Have you read Martin Plissner's rollicking memoir of his years as CBS News' executive political director, The Control Room, which explains the origin of the myth of the suppressed voter turnout? In 1980, NBC used exit-poll results and other voting data from the eastern states to call the presidential election at 8:15 Eastern time in favor of Ronald Reagan—almost three hours before the West Coast polls closed. A few days later, the California secretary of state claimed that the media projections turned "would-be voters" into "non-voters," and said that "election volunteers did not show up and voter information phones stopped ringing during a time when they are usually all tied up." Congressional Quarterly reported that many voters abandoned their places in polling lines when they received word of the Reagan victory.
Although the voters-leaving-the-polls story has cemented itself in our political folklore, it has never been substantiated. CBS News even assigned reporters to document the vanishing hordes created by the NBC projection and had no luck in finding them. "Newspaper reporters who wrote of such happening were queried as to the time and place," writes Plissner. "Again, no luck. They had only heard from someone else, or read somewhere, that it had happened somewhere."
According to Plissner, academic studies don't agree on whether projections change voting patterns. But the surveys seem to agree that most western voters believe that other voters are likely to change their mind about voting after hearing a projection, but not them. That's a pretty shaky stick upon which to hang your hat.
Yet the myth continued to grow. After the 1980 election, CBS crews covering California stories encountered hostility over the projection issue, Plissner reports. "On one such occasion, a CBS producer, in an aside to a colleague, muttered, 'Vote early or move,' " he writes.
Exactly! If, on a long shot, exit polls and projections do indeed suppress turnout, voters can always flummox the media by voting early or requesting an absentee ballot. There! I've solved the problem! Take my advice, Scott, and cast an absentee ballot (as I routinely do), and you'll never have to worry about exit polls tarnishing your vote. There is no reason to convene a Pew seminar to examine the ethics of releasing exit-poll data or to draft exit-poll legislation.
Even if exit polls are a civic menace, there are other, less drastic remedies than curbing the press. Internet voting would make exit interviews next to impossible, and uniform poll closings would similarly undermine the projectionists.
Before I sign off, let me take a few friendly swipes at your arguments.
1) You write that you're a minimalist on civic journalism: that it's not the media's job to boost turnout but that one of its jobs is not to depress it. Want to rethink that one? If, as you posit, the test case is Weimar Republic election of 1933, wouldn't it have been a good idea to suppress the Nazi turnout? I don't think journalists should worry about whether their stories encourage or discourage voters. That's the job of the League of Women Voters, who preach the neutral, every-vote-is-worth-casting religion. Likewise, it's the job of advocacy groups and the political parties to encourage turnout on one side and depress it on the other. Our job is to tell the truth. And if the truth hurts, society should find ways to accommodate us. (See my policy recommendations above.)
2) I'm amused by your discussion of "Gore voters" who stayed home. If they're Gore voters who didn't vote, then I'm a ballerina who doesn't dance.
3) Please don't be sore if I don't take the bait about revealing military positions or clues that might solve a murder. Those are really rotten analogies because they're about life or death. You can do better.
Lastly, let's not forget that we're discussing two issues: 1) Should exit polls be released before the polls close? I say that's up to the creators of the polls: If they don't want people gossiping about the numbers they should keep their secret secret. If I were running the exit polls, I'd run them when I got them. 2) Should news organizations project the winners of contests before all the polls close? Once again, it's up to the news organization, and once again, I'd project the winners as soon as I could do so confidently.