Jan 1 2001

Election Night Meltdown

Media monopoly contributed to exit-poll errors

No sign of journalistic caution attended the pulsating graphic Fox News flashed across screens late on Election Night 2000:

George Walker Bush,

43rd President of the United States

Tom Brokaw was no more circumspect in interrupting NBC’s broadcast to announce, “George Herbert—George Walker Bush wins!” And CBS’s Dan Rather left no doubt when he announced: “Sip it, savor it, cup it, photostat it, underline it in red, press it in a book, put it in an album, hang it on the wall. George W. Bush is the next president of the United States.” Rather had earlier made the unfortunate boast, “If we say somebody’s carried a state, you can pretty much take it to the bank.”

In fact, the networks’ premature call of the decisive state of Florida for Bush—preceded by an erroneous awarding of the state to Gore—shows that election night projections are anything but bankable. While responsibility for what is put on the air rests ultimately with the journalists who put it there, much criticism for blowing Florida projections has focused on faulty exit-polling data provided by the secretive and little-known Voter News Service (VNS).

A consortium financed by ABC, CBS, NBC, CNN, Fox and the AP, VNS has operated as a virtual monopoly providing exit-polling data and election returns to subscribing media outlets since 1990. (The L.A. Times is apparently alone in having an independent exit-polling operation.) Staking out 1,400 precincts in all 50 states, VNS surveys voters leaving the polls, and checks their responses against actual returns from 3,500 sample precincts across the country. This information, including projections, is then distributed via computer to subscribers. VNS had 45 exit-polling locations and 110 sample precincts in Florida for Election 2000 (AP, 11/17/00).

Although final assessments of what went wrong are still being prepared by television news operations, most reports fault VNS for providing the bad numbers that resulted in early calls projecting Gore as Florida’s winner. NBC and its cable sibling MSNBC declared Florida for Gore at 7:49 p.m. (EST). Over the next 13 minutes, CNN, Fox and CBS followed suit, with ABC not making the false call until 8:02 p.m.

The Panhandle problem

Networks began projecting Florida for Gore 11 minutes before the polls closed in Florida’s western Panhandle, which unlike the rest of the state is in the Central time zone. This action prompted protests and threats of congressional hearings by Republicans, who made the improbable charge that the erroneous call discouraged up to 20,000 voters from voting in the heavily Republican region.

Fred Barnes, co-host of Fox News Channel’s Beltway Boys (11/11/00), said he’d heard reports “that as many as 20,000 voters actually left the lines. They were lined up late in the day and went home after hearing that Florida had gone the other way.”

AP (11/14/00) quoted one Panhandle resident as saying the early projections “robbed me of my right to vote. . . . I figured it wouldn’t do me no good to go vote.” AP noted, “Watson decided not to make the trip of about 20 minutes to his polling place.” It’s a good thing too; he would have been disappointed; the polls would have been closed for at least nine minutes by the time he got there.

Given that Panhandle residents voted throughout the day at the rate of about 500 per minute, it’s extremely unlikely that 20,000 of them ever planned to vote in the last 11 minutes of the day. With a very competitive Senate race on the same ballot, it’s questionable how many would have chosen not to vote even after being falsely assured that the presidential race was no longer in doubt. One might even suspect that the networks’ repeated assertions, from 7:00 to 7:49 p.m. Eastern time, that the Florida race was both too close to call and crucial to the national results did more to boost turnout in the Panhandle than the 11 minutes of the wrong call did to depress it.

All of the outlets corrected their Gore gaffe within minutes of each other, beginning with CNN just before 10 p.m. Some four hours later, at 2:16 a.m., Fox News was first to project Florida’s 25 electoral votes for Bush, prompting the network’s declaration of Bush as the next president. Within four minutes the other networks followed.

Networks insist they made their own independent decisions in wrongly calling Florida for Bush early Wednesday morning, but the timing suggests that the hasty calls were the product of faulty numbers, as well as a rush to follow Fox’s lead (AP, 11/17/00). Inaccurate tallies put out by Florida’s Volusia County may have also contributed to inaccurate projections.

The emphatic but premature declarations of a Bush presidency are more likely than the earlier Gore flub to have had a real impact. For one, they prompted Al Gore to concede, cementing for many the early images of Bush as the president-elect, and of Gore, once he’d retracted his concession, as the “sore loser.” Such impressions may have significantly altered the dynamics of the post-election wrangling.

Rejecting the call

Against the claim that competitive pressures led to a rush to judgment on election night, it’s important to note that the networks’ collective failures were partly due to a lack of competition. Before 1990 each network provided its own exit polling. Pitching in to fund VNS was justified at the time as a cost-cutting measure arising from pressures to be more profitable. After the election, according to Broadcasting and Cable (12/4/00), the American Antitrust Institute “asked the Justice Department to break up VNS, blaming the election-night mishaps on the absence of competition for data among networks.”

Of course, whenever polling projections are the centerpiece of network coverage, with definitive language being used to “call” elections based on a fraction of the results, it may boost ratings, but it’s not factual reporting.

Yet in all the discussion about election-night media performance, few commentators question the practice of making projections the central focus of coverage. In one exceptional review, Washington Post editor Steven Luxenberg wrote (11/12/00), “We in the media had interrupted, for our own reasons and our own needs, a functioning system that hadn’t yet finished its work.” Luxenberg offered this advice for the future: “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.”