Exit Polls

12 Key excerpts on "Exit Polls"

• 

  eBook - PDF

  The SAGE Handbook of Public Opinion Research

  • Wolfgang Donsbach, Michael W Traugott, Wolfgang Donsbach, Michael W Traugott(Authors)
  • 2007(Publication Date)
  • SAGE Publications Ltd

    (Publisher)

  THE IMPORTANCE OF Exit Polls Exit Polls are polls of voters , interviewed after they have voted, and no later than Election Day. They may include the interviewing before Election Day of postal, absentee and other early voters. Exit Polls are now conducted in many countries, and they may not always be conducted at the polling place, although most are. There are differences in Exit Polls. Some may be conducted to predict election results and others may be used for later, academic and analytic purposes. Exit Polls can serve three different func-tions that are not mutually exclusive: pre-dicting election results, describing patterns of voter support for parties, candidates, and issue; and supporting extensive academic research efforts. The main difference between the analysis categories is the speed with which the results are formulated and disseminated. Exit Polls have become part of the ordinary political discourse. Perhaps the first was conducted inadvertently in the United States in 1964 by Ruth Clark. Clark was a well-known newspaper researcher, who began her research career as an interviewer. In 1964, she worked for Lou Harris, and was sent to conduct interviews in Maryland on its primary election day. Tired of door-to-door interviewing looking for voters, she decided to talk with them as they left the polling place. As she put it, ‘I told Lou what I had done, and by the [Republican] California primary in June, the exit poll was put to full use,’ with Barry Goldwater voters dropping blue beans into a jar, while Nelson Rockefeller voters dropped red beans (Rosenthal, 1998, p. 41). 2 Pollster Lou Harris and statistician David Neft, hired as consultants for CBS News at the start of 1964, used a set of ‘key precincts’ to project election outcomes. While Ruth Clark may have invented the exit poll, it became a staple of news election Exit Polls AND PRE-ELECTION POLLS 575 coverage in 1970s and 1980s.

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Political Campaign Communication

  Inside and Out

  • Larry Powell, Joseph Cowart(Authors)
  • 2017(Publication Date)
  • Routledge

    (Publisher)

  Vulnerability surveys are similar to benchmark surveys in terms of length, but differ in intent. This questionnaire is designed to measure the likelihood that an incumbent can be defeated by a strong challenger. If the vulnerability survey indicates that the challenger has a chance, it can help them develop a strategy for the campaign and assist in their fund-raising efforts. Sometimes, the vulnerability survey is done on behalf of the incumbent with the intent of identifying potential points of weaknesses; in those cases, the incumbent typically uses the information to reduce those weaknesses before a challenger has an opportunity to take advantage of them.

  Exit Polling

  One popular form of polling for media outlets is exit polling. Exit Polls are on-site surveys conducted on the day of the election with voters as they leave the voting booth. Their purpose is to identify for whom they voted, and what factors influenced their voting decision. For campaign purposes, this is relatively useless information; the campaign is over, and it is too late to use the information for strategic adjustments. However, these data can be useful to media outlets by allowing them to project winners before the polls have closed and to provide some type of insight as to why one candidate won and another lost.

  Exit polling was first developed by Warren J. Mitofsky for CBS News in the early 1970s and first used for election-night coverage of the 1972 election (Clymer, 2006). He then worked with the New York Times to design a broader method for the 1976 election. Since then, the national media have paid for a joint exit-poll system to aid them in election-night coverage of national elections.

