This article is based on the introduction to Chapter 4 ("Biased polls or biased count?") of Was the 2004 Presidential Election Stolen? Exit Polls, Election Fraud, and the Official Count by Steven F. Freeman and Joel Bleifuss (New York: Seven Stories Press, 2006). It is used here with permission. .
What is an exit poll?
Exit polls are questionnaires administered to voters upon leaving the polling place to either help news organization prepare their election news coverage (q.v. National Election Pool) and/or to help determine whether votes are being counted as they are cast (q.v. Election Verification Exit Poll). 
How are exit polls conducted?
On Election Day, one or more interviewers report to selected precincts. From the time the polls open in the morning until the polls close, interviewers select exiting voters at consistently spaced intervals (for example, every third or fifth voter). These voters are given brief instructions to fill out a short questionnaire, and asked to deposit it in a container similar to a ballot box. 
Occasionally, EVEPs attempt to poll every voter so as to rigorously attempt to verify results beyond any doubt. EI has obtained response rates as high as 91% in some cases.
Exit poll accuracy
Properly conducted, exit polls mirror actual Election Day voting with a high degree of reliability. Around the world, exit polls have been used to verify the integrity of elections. The United States has funded exit polls in Eastern Europe to detect fraud. Discrepancies between exit polls and the official vote count have been used to successfully overturn election results in Ukraine, Serbia, the Republic of Georgia and Peru. 
In nations with transparent voting practices (e.g., votes cast on paper ballot and counted in full public view), election results predict official results with remarkable accuracy. For example, in Germany, as soon as the polls close, polling agencies release prognoses that have proven highly reliable. In the three most recent national elections there, poll percentages diverged from official counts by an average of only 0.26%. They have been almost as accurate for the German vote in the European Parliament Elections (table 2), averaging 0.44% differential from tallied results over the past three elections. [1: Appendix A]
Until the 2000 election and the introduction of electronic voting, the only worry was that they were too accurate! Political scientists George Edwards and Stephen Wayne wrote:“The problems with exit polls lie in their accuracy (rather than inaccuracy). They give the press access to predict the outcome before the elections have been concluded.” 
Why Exit Polls are so accurate
Properly conducted, exit polls should mirror actual Election Day voting with a high degree of reliability. Around the world, exit polls have been used to verify the integrity of elections. The United States has funded exit polls in Eastern Europe to detect fraud. Discrepancies between exit polls and the official vote count have been used to successfully overturn election results in Ukraine, Serbia, the Republic of Georgia and Peru. In nations with transparent voting practices (e.g., votes cast on paper ballot and counted in full public view), election results generally predict official results accurately. 
Exit polls eliminate most sources of general polling error. Unlike telephone opinion polls that ask people which candidate they intend to vote for several days before the election, exit polls are surveys of voters conducted after they have cast their votes at their polling places. In other words, rather than a prediction of a hypothetical future action, they constitute a record of an action that was just completed. Exit polls will not be skewed by the fact that some groups of people tend not to be home in the evening or don’t own a landline telephone. Exit polls are not confounded by speculation about who will actually show up to vote, or by voters who decide to change their mind in the final moments. Rather, they identify the entire voting population in representative precincts and survey respondents immediately upon leaving the polling place about their votes. Moreover, exit polls can obtain very large samples in a cost-effective manner, thus providing even greater degrees of reliability. 
The difference between conducting a pre-election telephone poll and conducting an Election Day exit poll is like the difference between predicting snowfall in a region several days in advance of a snowstorm and estimating the region’s overall snowfall based on observed measures taken at representative sites. In the first case, one is forced to predict future performance on present indicators, to rely on ambiguous historical data, and to make many assumptions about what may happen. In the latter, you simply need to choose your representative sites well. So long as your methodology is good and you read your measures correctly, your results will be accurate. 
Who conducts exit polls?
In most countries, competing media and political groups fund several independent exit polls. As noted above, the U.S. as well as other countries and independent agencies have funded Election Verification Exit Polls.
In the US, all major media firms act as a single block. Exit polls for federal elections are conducted by the National Election Pool (NEP), a consortium of six major news organizations (ABC News, the Associated Press, CBS News, CNN, FOX News and NBC News. NEP exit polls, and their predecessors, however, have been widely criticized for wide disparities with official numbers and, simultaneously, for leaking "uncorrected" results and for covering up those disparities. (NEP "corrects" their survey data so as to make their numbers conform with the count -- which, of course, undermines completely its potential as an election verification tool. The handful of people who now see actual survey results are, literally, quarantined and sworn to secrecy.  Until recently, many independent newspapers and local broadcast stations used to conduct limited exit polls, but none that we know of do so any more.
Election Verification Exit Polls
Election Integrity launched the first Election Verification Exit Poll in 2006 and did so nationally in 2008. Two factors discouraged continued use:
- With the election of Barack Obama in 2008, funding for these EVEP exit polls dried up.
- With widespread early voting and vote-by-mail, the efficacy of EVEPs as tools for verifying results greatly diminished.
Articles and resources
 Freeman, Steven F. and Joel Bleifuss, Was the 2004 Presidential Election Stolen? Exit Polls, Election Fraud, and the Official Count, (New York: Seven Stories Press, 2006) Chapter 4. Biased polls or biased count? Election Integrity's page on Exit Polls.
Steven F. Freeman and Joel Bleifuss Was the 2004 Presidential Election Stolen? Exit Polls, Election Fraud, and the Official Count (New York: Seven Stories Press, 2006)
Ken Warren In Defense of Public Opinion Polling (Cambridge, Mass:Westview Press, 2003).
Edison Media Research and Mitofsky International National Elections Pool: http://www.exit-poll.net/faq.html
Election Integrity: http://www.electionintegrity.org/
Freeman, Steven F., Presentation to the American Statistical Association, Philadelphia, October 14, 2005
NEP Exit poll data is stored and made available (without precinct or county identification) through the University of Connecticut’s Roper Center a major center for polling data and reports of public opinion. http://www.ropercenter.uconn.edu (This same exit-poll data is also available through Survey Research Center at the University of Michigan’s Institute for Social Research. http://www.isr.umich.edu/src) Election Integrity: http://www.electionintegrity.org/ makes primary data available from its Election Verification Exit Polls