Once again, television networks are looking to Election Night with some apprehension, trying to avoid a repeat of the past two presidential cycles when initial exit polls suggested victories for John Kerry and Al Gore.
Net execs say this time around they’re deploying a series of measures to prevent the early leak of polling information, as well as to make adjustments for skewed sampling of respondents as they leave the polling booth. With at least a dozen battleground states where votes for the two candidates are expected to be close, networks will have to resist the competitive temptation to project winners based on exit polls, said ABC News spokesman Jeffrey Schneider.
“In projecting races in those states, our desire is to be right,” Schneider said. “If that means being more thoughtful and steady, that’s fine.”
In 2000, exit poll data led to the networks calling Florida for Gore, only to pull back when the actual vote count came in. Of course, they also had to retract their later call for George W. Bush after the hair-thin results threw the entire election in dispute.
In the 2004 election, initial exit polls gave flawed and incomplete information about voting in dozens of states. When the data leaked over the Internet, many assumed that Kerry was on his way to a resounding victory, to the point where even his campaign manager said to him, “Can I be the first to call you Mr. President?”
Exit polls are “really just one ingredient” of Election Night coverage, Schneider added.
Newsies say they have refined their exit polling methods since 2004, when a post-election investigation concluded that the skewed projections were likely due to sampling in some precincts where Kerry voters “were participating in the exit polls at a higher rate than Bush voters.”
Somerville, N.J.-based Edison Media Research, which has been hired by a consortium of NBC, ABC, CBS, Fox, CNN and the Associated Press to conduct the exit polls, will conduct its analysis in a closed-off room to prevent information leaks. The system seemed to work well in the 2006 midterm elections.
“They’re a survey, and like any other survey there is sampling error, which we try to minimize,” said Joe Lenski, co-founder and exec VP of Edison Media Research.
Exit poll results are just one of three information bases that EMR will provide to the news orgs. “We have actual vote-count reporters at the precinct level and the AP reports on vote counts at the county level,” Lenski said. All three sets of results are fed to subscribers.
“All the networks use the exit polling with actual vote counts and the historic voting trends as well as the latest pre-election polling in their model for projecting,” said one news exec. “But if the model doesn’t produce a very clear and obvious outcome, you don’t call the winner.”
Some people don’t answer truthfully when asked whom they voted for, and sometimes pollsters inadvertently talk mostly to supporters of one candidate.
“The lessons of ‘04 that have been applied to exit polls since then are that interviewers need to have better training, that you need to be cautious in looking at the results,” CBS News polling director Kathy Frankovic told Politico.com.
Exit polling is most useful, Lenski said, in determining voter demographics in retrospect. “They’re very useful, for example, in the primaries because they tell you what income bracket, what sex, basically what kind of person voted for whom. There’s really no other way you can get that information except by exit polling.”
Some critics of exit polls have decried media use of them as an impediment to voter turn-out — particularly if it looks to be a landslide at the top of the ticket for one candidate or another. When networks are reporting that one particular candidate is ahead, voters of the other candidate may not bother to vote.
But Jack Shafer, media critic for the online magazine Slate, has rejected that claim, arguing that evidence of vote suppression or depression by exit poll reporting is far from conclusive. Shafer has even called for networks to disregard the embargo on reporting exit poll results (usually until voting stations close) because the numbers constitute news, if not a reliable indicator of a winner.
“They help the networks make good guesses, and (the networks’) track record is good, except when it isn’t,” Shafer said.