Is TV To Blame? Well, Let’s Go to the Videotape

By Steven Luxenberg

The fault wasn’t in the data. The fault was in ourselves and the artifice that we in the media—primarily the TV networks—have constructed to “call” races on Election Night.

For about 90 minutes in the wee hours of Wednesday morning, from 2:20 a.m. until 3:50 a.m., the TV networks called George W. Bush the president-elect of the United States. He heard it. So did thousands of supporters huddling in the rain outside the state capitol in Austin and millions of viewers overseas. So did world leaders who sent congratulatory messages. And so did his opponent, Vice President Gore.

It’s easy for members of my profession to dismiss what happened as the inevitable result of competitive pressure and human error under an intense deadline. If we’re looking for someone and something to blame, we might say, as Carolyn Smith of ABC News’ Election Decision Desk did, “The data failed us in Florida.” And if we’re trying to make amends, we might apologize, as nearly all the network anchors did. “We don’t just have egg on our face,” said NBC’s Tom Brokaw, “we have an omelet.” On CBS, Dan Rather said, “We do the best we can on these calls. But we have to stand up and take responsibility and be accountable.”

This was not just another media blunder, however. The premature projection of a Bush victory—by Fox, CNN, NBC, CBS and ABC—had real consequences beyond the surreal spectacle of Gore making a concessionary call to Bush and then retracting it. It made the Bush campaign furious at Gore’s turnabout, and it left Bush and the Republicans feeling that Gore was a sore loser who was refusing to concede a legitimate victory. More importantly, it helped create the erroneous and dangerous impression that the nation had already entered—as historian Michael Beschloss said on ABC during the confusion—an “electoral crisis.”

There was no crisis at that moment (although we may see one before the business in Florida is over). All that had happened was this: We in the media had interrupted, for our own reasons and our own needs, a functioning system that hadn’t yet finished its work.

Much of the country was asleep when these strange and unnecessary events unfolded, so it’s worth recounting exactly what happened. The “decision desks” at the networks were receiving and evaluating a constant stream of data from Voter News Service, the consortium they and the Associated Press created in 1990 to handle exit polling and vote counting. At 2:16 a.m., Fox News called Florida for Bush. The rest of the dominoes fell, one by one, over the next three minutes. With definitive language like “It’s official” and “Bush wins,” the networks turned Florida red for Bush on their colorful electoral maps. The pre-loaded video graphics flashed the words that the governor had been waiting 18 months to read: “George Walker Bush, 43rd President of the United States.”

The crowd in Austin, watching on a huge TV screen, erupted into delirious cheering. The crowd in Nashville became funereal. The networks then leapt to cover the reaction that they had just created. The analysts began discussing how “President-elect Bush”—no longer “the Texas governor”—would conduct his transition.

It was a stop-in-your-tracks moment. For hours, the networks had been saying—rightly—that Florida was too close to call, that all the votes would need to be counted, that the presidency might turn on absentee ballots that were arriving from overseas. They had already been burned once—having given Florida to Gore early in the evening and then backing off just before 10 p.m.—and they had made clear that they weren’t going to make that mistake again.

They had forgotten the lesson they had just learned. Votes were still uncounted in Florida, and yet the networks put their faith in the same computer models, exit polls and mathematical formulations that had led them astray a few hours before. Only ABC’s Peter Jennings showed any true hesitation, but it didn’t stop his network from going along. It never seemed to occur to anyone to say to viewers, “We now have new data that, if accurate, shows Bush winning Florida. Now, this is just a projection, based on the available information. But if it turns out to be true, then the governor would have enough electoral votes to send him over the top. We all know how projections can go wrong, so at this late hour, with 96 percent of the Florida vote counted and after that awful mistake we made earlier, we want to wait for the election officials there to finish. It isn’t our job to declare the winner.”

Imagine for a moment that it’s Election Night again, and the networks have taken this approach. They don’t rush to be first to declare Bush president, so Gore never calls him. The Florida law that requires a recount (if the vote margin is less than 0.5 percent) kicks in before Gore has a chance to create the impression that he is the one dragging out the election. The candidates are left in their proper position, awaiting the outcome, rather than put in the uncomfortable position in which Bush and Gore found themselves at 3 o’clock in the morning.

The networks’ call had a huge psychological effect on the electorate and the candidates, with political and historical ramifications. Voters might have shrugged off a long count in Florida as an acceptable delay in a contested election. But a “reversal”—taking back something that was “decided”—fed the notion of a tainted result, contributed to the overheated rhetoric on both sides in the past few days and helped fuel the sense of a country in crisis.

There’s nothing inherently wrong with exit polls or computerized voting models. They can be useful tools for understanding the electorate and how it voted. But we are letting them become a substitute for the real thing. Millions of dollars are spent to reveal something that would be known anyway a few hours later. And in a case where a race comes down to less than one-half of 1 percent out of 6 million, what could be more exciting than reporting the totals in each Florida county as they are announced?

The live nature of television magnifies this kind of mistake. At The Washington Post, after the networks called the election for Bush, we prepared a story and a new front page with the headline, “Bush Appears Victorious.” Fortunately for us, it takes about 15 minutes before a new front page can be put on our eight presses; that gave us time to do more reporting. When our executive editor learned that Bush’s lead was shrinking, he killed the front page.

Our readers weren’t watching as our editors agonized over what to do. But it’s not that hard to get it right, if we commit ourselves to covering the news rather than making it. I discovered this myself one night, in 1978. I was working at the Baltimore Sun and was assigned to observe the voting tally for the Democratic gubernatorial primary. It wasn’t supposed to be a close race. The last poll before the election had the incumbent acting governor, Blair Lee, leading an upstart challenger, Harry Hughes, by 14 percentage points.

Reporters around the state were phoning in vote totals as soon the counties reported them. Hughes was running ahead, astounding everyone, but the tally from Montgomery—Lee’s home county—wasn’t complete yet. It was possible that a heavier-than-expected vote for him there could turn the election, but it seemed increasingly unlikely. The reporter writing the story was on deadline for the edition that went to most subscribers, and the editors wanted to know—as soon as possible, dammit!—when to call the race for Hughes. We waited until it was clearly impossible for Lee to overcome Hughes’s lead.

The final outcome wasn’t even that close—Hughes won by 19,000 votes out of 573,000 in the four-person race. So you could argue that we were being a bit too cautious. But I never thought about it that way: My assignment was to cover the unofficial returns as they were reported, not to project a winner.

There will be calls for better modeling, better ways to find discrepancies, better safeguards to avoid what happened in this remarkable election. That’s fine. But let’s not call the election better next time.

Let’s not call it at all.

© The Washington Post; printed here by permission

The Washington Post
Sunday Outlook
November 11, 2000