Should Exit Polls Be Released Before Voting Ends?

Scott Shuger is a Slate senior writer and author of “Today’s Papers.” Jack Shafer is Slate’s deputy editor, “Press Box” columnist, and resident embargo-breaker (click here to catch up on the controversy and to read Shafer’s op-ed that ran in the Wall Street Journal).

Hi again, Jack:

I have not read the Plissner book, but I’m familiar with the California-in-1980 scenario, and all I can say is that CBS didn’t contact me. Because I was then a California voter and upon arriving at my polling place and seeing a long, slow line and having just heard that Reagan was the projected winner, I moved to the next thing on my to-do list. Now, granted, if I’d been more committed to other state and local races (partial excuse: I had just come back from being out of the country for the entire year and hadn’t been able to learn about them), or if I’d been a better game theorist, or if I’d simply gone on to my next errand and then come back when no doubt the line would have thinned because other people would have left because they were probably going to reason like I actually did, then I still would have voted. But instead I did the human thing—I reasoned to a predictable level and no further. The point is that our reasoning about voting and the press should take account of what, being human, people are likely to do and how they’re likely to feel. Hey, people don’t turn out to vote as much when it rains either. Hmmm, war and peace and jobs and education on the one hand vs. wet shoes on the other. That makes utterly no sense, but if there was some sort of cloud-seeding program going on, I’d still tell the seeders to skip Election Day.

It’s plausible that my 1980 experience was generalized, and it’s incumbent upon your position to show that it was not. But in all your writing on this topic up through Plissner, I haven’t seen a detailed reference to a definitive empirical study of the relationship between election projections, particular outcomes, and turnout. I don’t know if there is one. If there is, let’s get it out there and if there isn’t, there should be. If such a study were to show that the projection effect is negligible, then I would accept both the early release of exit polls and the networks’ publicizing of them. If not, then not. In short, I think this is an empirical issue and (until now at least) you seem to think it’s a priori. And because voting is a messy affair tapping into all sorts of motivations, perceptions, and misperceptions, I don’t think there are too many a priori arguments about it that are worth much.

You’re right to point out that there are some alternatives to the current scheme of things that would moot this controversy: Everybody could move to the East Coast; everybody could vote early; everybody could vote by absentee ballot; and, eventually, everybody could vote online. Regarding the most plausible of these, I’m not sure I’m thrilled about eliminating voting’s public profile. There’s something reassuring about seeing our schools, churches, and libraries periodically filled with voters. On most days in most courtrooms, there are no spectators, but it seems important to our legal system that there could be. But this worry isn’t a deal-breaker, and so I suppose I could endorse all-online voting. But that’s many years off, don’t you think? (I’d also endorse just banning exit polls and indeed banning polls for a while immediately before an election, as is done in some European countries.) And so, until the electorate can be convinced to vote absentee en masse, that leaves us with the current problem.

To address your other comments: 1) My position is that, at our level of democracy, I think the press should abstain from suppressing turnout but isn’t required to boost it, although in more dire circumstances (as in 1933 Germany) it would be. 2) You seem to imply that a Gore voter is a Gore voter only if he actually votes for Gore. That’s like saying only burning matches are a fire hazard. Democracy and the press, like fire safety, need a more inclusive sense of possibility than that. 3) You say that my prior examples only show that the press ought to suppress the truth where lives are at stake. Well, first off, in my example where the press should not publish a fact about a murder, if the state doesn’t have the death penalty this is an example of justified suppression where no life is at stake—suppressing will not revive the victim. Another similar example: In the case of a plane crash, newspapers routinely withhold publication of the names of the victims until their next of kin can be notified, even though publishing them would not cause any deaths. Or are you against this practice too? And finally and most important of all: In every presidential election, lives are at stake, many more than in a murder case or plane crash.