  Exit polling is not a foolproof system. It is subject to the same margin of error (usually plus or minus five percentage points) that limits regular polling approaches. Furthermore, since it is physically impossible to have interviewers at every voting site, the validity of the system depends on the representativeness of those voting precincts that are selected for interviews, which introduces a further possible margin of error. Those precincts or election districts are often chosen on the basis of turnout in previous elections, without considering whether the nature of the precinct has changed since that time. Even if representative precincts are chosen, if those precincts do not have voter turnout rates consistent with the election as a whole, its results may be invalid. Given all of these problems, it is really remarkable that exit polling is as effective as it is. Most of the time, the national networks get it right, and correctly project probable winners. Still, they occasionally miss. In 1984, for example, the national networks projected Senator Jeremiah Denton (Rep.-Ala.) to win reelection; when the votes were finally counted, Democratic challenger Richard Shelby narrowly defeated him. In 1996, incumbent Senator Bob Smith (Rep.-N.H.) was declared the loser, but he narrowly won reelection. In the 2000 election, exit-poll projections for Florida led to a series of wrong predictions by the television networks. Early on election night, all of the major networks declared Al Gore the winner; they subsequently rescinded that and later credited the state to George W. Bush’s win column. Eventually, they withdrew both predictions; the vote count was so close that it took weeks to decide a winner.

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Polls and Politics

  The Dilemmas of Democracy

  • Michael A. Genovese, Matthew J. Streb, Michael A. Genovese, Matthew J. Streb(Authors)
  • 2012(Publication Date)
  • SUNY Press

    (Publisher)

  Out of this volatile mix of media and politics has grown the highly controversial offshoot of scientific polling called the “exit poll,” which collects information gathered from large samples of citizens as they leave their polling places. Exit Polls ask, first and foremost, who the voters have cast their ballots for, and also some information about the voters themselves to enable analysts to get at class, race, and religious cleavages that may be at work in the election. They typically also include a few questions about perceptions of the candidates’ qualities and the voter’s positions on some of the issues that were prominent in the campaign. To meet their needs the media invented these Exit Polls, which have now become fixtures of national, and many state, elections. The Exit Polls represent quite a substantial investment by the media. In 2000, the Voter News Service, which is the major exit poll organization and about which I will say more below, had a budget of $35 million. With this budget, they harvest information that is to serve the two principal goals of election night coverage: to tell us who is win- ning and to explain the voters’ decisions. For this investment, one would think that smart businesspeople like those running the network news organizations would see a favorable return in terms of a higher level of precision and depth of explanations of election night reporting. Instead, for many postelection commentators, the chief result of exit polling has been to debase democracy and to show the recklessness of media in their pursuit of ratings. Consider some of the language used to describe the media’s contri- bution to Election Night 2000.

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Encyclopedia of Survey Research Methods

  • Paul J. Lavrakas(Author)
  • 2008(Publication Date)
  • SAGE Publications, Inc

    (Publisher)

  Thus, despite what many people believe, the election projections announced on television in many cases are based very little, or even not at all, on the exit poll data. Although it is these projections of election winners that most people think of when they think of Exit Polls, the second use of exit poll data is arguably more influential on the geopolitical entity and the citizenry being measured. Thus, Exit Polls are important, not so much because they are used to help make the projections reported by the major television networks on Election Night, but because the information they gather about the voters’ demographics and attitudes toward the candidates and the campaign issues provides very powerful and important explanations about why the electorate voted as it did. It is only through the use of accurate exit poll data that the so-called mandate of the election can be measured and reported accurately without relying on the partisan spin that the candidates, their campaign staff, and political pundits typically try to put on the election outcome. For example, in 1980, Ronald Reagan’s strategists described his sound defeat of Jimmy Carter as a “turn to the right” by American voters and as an impetus for a conservative legislative agenda for the new Congress. In contrast, 1980 exit-poll data showed there was no ideological shift among American voters. Instead, they were primarily concerned about President Carter’s inability to influence the economy and settle the Iran hostage crisis, and they wanted a new president whom they hoped would do a better job in reducing inflation

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  The SAGE Handbook of Electoral Behaviour

  • Kai Arzheimer, Jocelyn Evans, Michael S. Lewis-Beck, Kai Arzheimer, Jocelyn Evans, Michael S. Lewis-Beck, Author(Authors)
  • 2017(Publication Date)
  • SAGE Publications Ltd

    (Publisher)

