyed quicker calls, but it's hard to see how the Republic would have been served by a call at 1 a.m. instead of 2 a.m." Given scars from 2000, the 2002 collapse of the Voter News Service exit-polling system and the "underestimation of the Republican vote in 2004," Pitney thinks the networks were "appropriately careful."

But he would have enjoyed a bit more variety. "They could have talked about some of the individual races for Congress and the Senate," Pitney says. "After awhile, you learn all you need to learn about provisional ballots in the state of Ohio. What would have been more edifying for viewers is a discussion of how redistricting in Texas" led to GOP gains in the House of Representatives.

The National Election Pool--a media-owned consortium that replaced Voter News Service and is composed of the principal TV networks and the Associated Press--sent the first wave of exit-poll data to subscribers around 2 p.m. on November 2. Other media subscribers also pay to see the results, and participants pledge not to publicize or leak the early numbers.

But leak they did. The figures were whizzing through cyberspace by mid-afternoon. Jack Shafer, Slate's media critic and editor at large, explained in his Web post that "Slate believes its readers should know as much about the unfolding election as the anchors and other journalists, so given the proviso that the early numbers are no more conclusive than the midpoint score of a baseball game, we're publishing the exit-poll numbers as we receive them."

The surveys, which created a false impression that Kerry was rolling toward victory, had the "biggest partisan skew since at least 1988," said a November 5 article in the New York Times, which had obtained a copy of a report by the system's creators exploring why the exit polls were misleading.

Although those early numbers did not predict final results, Shafer stands by his decision to post them. "The networks were every bit as remiss this year in telegraphing what was in these confidential exit polls that they're not supposed to reveal as in previous years," Shafer says. "They need to be demystified. It's better for my readers to know why a television broadcaster seems to be so up on Kerry at 5 o'clock."

Political bloggers quickly seized on the numbers. In a hip "Marketplace" graphic on November 4, the Wall Street Journal interspersed updates from bloggers and network television throughout the afternoon and evening. "PA and Ohio margins widening for Kerry, Florida, Wisconsin tie in early returns," the blog Wonkette.com reported at 9:10 p.m.

As the Journal article noted, WSJ.com, the newspaper's Web site, posted a mid-afternoon article saying exit-poll data "purported to give Mr. Kerry an early lead in several key states" and linked to a blog with the numbers. The Journal's Web site also questioned their validity.

"I know there'll be a lot of gnashing of incisors tomorrow about blogs leaking leaked exit polls," said a November 2 post by Jeff Jarvis, the blogger behind buzzmachine.com, who also runs online services for Newhouse Newspapers. "To hell with the gnashing. We're all big boys and girls. We can decide whether to (a) believe it and (b) vote on our own. Let information be free. Let us know what big media and big politics know. Transparency, man, transparency."

Jarvis, a former Sunday editor for New York's Daily News and former TV critic for People and TV Guide, argues blogs "were saying what they knew and the journalists were not saying what they knew." He says "big media" should "tell us what they know: 'The exit polls look like Kerry, but boy, they were wrong last time.' "

But Kevin Drum, who blogs for Washington Monthly, a small but influential politics and policy magazine, thinks the networks rightly avoided exit polls and the bloggers rightly published them. "To me, it's just data," says Drum, who shared early exit-poll data in one post. "If you're interested in looking at it, go ahead and look at it." But the networks are watched by millions of people, and if you're among those viewers, "you're sort of forced to see it."

Jonathan Dube, the managing producer for MSNBC.com and publisher of CyberJournalist.net, notes that Slate and Drudge also published exit polls in 2000. "It's more of a phenomenon that was created by the Internet, and Weblogs are just one way the information is being distributed," Dube says.

But he thought blogs' coverage on election night "got very skewed based on exit polling they were publishing early on." While one of the virtues of bloggers "is that they are independent and will publish what they want, when they want, certain bloggers did a good job of putting certain caveats" on the numbers, Dube says, and "others didn't do quite as good a job."

Network election coverage began at 7 p.m., with each unveiling special sets notable for glitz and excess. NBC coined the silliest set name, dubbing its sky booths outside Rockefeller Center "Democracy Plaza," a phrase which recalled--hopefully inadvertently--administration platitudes such as the "coalition of the willing," "they hate our freedom," and Bush's favorite: "Democracy is on the march."

