Exit Poll

views updated May 29 2018

Exit Poll

PROBLEMS WITH U.S. PRESIDENTIAL ELECTIONS

Exit polls are surveys based upon voter interviews immediately after they have finished voting. The exit polls play a significant part in media projecting election winners and their margin of victory. The polls are used mainly in major elections, usually concerning national or state candidates. Exit polls are part of the pervasive status of surveys and polls in American and other societies. Exit polls have been used in elections throughout the world, such as in 2005 in the Belarus and Ukraine and 2006 in Mexico. Media organizations purposely utilize exit polls to understand voters’ choices and explain issue, partisan, and demographic differences in electoral outcomes. This information serves as a means of validating elections.

Exit polls are highly accurate because they remove main elements of potential survey research error problems by accounting for who actually votes and avoiding undecided voters. Problems for exit polls include a sample estimate of voter turnout, which may be incorrect for one or more candidates. Further, these polls do not consider absentee voting and other forms of early voting that do not require individuals to be at a polling place.

HOW CONDUCTED

An exit poll is conducted by selecting sample precincts in a state or other electoral area based upon historical voting data to judge outcomes of voting data. In each of the selected precincts, an interviewer requests a voluntary interview from a systematically selected voter; for example, every third or every tenth voter. Each voter completes a thirty- to forty-item questionnaire, which is placed in a box.

In each election cycle new questions are developed. The information is collected, tabulated, and transmitted to a central facility that tabulates a state or other electoral area total. In the 2004 U.S. presidential election nearly 150,000 people in nearly 1,500 precincts were interviewed in every state except Oregon. In 2004 the National Election Pool (NEP) also conducted a national sample of polling places to supplement exit poll data.

Prior to the 1990 U.S. midterm elections, each major network—ABC, CBS, NBC, CNN, and the Associated Press—conducted separate exit polls. For financial reasons, they created a consortium (also joined by Fox News) to conduct a single national exit poll, called News Election Service, and later called Voter News Service. After problems encountered with the 2000 presidential election, the 2004 presidential election exit poll was conducted through NEP by Edison Media Research/Mitofsky International. Exit polling results are delivered in three different times during election day, including mid-afternoon on that day. Multiple national exit polls can have different results based upon different sampling frames and different polling places. In other nations, major media organizations sponsor similar survey organizations to conduct exit polls. Media networks use exit poll data and actual vote returns from some exit poll sample precincts to project electoral victors. To project the outcome, exit poll findings must exceed the margin of sampling error in the sample design, generally plus or minus 4 percent at a 95 percent confidence level for state polls, and plus or minus 3 percent at the same confidence level for the national exit poll. Data within the sample design parameters is “too close to call.”

PROBLEMS WITH U.S. PRESIDENTIAL ELECTIONS

High profile exit polls in U.S. presidential elections have encountered difficulties in analysis and reporting of information, especially in the 1980, 2000, and 2004 elections. In 1980 television networks correctly predicted Ronald Reagan’s victory over President Jimmy Carter prior to many election polls remaining open in the midwestern, mountain, and western states. Individuals argued that this projection diminished those individual’s willingness to vote. In 1985 Congressional hearings—U.S. House Task Force on Elections—were held and resulted in network executives voluntarily promising to not release exit poll or election projections for a state prior to its poll closing time.

In 2000 the winner of Florida’s twenty-five electoral votes would obtain the majority of national electoral votes and be the victor. On several occasions networks changed their exit poll projection for this state. First, they projected Democratic vice president Al Gore, then retracted their projection as actual vote return data arrived. At approximately 2:00 a.m. Eastern Standard Time (EST) they projected Florida for Republican George Bush, and about an hour later retracted it. Bush won Florida by 527 votes after a U.S. Supreme Court decision. The Florida vote was “too close to call” through exit polls, and individuals incorrectly reported their vote based upon the “butterfly ballot” or other difficulties, such as hanging chads.

In the 2004 Presidential election, there was confusion and controversy concerning reliance upon exit polls. As late as 9:15 p.m. EST, NEP indicated that Senator John Kerry was leading President Bush, 51 percent to 48 percent. The final outcome was President Bush 51 percent and Kerry 48 percent. It has not been clearly determined why the exit polls overstated the Kerry vote in several states.

SEE ALSO Elections; Polling; Survey; Voting