Exit pollsters today will be closer to voters — and likely closer to final results

The exit pollsters who feed estimates of election results to major media outlets literally will be closer to Minnesota voters today than they have been for at least two decades.

In October, a federal judge barred the state from enforcing a Minnesota law requiring exit poll interviewers to stand at least 100 feet away from buildings where votes are being cast.

Does that mean someone will peer over your shoulder while you mark your ballot?

Nope. The state now maintains that the interviewers cannot be in the room where people actually vote, said Beth Fraser, director of governmental affairs for the secretary of state’s office.

In other words, an exit interviewer could catch you just outside the door of the voting room and ask you to fill out a short, anonymous questionnaire about your vote.

As a practical matter, relatively few Minnesota voters are likely to notice any difference today as a result of the court tussle between the state and the national news organizations that commission exit polling.

National sample will include 45 Minnesota precincts
Survey researchers for the National Election Pool have randomly selected slightly more than 1,300 precincts nationwide for the polling. In Minnesota, that will include 45 of the state’s 4,100 precincts, according to a document in the court case (PDF).

What’s more, only a small fraction of the voters at a chosen precinct will be asked to fill out a survey form. In a second layer of random selection, pollsters typically approach every “nth” voter where “n” might be 20 or 50 or whatever number fits the survey’s statistical model.

If there is any real difference for most voters, it would come in increased accuracy of the information the exit interviewers gather. The news organizations that challenged Minnesota’s 100-foot rule claimed that in 2004 and 2006, error rates could have been cut in half if interviewers had been allowed to stand within 25 feet of the polling place to randomly select the voters who were invited to complete exit questionnaires.

Accurately projecting the winners is a big deal for those of us who can’t wait to see the outcome of today’s historic election.

Here’s roughly what will happen.

At 4 p.m. Central time, the first batch of exit poll results will be distributed to news organizations that participate in the National Election Pool: ABC, CBS, CNN, Fox, NBC and the Associated Press. They have mutually agreed not to release or publish any data that would characterize the outcome of a race in a given state until all of the polls have closed in that state.

Meanwhile, the exit pollsters also will watch tabulated votes as they are available from precincts, counties and states, said Rob Daves of Daves & Associates Research in Minneapolis. He is a past president of the American Association for Public Opinion Research and former director of the Minnesota Poll.

As the evening goes along and more actual vote counts become available, the exit data will be weighted to fit the actual results, bringing the tallies closer and closer together, Daves said.

By about 8 p.m., the action shifts to newsrooms where editors and their hired experts will have seen enough data to begin deciding how to call a given contest. Any difference in how and when one network or another calls a close race is the result of that internal decision-making process, not the polling. They’re all working from the same data, Daves said.

Even with poll  improvements, newsrooms still wary of pitfalls
Even with Minnesota’s 100-foot rule set aside for the day, newsrooms must watch out for pitfalls in the polling.

In 2004, early national exit poll results had Sen. John Kerry beating President Bush. We now know, of course, that was not the true outcome. It turned out that Kerry supporters were more likely to vote early and to agree to participate in the poll.

Each election is different, and this year could be strikingly different. But in general, highly educated voters are more likely to agree to fill out an exit poll questionnaire.

So are voters who were actively involved in campaigns: the donors, door-knockers and phone callers. To the extent that one side had more volunteers than the other, that could skew the exit poll findings.

A Fox News/Opinion Dynamics poll (PDF) conducted last week suggests that supporters for Democratic Sen. Barack Obama will be more willing to fill out exit poll questionnaires today than those who plan to vote for Republican Sen. John McCain.

Seventy-one percent who favored Obama said they would agree to spend 10 minutes filling out an exit poll questionnaire. Just 64 percent of McCain supporters said they were likely to do so.

The greatest challenge this year will be polling the early voters, who could account for as much as 30 percent of the ballots cast in the nationwide election, especially in Colorado, Oregon and Washington.

Pollsters have tried to capture those voters in telephone surveys, a leader of the company that conducts the exit polling told the Pew Research Center.

“In the national survey, and in 15 other state surveys, we’re doing telephone interviews the week before the election of people who have already voted or are telling us they are definitely going to vote before Election Day,” Joe Lenski, co-founder and executive vice president of Edison Media Research, said in the Pew interview.

The news organizations in the National Election Pool hire Edison to do the research in partnership with Mitofsky International.

Beyond Election Day use, polls have long-term research value
Calling the winners is only the first use of the information gathered. More valuable is the long-term analysis scholars and researchers will make of the data for years to come, said Daves, the polling expert in Minneapolis.

When you cast your vote, the only information you leave about yourself is your candidate preference. At the level of your precinct, political scientists can add a fair amount of context from census data and past voting trends.

For example, they’ve known that certain precincts in traditional Republican strongholds of Edina and Minnetonka have tilted more toward Democrats recently. And they can tap census reports for corresponding income and ethnic trends in those precincts.

But that still doesn’t tell anyone how women compared with men in their ballot choices, or how young people turned out and voted.

Exit polls are the key source of those finer demographic points. As such, they are valuable to political scientists, parties, interest groups and organizations representing all kinds of voter blocks. For example, the group Women’s Voices Women Vote said on its website that it expects exit poll results for unmarried women voters this year to reveal a wealth of information about political preferences as they relate to economics, age and other factors.

“The only way we know the demographics of the electorate and the reasons why the voters did what they did is because of exit polls,” Daves said. “So when people choose not to participate, they are not making their voices heard.”

Meanwhile, exit poll interviewers are not the only people who may be curious about your vote today. News reporters and other observers no doubt will hover around many precincts hoping to catch Election Day anecdotes and color, if not scientific assessments of the vote.

The secretary of state’s office maintains that the 100-foot rule still applies to those who are not working for the six national media organizations that won the preliminary injunction in federal court last month, said Fraser, the office’s director of governmental affairs.

Minnesota law permits journalists to enter polling places to observe voting. But they’re not supposed to speak to voters until they are 100 feet from the polling place. Same goes for die-hard candidates and campaign workers.

We’ll have to wait for the next legal steps to see how these rules stand over time.

U.S. Chief Judge Michael Davis’ preliminary injunction applies to today only. The court has yet to decide the underlying lawsuit in which the six national news organizations argue that the 100-foot restriction infringes on their First Amendment rights.

And then there’s the prospect that the Minnesota Legislature could revisit the whole question of who can stand where on Election Day.

Sharon Schmickle writes about national and foreign affairs and science. She can be reached at sschmickle [at] minnpost [dot] com.

Comments (1)

  1. Submitted by John Olson on 11/04/2008 - 07:13 am.

    CNN reporting that Dixville Notch in NH votes 15 to 6 in favor of Obama. First time since 1968 that they have voted for a Democrat.

    It will be a long, historic day. My elderly parents (safely sequestered in another state) have both said that this election reminds them a lot of the 1960 election of JFK. One can only imagine what that campaign might have been like if there had been cable news and internet going 24/7 at the time.

Leave a Reply