  Some problems can be inherent to a methodology: for example, the now widely used opt-in internet panels rely on the assumption that those who the polling company recruit to their panel are representative of the broader universe of voters. This assumption can fail in many ways: for example, panel volunteers will usually be more interested in politics, which may impact on their choices, and some groups with systematically different preferences may be under-represented in the panel. One British internet panel, analysing its ‘polling miss’ in 2015, found evidence of both – the young voters they managed to recruit were systematically different to those they did not, due to higher political interest, and they systematically under-represented the over-75s, who voted in a distinctive way on polling day in 2015 (Rivers and Wells, 2015). Biases can also emerge due to postsampling techniques, such as those used to identify ‘likely voters', which can produce an unstable sample (Erikson et al., 2004) or generate bias due to respondents’ excessively optimistic estimates of their chances of voting (Mellon and Prosser 2015). Respondents can also misreport vote intentions – one particularly notorious example being the ‘shy Tory’ phenomenon in the 1992 British election – many voters who ended up backing the governing Conservative party were apparently unwilling to admit their support during the campaign and responded ‘won't say’ when asked their vote choice (Jowell et al., 1993).1

  Polls after votes: forecasting outcomes from exit polls

  While pre-election polls measure voters’ future intentions, exit polls capture the decisions they have just made. The polls–vote relationship therefore becomes a ‘vote–vote’ relationship of sorts – between the reported vote choices of the exit poll sample and the vote choices made by the electorate overall. Exit Polls are frequently employed to inform live coverage and analysis of the election result as broadcasters wait for votes to be counted, and can be used to build forecast models to anticipate results due to be declared hours later. The highest profile British election ‘forecasting’ model is the one used to predict outcomes when the polls close at 10 p.m., which then drives election night coverage until the results begin to arrive several hours later (Kuha, 2015; Firth, 2011; Curtice et al. 2011; Curtice and Firth, 2008; Brown and Payne, 1984). These models employ information from specially designed Exit Polls to provide an initial estimate of the likely outcome, and then update themselves on a rolling basis with new information from individual seat results as these are announced.2

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Political Communication in Britain

  Campaigning, Media and Polling in the 2017 General Election

  • Dominic Wring, Roger Mortimore, Simon Atkinson, Dominic Wring, Roger Mortimore, Simon Atkinson(Authors)
  • 2018(Publication Date)
  • Palgrave Macmillan

    (Publisher)

  But their efforts are directed entirely towards predicting the number of seats that each party will win, and the poll’s design reflects this. The sample is not nationally representative: interviews are mostly concentrated in marginal constituencies (and as far as possible are conducted in the same polling districts as at the previous election), since this provides the most useful data for making the prediction. The interview consists solely of asking the participant to complete a duplicate ballot paper and place it in a ballot box, and is over in seconds: there are no demographic questions and none about election issues or personalities. This very short interview length probably helps to maintain the exit poll ’s high response rate, usually over 80% (a lengthy questionnaire might make it much harder for interviewers to persuade the randomly-selected voters to participate), contributing to its excellent record of accuracy in its predictions. 2 But it offers no evidence on who has voted which way, or why. In the past, there have occasionally also been analysis Exit Polls conducted at British elections, but the last was a number of years ago. Exit Polls are expensive exercises, and the media presumably feel that they would not make sufficient use of the data to justify the cost; academics, for whom depth of content in the data is more important than the speed with which it can be produced, find the British Election Study a better use of the available research funds. But for reliability in measuring the basics of electoral behaviour, Exit Polls have important advantages. An exit poll interviews people known to have voted, and does so within minutes of their vote

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Polling UnPacked

  The History, Uses and Abuses of Political Opinion Polls

  • Mark Pack(Author)
  • 2022(Publication Date)
  • Reaktion Books

    (Publisher)

  69

  As well as concentrating on changes, the exit poll also considers any systematic variation between different types of constituencies and produces different estimates of vote share changes in those different types. This methodology explains why the headline results read out just after the polls have closed at 10 p.m. on an election night are all about seat numbers, not national vote shares. The methodology is tuned to produce the former, not the latter. In that respect, the exit poll is the opposite of a regular voting intention poll.

  Exit Polls: The USA

  American Exit Polls deserve their own section because, frankly, they are not very good. In other countries where an election count’s accuracy or fairness is in doubt, it is common to see and use Exit Polls as a safeguard, giving numbers to test against the actual result. Yet in the United States, ahead of the 2020 presidential election, the blunt verdict on U.S. Exit Polls from a question-and-answer piece run by the Princeton Election Consortium was that they are not up to this task:

  M. D. writes: In third world countries, Exit Polls are compared to election results to see if there has been election fraud. Is there anything set up this time in America to compare results to how people thought they voted?

