Why is this night different from all other nights? On all other nights, the media go after the big story, find out as much as they can about it, and report what they know. But on Election Night, the media strenuously avoid the big story, try not to find out too much about it, and are careful not to reveal what they know.
On all other nights journalists try to get the story before anyone else does. This is called a "scoop" and is generally regarded in the profession as a good thing. But on Election Night journalists forswear scoops and brag about their self-restraint.
On all other nights citizens prefer that journalists tell them the truth and become quite agitated if they believe they are being kept in the dark or lied to. But on Election Night citizens demand to be kept in the dark and become agitated if they are told the truth.
There is only one big story Election Night: Who won? Even under the best of circumstances, this is a problem for the TV folk. Never is there more time to fill and less material to fill it with. Party conventions are as narratively rich as The Sopranos in comparison. And thanks to exit polls, the people reporting the election generally know the likely outcome many hours before they report it. But journalists, politicians, commissions of high-minded worthies, and many ordinary citizens are under the delusion that it corrupts the electoral process for the media to report what they know.
So, every election cycle the self-imposed rules get tighter about when the networks may "call" an election. At three seconds past the appointed hour, the result is announced as if the heavens had suddenly opened to reveal it. A decade ago the Voter News Service was formed to monopolize the exit-poll business and make it harder for any one network to get the story first. (By no coincidence, doing it jointly also saves them pots of money.)
After the Florida double-miscall disaster of 2000, the networks spent millions revamping the VNS computer system and vowed extra-special caution about calling any race too soon. A couple of them went so far as to "cloister" their exit-poll analysts (as the New York Times put it) to prevent the information everyone wants to hear from leaking onto the airwaves prematurely. When I worked at CNN in the 1990s, there were special terminals on Election Night where you could punch in any major race and get the probable result. But you not only could not reveal this information to viewers—you could not even acknowledge that you had it. Oh, what tangled webs we weave when first we practice to deceive.
In any event, the new VNS computers didn't work. As Marty Plissner explained Wednesday in Slate, the networks did have the raw exit-poll numbers, which they were not supposed to use until polls closed, but they didn't have the rich demographic data, which is what the exit polls are supposed to be for. ("Women under 5 feet, 2 inches who drink gin more than once a week are tilting Democratic tonight, Dan.") I don't know about you, but the election-night punditry struck me as no more or less Delphic than in earlier years, when more data was available and the rules about using it were looser.
After all, even without exit polls, there still are polls of all sorts conducted right up to the day before the election. And there is no anathema on using these—indeed they are the basis for most election-night commentary. The main difference between a poll taken the day before an election and a poll taken as people exit the voting booth is that the exit poll is probably more accurate. So, the net effect of keeping exit-poll results off the air for hours is to make election coverage less accurate on average, not more so.
Of course it is generally felt that people can be trusted to understand that a non-exit poll is just a sampling and may not be accurate. And no one seems to feel that his or her vote has somehow been "stolen" by a regular poll showing that a particular race is a foregone conclusion. But about exit polls, people are regarded—and millions apparently regard themselves—as incapable of such understanding.
Arguing with superstition is generally futile, but let me try one more time.
Look, despite pious civic propaganda, it's just not true that your one vote could determine the result. That never happens, even in Florida. Individually, our votes don't matter, but cumulatively they do. That is a logical conundrum, and you can decide for yourself if it makes voting worthwhile. But whatever you may decide about that is just as true (or just as false) no matter what time you vote. And it is just as true (or false) whether you know how others have voted or you don't. Exit polls, in short, have nothing to do with it. They cannot steal your vote.
Fairness to individual late voters aside, does reporting of exit polls while people are still voting sometimes affect the result? There is no logical reason why this should happen—are you more likely to drop out of line if you hear that your candidate is projected to lose or to win?—and no evidence that it does happen. But isn't it bad if people don't vote, even if it doesn't change the result? Well, maybe. But promoting superstition, ignorance, and deceit seems worse, and restrictions on exit polls involve all three.
Now that VNS has blown it for the second time in a row, maybe some network will say, "Screw this. From now on we're going to find out the one thing people want to know, we're going to find it out as soon as we can, and we're going to tell people as soon as we find out." What did they used to call that? Oh, yes: news.