CNN's set was the most distracting. Not only were video wall graphics and precinct numbers often difficult to see, but the rented Nasdaq site in Times Square looked so much like a running track that it seemed to beckon an earnest Blitzer to relax and take a lap. A "Today" show-style glass window behind the anchors allowed anxious viewers at home to see vapid viewers outside the studio waving to the watchers.

Fox News Channel provided the clearest and most helpful graphics for election night returns with its "Election Ticker," which continuously displayed raw vote numbers, percent of the vote for each candidate and percent of precincts reporting.

Michael Barone, a Fox News contributor and senior writer for U.S. News & World Report, analyzed the tabulated vote in swing states such as Florida and Ohio county by county to explain where votes counted so far had originated, why that appeared to benefit Bush and how Kerry still might be able to prevail. Anchor Brit Hume handled questions deftly, keeping Barone's discussion of the numbers accessible to viewers who lack his intimate knowledge of the political landscape.

As Shafer argues, early network coverage did indeed allude obliquely to preliminary exit polls that favored Kerry. Fox News commentators initially looked "stunned and somber as they hinted that early voter surveys showed Mr. Kerry doing better than expected," the New York Times' Stanley wrote the next day.

On CNN, "Crossfire" pundits downplayed Bush's chances in Ohio. Conservative cohost Robert Novak announced Republicans in the state were "very pessimistic." They're "looking at the returns; they think it's going to be very tough for Bush" to carry the state, Novak said.

On MSNBC's "Hardball," reporter Lisa Myers enigmatically informed viewers in the 8 o'clock hour that Bush's campaign was finding "that the exit-poll data is significantly underrepresenting the Republican vote."

Despite these veiled hints, the networks held back on risky calls. Linda Mason, the CBS News executive in charge of its "decision desk," characterizes her network's approach as "cautiously aggressive." CBS was the first to call states including South Carolina, Virginia, Mississippi, Pennsylvania, Colorado, Michigan and Minnesota.

CNN Political Director Tom Hannon says he was not monitoring "call sheets" that night tracking what time other networks projected winners. "When the other networks called [a state] was not a factor in decision making," Hannon says.

As midnight neared, the networks offered markedly different approaches to projecting winners in pivotal states.

At 11:39 p.m., ABC called Florida for Bush, followed by CBS. CNN waited until 12:10 a.m., Fox until 12:22. "The race there was very tight," says Mason, who wrote a report for CBS on the 2000 election fiasco. "Of course, Florida holds a certain symbolism for us, so we wanted to be super careful."

Marty Ryan, Fox's executive producer of political programs, says "certainly it was in the back of our mind what happened in 2000," when his network was the first to call Florida for Bush. Ryan says the four-person decision team at Fox wanted to be particularly careful because Florida had extensive early voting. "There was no early gauge to assess how early voting went," he says, and the exit polls became "problematic."

At 12:41 a.m., Fox was the first to call Ohio for Bush, followed by NBC and MSNBC around 1 a.m.

Fox News Senior Vice President John Moody consulted with a pollster and two statisticians hired by the network and decided Bush's more than 100,000-vote lead in Ohio gave him an almost insurmountable edge over Kerry, who would need a huge majority of the as-yet uncounted provisional ballots to triumph. (Provisional ballots are cast by registered voters who have moved but not updated their registration or believe themselves to be eligible but do not appear on the rolls.) "It became a little bit of a math problem" to see how Kerry could prevail, Ryan says.

But ABC, CNN and CBS declined to call Ohio for Bush. At CNN, the decision team was hearing conflicting reports about the number of provisional ballots, with Ohio's secretary of state estimating about 170,000 and Democrats projecting some 250,000. "There was a clear possibility that there could be a dispute over the number of ballots," CNN's Hannon says. "Had we known exactly how many provisional ballots there were with certainty, I think we would have made a call."

CBS declined to call Ohio because Mason and her team believed as many as 250,000 provisional ballots could be outstanding. "In addition to that, there were all those lawyers in Ohio ready to sue on various other counts," Mason says. "It looked like this could be [a repeat of] Florida and be in turmoil for weeks. It would be silly for us not to take all this into account."