  Sam replies: Can’t exactly do that because Exit Polls are themselves not very well weighted, i.e. they get adjusted to reflect the outcome. Better to use opinion polls, though I think that is not standard practice.70

  This is not because Exit Polls are a new thing in America, without enough experience yet accumulated for them to do well. In fact, America was home to the first exit poll.

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  The Power of Polls?

  A Cross-National Experimental Analysis of the Effects of Campaign Polls

  • Jason Roy, Shane P. Singh, Patrick Fournier(Authors)
  • 2021(Publication Date)
  • Cambridge University Press

    (Publisher)

  This style of reporting limits coverage of more substantive issues, focusing instead on how well candidates or parties are doing: who is in front, who is behind, who is gaining momentum, and who is losing ground. Public opinion polls provide much of the fodder for this style of journalism, which has been remarkably intense during electoral periods— precisely the time when voters are called upon to weigh the political options available to them and to select their representatives. In addition, the rise of data journalism has contributed to the presence of polls in media reports during elections (Coddington 2015; Splendore et al. 2016). Altogether, there is little doubt that polls have become a dominant part of contemporary election campaigns. 1 With the term public opinion poll, we are referring to information (e.g., vote intentions, policy preferences, issue positions) drawn from a sample of a larger population of citizens. In what follows, our use of the term refers exclusively to a measure of public vote intentions in the lead-up to an election. 1 The Power of Polls? Given the prominence of election polls across democracies, their accur- acy warrants consideration. Unlike most attitudes and behaviours estimated with surveys, it is possible to compare campaign poll results to the thing they are meant to estimate: the distribution of party or candidate support in the voting population on election day. When election outcomes differ from pre- election poll projections, these polling errors reflect a lack of accuracy (Jennings and Wlezien 2018), and they receive considerable attention. For example, a straw poll by the Literary Digest magazine famously predicted the wrong winner in the 1936 American presidential election (Alf Landon rather than Franklin Roosevelt), while newcomer poll houses headed by Crossley, Gallup and Roper that employed more rigorous sampling methods made quite accurate predictions (Gosnell 1937).

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Opinion Polls and Volatile Electorates

  Problems and Issues in Polling European Societies

  • Matt Henn(Author)
  • 2019(Publication Date)
  • Routledge

    (Publisher)

  In the current era, where electorates are typically becoming increasingly volatile, instrumental, sensitive, and receptive to political news and events, the misuse of data in this way has serious implications for electoral politics. Given the prevalence of tactical voting in countries like Britain and elsewhere in Europe, the increasing voter awareness of and exposure to these poll reports may have an influence on people's voting behaviour.

  The often contradictory relationship between the pollsters and the mass media organisations who sponsor them, is summarised by Weiner (1977, p.675):

  The tension between academic survey techniques and the pressures of journalism has remained the prime characteristic of polling organisations. While the pollsters are indeed concerned with the 'scientific measurement' that their promotion suggests, their concepts and operations are constrained by journalistic conventions and resources. Polls must produce news - items of immediate interest such as explicit election forecasts - with limited time, staff, and money.

  While pollsters may attempt to cajole their clients into a more informed, professional and valid use of polls, the underlying needs of the media to maximise profits and their share of the audience often results in the use of polls as part of a strategy to provide entertainment, rather than to educate the public on political issues. As pollsters Webb and Wybrow (1974, p. 265) have observed:

  Our main sponsors, the newspapers and T.V. are continually pressuring us to forecast seats, and are likely to supply their own forecasts in the face of our resistance.

  Worcester reports for instance that at the 1992 General Election, the public "...were clearly misled by the seat projections done by the broadcasters and their psephologists, not the Exit Polls" (1992a, p.3). Thus, the Harris exit poll pointed to a Conservative majority of 12 seats (compared to an actual result of a 21 seat majority), ITN massaging of the data actually suggested that the Tones would be 21 seats short of a majority.

