Press Box

Exit-Poll Fetishism

As you channel surf from MSNBC to Fox News to CNN this afternoon seeking news about the Arizona and Michigan Republican primaries, I invite you to help me collect the Best Performances by TV Journalists Who Feign Ignorance About Who’s Winning.

They’ll hem and haw about who might be leading. They may even add a little suspense to the broadcast by noting that there’s a big turnout of voters, and how that could be good news for McCain. Or bad news for McCain. They’ll tug their forelocks, but they’ll never confess that they already know who’s winning based on the exit-poll data they’ve seen, which is produced for the networks and newspapers by the Voter News Service. In short, the broadcasters will be acting.

Some will craft horrible performances, such as the one Brit Hume turned in the night of the New Hampshire primary. As the Washington Post reported, Hume told his viewers at 6 p.m., an hour before most polls had closed, that “all signs point to a very good night for John McCain.” Hume, of course, knew from the VNS exit-poll data that McCain had fannywhacked Bush bloody. If Hume had any sense, he’d turn in his Screen Actors Guild card.

Likewise, the New York Times caught NBC’s Lisa Myers blowing her lines on MSNBC, when at 6 p.m. she mused about how George W. Bush would face “a loss, particularly if he loses big.” Of course, Myers–an honest journalist and a lousy method actor–knew there was no “if” about the loss, and it showed.

Perhaps the Greatest Acting Performance by a Network Anchor on Election Night of all time was captured by the Los Angeles Times (Feb. 17, 1988). At 6:57 p.m., when the New Hampshire polls were still open, Peter Jennings went on the air and said:

“Based on everything we see so far, that’s the total vote counted so far, the key precincts in some of the areas around the state and even some of the exit polling information … George Bush is going to be able to pull it out by just that.” Jennings then pinched his thumb and index finger close together.

The reason TV reporters preen and pretend all Election Night about who is winning has less to do with theatrical ambition than it does with the news embargo the broadcasters have yoked themselves to as members of VNS, the consortium of ABC, CBS, NBC, Fox, CNN, and the Associated Press. On Election Nights, VNS collects and distributes polling data to its members and about 150 other news organizations, including most major newspapers, and VNS members are prohibited from spilling the exit-poll beans before most of the polls close.

The VNS consortium does its best to avoid scrutiny from the press. It doesn’t have to try too hard. An exhaustive Nexis dump on VNS reveals that in recent years only USA Today and Roll Call have taken even a remote interest in what this monopoly provider of election data is doing. One suspects that the New York Times, the networks, and other subscribers avoid reporting on VNS for the obvious reason: They don’t like to discuss themselves in front of their audience. One also imagines that they’re secretly ashamed of their participation in the VNS information cartel. Journalists ordinarily love to compete, and the non-competitive quality of VNS must rankle editors as much as it pleases green-eyeshades.

The networks basically did what they wanted to with their data until 1980, when their projection of Ronald Reagan as the winner of the presidential contest–hours before the West Coast polls had closed–angered Democrats. The Dems blamed the losses of several Democratic representatives on the unwillingness of Democrats to go to the polls once they knew Jimmy Carter had lost. But the best academic work on the 1980 turnout, by George Washington University professor William Adams, found that fewer than 3 percent of potential voters stayed home because of the Reagan projection. So much for knowledge of the probable winner suppressing the turnout!

The ritual suppression of Election Night exit-poll data became journalistic religion in 1985, when the broadcasters came under attack from Congress, which believed that projections based on the use of exit-poll data would tamp down voter turnout. In a compromise move, the networks agreed to hold back projecting statewide winners until most of the polls in the state closed. For the most part, the agreement has stuck.

The embargo places a terrible burden on reporters, who are paid to disseminate information and are rotten at keeping secrets. On Election Day, the VNS exit-poll data cuts a wide sluice through the press corps and washes down on anybody with serious interest in politics, as reporters ignore the official embargo and share it with their colleagues, friends, politicians (who do their own exit polling), and other members of the elite. The freewheeling distribution of the forbidden knowledge makes a mockery of the embargo and makes chumps out of most voters, who have been aggressively following tracking polls and other measurements of support for their candidates and want the news to keep coming.

Only because VNS does such a rotten job of maintaining its own embargo was I able to get the numbers from the New Hampshire and South Carolina primaries and can I report the first wave of raw exit-poll results from the Michigan and Arizona contest. One caveat: These are raw numbers, and should not be regarded as projections of the ultimate winners. Projections are another beast, made by combining exit-poll data, previous polling data, a measurement of the size of the return, and other information. It’s a job that VNS and the networks are pretty good at.

Michigan 2 p.m. exit-poll numbers:

John McCain:

48 percent
George W. Bush: 46 percent

(Whoops: For about 30 minutes I had the wrong data up for Bush. My apologies.)

Later numbers from Michigan:

John McCain: 47 percent
George W. Bush: 47 percent

These numbers parallel the last Zogby poll going into the Michigan race, which were: John McCain, 48 percent; George W. Bush, 46 percent.