But CBS did call Nevada at 3:45 a.m. Before the announcement, Mason told CBS News President Andrew Heyward that her team planned to call Nevada for Bush, and she expected NBC and Fox to follow suit. On those networks, the Nevada call would push Bush over 270 votes in the Electoral College, securing his reelection. Because CBS had not projected a winner in Ohio, it would not be in a position to announce Bush's victory. Does that bother you? she asked Heyward. No, he replied.

ABC called Nevada for Bush about 10 minutes later. Mason expectantly watched NBC and Fox, but neither called Nevada.

"Everybody has a little better knowledge or confidence in different states they look at and concentrate on different things," Fox's Ryan says. The Fox team felt New Mexico, Iowa and Nevada each had "different issues" that precluded projecting a winner. In Nevada, the team was concerned about where the votes were coming from and how that might impact the overall margin.

At about 2:30 a.m., Moody and Ryan went to the studio to explain the situation to Brit Hume and to tell him that unless something occurred to resolve their concerns in the next half hour, they would hold off until morning.

Bush's Electoral College margin on Fox, NBC and MSNBC hovered the rest of the night at 269 votes--a number that would have allowed Kerry to tie, at best, and thrown the election into the GOP-controlled House of Representatives. No network awarded Bush the full 270 votes.

Major newspapers, too, avoided making definitive declarations in their headlines. "Bush nears victory but Ohio count is in dispute," said USA Today. "Bush Holds Lead/Kerry Refuses to Concede Tight Race," said the New York Times.

Papers around the country echoed those themes. "Down to Ohio" declared the Akron Beacon Journal. "WE WAIT/ Bush close to victory, Kerry clings to Ohio," said the Arizona Republic. "SQUEAKER!" announced the Charlotte Observer.

Like their TV counterparts, many newspaper reporters saw early exit polls, and some crafted preliminary stories relying on those numbers.

Jack Torry, a Washington reporter for Ohio's Columbus Dispatch, helped write two different analysis stories on election night. The first, never published, was based on late afternoon exit polls and explored how Kerry won the election, a "referendum on President Bush."

"The more I kept checking with Republican sources who I really do trust, the more I began to wonder: Could those exit polls be right?" Torry says. He and his colleagues scrapped the first piece around 9:30 p.m. Their second analysis, which actually appeared in the paper, examined why the race was so tight.

But not all early misfires based on exit polling stayed out of print. The November 5 story by the New York Times' Jim Rutenberg referred to a Times snafu. "The New York Times removed an analytical piece about the vote based in part on the Election Day survey from its later editions," Rutenberg wrote.

That story, by R.W. Apple and Janet Elder, also appeared in the International Herald Tribune. Bush's "bid for re-election was weakened by his failure to compete on even terms with Senator John Kerry among the millions of new voters who cast ballots on Tuesday, preliminary data from exit polling indicated," said the Herald Tribune version. The article said "Kerry clung to narrow exit-poll leads" nationally and in Florida, Ohio and Wisconsin. It did not explain that these early surveys might be skewed or misleading.

New York Times Washington Editor Richard L. Berke, who oversaw his paper's election coverage, says the story, which also contained many voter interviews, aimed to give readers a sense of the demographics and voting patterns. "Traditionally, those sorts of things are more reliable in exit polls even if you don't know the outcome," Berke says.

But as the evening progressed, actual vote counts contrasted jarringly with exit-poll results. "It made us nervous, so we pulled the whole story," Berke says. "I wanted to reshape the story to focus on all the [voter] interviews, but we ran out of time, so we pulled the whole thing."

Although Berke "felt OK at the time about using the exit polls for demographic information and not using it to call states or the outcome," he wants to review how they're used on election night in the future. "I'd be lying if I said it didn't affect our thinking and our planning for what the outcome could be," he says. "We have to give second thought to how and if and whether we use the exit polls at all."

The close of an election inevitably, and rightly, ushers in a period of self-examination for political journalists--after they catch up on some much-needed sleep.

