Ever wonder about those exit polls they talk about on election night and how they can be so wrong?
The election returns tell us who or what garnered the most votes in which races. Exit polls attempt to explain the results. Sometimes, the explanation can run counter to our expectations. In 1998, exit polls showed Nevada’s Harry Reid doing suprisingly poorly among women voters. This year, they showed George Bush doing surprisngly well among abortion supporters.
I helped conduct exit polls for Edison Media Research and Mitofsky International this year.
What are “exit polls"? It’s pretty simple. An exit poll is a questionnaire completed by voters outside randomly selected voting places.
As voters leave their polling places, they are asked to complete a questionnaire with information about their candidate selections and their opinions on different issues. They then fold their survey and place it in a white, cardboard box with media-outlet logos printed on the side.
The exit poll I worked on consisted of 28 multiple choice questions and took an average of two minutes to complete. The questions ranged from who the voter supported to what was the last grade of school completed by the voter to subjective questions, like “Is the country more or less safe from terrorism compared to four years ago?” or “How do you feel about the U.S. decision to go to war with Iraq?”
Throughout Election Day, the polling companies sold the collected information to major media outlets and other interested parties. The news folks analyzed the data and projected election results. The Edison/Mitofsky exit poll conducted in Washoe County provided information to ABC, the Associated Press, CBS, CNN, FOX and NBC, among others.
Interviewers approach voters according to an interviewing rate. For example, an interviewer might be instructed to approach every third voter as she or he exits the polling location. The rate ensures that the desired number of interviews is completed over the course of the day.
If the designated voter refuses or is missed for whatever reason by the pollster, pollsters start counting again, starting with the next person.
The pollster’s day is divided into three call periods in which tallies are relayed back to the company. Refusals aren’t missed, either. A “refusals and misses” code sheet is color coded for each call period. The interviewers must guess sex, age and race—often simplified to black or non-black—and relay the totals back to the company. The total number of votes cast at the polling place is also requested, even though it is possible that the polling place official will be unable or unwilling to share that information.
Exit polling is subject to state, county or municipal regulation. Laws vary, but in general, they either limit proximity to the voting booth or the center itself, or prohibit interviewers from making statements that could influence a voter’s selections.
Surveying voters after they have voted is permissible and is a protected part of free speech. Nonetheless, a legal team is available on Election Day to assist interviewers with unforeseen issues, for instance, if an interviewer is asked to move to a distance where it is impossible to talk to voters. That legal assistance may have been prudent this year in Nevada, since Secretary of State Dean Heller’s office issued an Oct. 19 warning to Associated Press, ABC News, CBS News, CNN, NBC News and Fox News that the requirement to keep at least 100 feet from the entrance to a polling site can be applied broadly and not just to anyone who’s working on behalf of a candidate. Media entities disputed Heller’s interpretation of the law.
The message from the secretary of state wasn’t included in the Nevada package given to exit pollsters, but we were given a troubleshooting sheet that said in bold print, “Electioneering laws do not apply to you.” We were told that electioneering laws pertain only to attempts at influencing a person’s vote on behalf of a candidate or issue before that vote is cast, not interviewing voters after they have finished voting, and the interviewer can’t be working for any candidate or political party.
Even the physical location of pollsters can be considered a factor that can influence a voter’s thinking. Standing with people campaigning for or representing candidates can be a violation of electioneering laws. Pollsters are advised not to show preference for any candidate or party by doing such things as wearing a political button.
County registrar of voters Dan Burk told the Reno Gazette-Journal that not only would the sheriff and local police departments be on hand to deal with saboteurs, so would the federal Homeland Security Department.
State law prohibits anyone “to solicit a vote or speak to a voter on the subject of marking his ballot” within 100 feet of polling places. The penalty is a civil fine, punishable by up to $20,000 per incident.
Leaking exit poll information to friends or colleagues ahead of schedule is a sin in the world of exit polls, and incomplete information can distort the picture that the exit poll was intended to clarify.
For example, mainstream news organizations followed exit poll data to wrongly proclaim Vice President Al Gore president-elect in the 2000 election. Networks and cable news outlets were more conservative with their projections this time around.
While exit polls are indicators of many things, they shouldn’t be used to predict outcomes. This year, pollsters blamed bloggers for spreading news that gave a misleading view of candidate Sen. John Kerry in the lead.
The inaccuracy of the apparent results will now be argued, including theories that Kerry supporters went to the polls earlier than Bush voters, or Republicans were reluctant to talk to pollsters.
Even when used properly, exit polls have their pitfalls. They are no substitute for real votes. In the 1992 New Hampshire presidential primary, Pat Buchanan gained massive publicity and a huge boost to his campaign when exit polls showed Vice President George Bush (the Elder) winning the primary by only a narrow margin over Buchanan. The front-page headline in the New York Times was “New Hampshire rebukes Bush.”
But actual late-night vote counts showed Buchanan lost to Bush in a huge landslide of 16 percentage points. Few media entities bothered correcting the record, and those that did downplayed it (the Times ran an eight-sentence corrective story at the bottom of page A13). For the rest of the campaign, Buchanan, on the “strength” of his New Hampshire showing, was treated as a major contender.
At one Washoe County polling place this year, a man wore a Josef Stalin T-shirt that seemed to illustrate the risks of exit polling: "Those who cast the votes decide nothing; those who count the votes decide everything."