"Across the campaign, I thought there was a lot of very good coverage that demonstrated that newspapers and broadcast networks do learn more each cycle about how to cover campaigns," says Doyle McManus, Washington bureau chief of the Los Angeles Times. "You got a lot, I thought, of quite impressive coverage on phenomena like fundraising and the use of television commercials, and quite sophisticated demographic analysis by my colleague Ron Brownstein and others."

The media correctly portrayed the race as tight in the campaign's final days. "People were describing Ohio as a critical swing state; they were describing the race as a competition to get out the base vote on each side," McManus says. "In retrospect, all of that looks remarkably accurate."

McManus also credits newspapers with giving "quite full, detailed and sophisticated coverage to all the major issues," despite the usual carping from critics that journalists ignored issues in favor of the horse race. But "we need to find more ways to direct our readers to places on the Internet or elsewhere," McManus says, "so we can escape the trap of having written about Issue X in early September when our readers didn't get interested [until] early October."

Of course, McManus adds, the media did fall "prey to the temptation to chase rabbits," to fixate on what he calls "ephemeral controversies" such as the Swift Boat saga and the missing explosives in Iraq. "But each of those ephemeral controversies was a metaphor for a larger issue," he says. "They weren't trivial, just ephemeral."

Not everyone gives the media such high marks for their performance. National Journal's Powers faults the usual "herd thinking" for chilling more creative political journalism--and he's not talking about the sort of creativity displayed by CBS during the National Guard document fiasco.

Powers says coverage was "way, way too poll driven" and blasts polls as an "addiction" for the media. "It's very hard to come up with something new every day. And these numbers provide newness," Powers says. "It zaps a lot of resources and energy and reduces the campaign to numbers in a way that is not helpful to democracy."

Another Powers peeve: "Every four years, we have this story line about the youth vote--'The youth vote is going to be gigantic.' The story never pans out, but we kind of go through these ridiculous rituals."

And don't get him started on the media's "gigantic obsession" with red and blue states. "The red-blue theme is so overdone and also really reductive," Powers says. "We have this red-blue motif because we have a two-party system in which we go into the booth and are supposed to choose between two parties," not because everyone falls neatly into a red or blue category. "Yet we play this story up because it's sexy; it's easy; it's simple. It was easy to have that map after the 2000 election, but it portrays the country far too simplistically and does a disservice to the public."

The media also would better serve the public--and the English language--by bucking the prevailing clichés of each election season. In 2004, candidates "barnstormed" the country, appealing to one-time "soccer moms" who metamorphosed into "security moms" after the terrorist attacks. Pollsters combed those "red and blue states" to gauge the preferences of a "polarized" and "closely divided" electorate, wondering whether the "youth vote" could tip the election but flummoxed by the "cell-phone generation." In the end, as pollsters and journalists foretold, it all came down to the "ground game" in Ohio, "the Florida of 2004." Sort of.

Fundraising prowess fueled media attention, in some cases leading to sophisticated reporting, such as the Washington Post's two-day series in May exploring links between fundraising and access to the administration (see "Follow the Money," August/September), but often producing a mind-numbing array of horse-race, money-chase dailies.

During the political conventions, the broadcast networks abdicated their civic responsibilities in search of higher ratings from sexier reality shows. The cable networks galloped into the void, but too often padded downtime with shouting heads and insipid spin from party officials.

There was rampant media speculation throughout the campaign: When would a whipped Kerry admit defeat and drop out of the primaries? Whom would a triumphant Kerry tap as his running mate? ("KERRY'S CHOICE: Dem picks Gephardt as VP candidate," blared the New York Post in one of the season's more entertaining media meltdowns.) Might Bush drop Dick Cheney from the ticket? When would Kerry sideline Bob Shrum, the famed Democratic adviser and perennial presidential campaign loser? Did a home loss by the Washington Redskins the Sunday before the election presage victory for Kerry?

But the networks pulled themselves together for election night--at least for a few hours.

By 1:26 a.m. on November 3, hours before a victor had been officially declared in the 2004 presidential race, anchors and pundits, perhaps reeling from their night of restraint, already were speculating about Democratic candidates in 2008. On NBC, Washington Bureau Chief and "Meet the Press" host Tim Russert predicted a battle for the "soul of the Democratic Party."

After all, the next presidential election was only 1,462 days away.