  To create a space between themselves and their media clients, the pollsters have combined to form various associations. The purpose of these associations is to preserve their integrity as independent research organisations, to maintain high standards in data collection and dissemination, and to set up a mechanism for self-regulation. At an international level, the World Association of Public Opinion Research (WAPOR) exists to "...establish a world-wide meeting ground for those working in the area of survey research" (cited in Worcester, 1991, p.207). Similar associations include ESOMAR (Europe) and in Britain there is APQPO.4

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Presidential Polls And The News Media

  • Paul J Lavrakas, Michael Traugott, Peter V Miller(Authors)
  • 2019(Publication Date)
  • Routledge

    (Publisher)

  The Times's experience in 1992 is that any program of opinion research must be designed so that it is responsive, flexible, and adequate to the task it faces. Because the election is a year-long process, it is necessary to design a 12-month (or longer) series of polls that can provide a setting to and offer perspective on the individual phases of the electoral cycle. Because public reactions to personalities, issues, and events can change during the year, it is important to repeat a core of trend questions time after time in order to capture the dynamics of the campaign. Because long-term, middle-run, and short-lived forces can each affect the outcome, it is important to build in devices to study all three types of factors. Because American Presidential elections are finally decided by 50 separate state elections, it is important to poll in some of the crucial battleground states to see how the national contest is playing out in different locales. Because last-minute shifts in voter preference can and sometimes do occur, it is important to poll as close as is practical to Election Day with pre-election polling, then to document actual voter decision-making through exit polling on Election Day itself, and then to be prepared to call back the pre-election respondents afterwards in the event there are surprises in the election outcome. Because voter turnout can be critical, it is important to develop and improve ways to estimate the likelihood that individuals will vote. Because nonvoters are almost half the voting age population, it is important to find ways to hear their voices too. And because the quality of the polling data is crucial to all subsequent interpretation and reporting, it is important to invest in properly drawn samples, high-quality interviewing, and an experienced staff for question writing and data analysis. Finally, when it comes to the ultimate reporting of poll results, it is a boon and a pleasure to have so many fine correspondents, editors, and graphics artists on your team!

  Notes

  Many thanks to the staff members of the News Surveys Department at The Times

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Elections and Exit Polling

  • Fritz J. Scheuren, Wendy Alvey(Authors)
  • 2008(Publication Date)
  • Wiley

    (Publisher)

  Using data from the 2006 mid-term congressional elections, I conduct an analysis that ' Pre-election polls are conducted between elections and often used to assess the ' feasibility of a particular candidate's run for office. In addition to other questions, they often ask respondents, If the election were held today, who would you vote for in the upcoming primary (or election)? Many poIlsters conduct successive pre- eIection polls, building up to the election, to examine potential voting trends and how they are impacted by current events. Harrison (2007) examines pre-election polls for the 2006 mid-term elections in an electoral context, to determine which factors are most likely to impact results. 102 MID-TERM ELECTIONS: 2006 accounts for factors related to both survey methods and factors involving the dynamics of elections. Pre-election polls represent an important institution in contemporary politics, journalism, and campaigns. During the course of an election, the accuracy of electoral forecasts, or the predictive accuracy of a pre-election poll, represents an item of substantive concern. In a different capacity, survey researchers and methodologists look to pre-election polls as a rich and valuable source of information about survey methods. Political polls are, after all, surveys. In contrast with many surveys, pre- election polls-at least those close to an election-which attempt to forecast a vote can be externally validated by actual election results. Though the study of the relationship between survey methods and the accuracy of pre-election polls dates, at least informally, to the early days of newspaper polling, attention was formally focused on measures of the accuracy of pre-election polls af- ter newspapers incorrectly predicted the winner of the 1948 presidential election.

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  The UK General Election of 2010

  Explaining the Outcome

  • Justin Fisher, Christopher Wlezien(Authors)
  • 2013(Publication Date)
  • Taylor & Francis

    (Publisher)

  There are two further details of the methodology of the poll that we should explain in a little further detail, not least because the position in 2010 was somewhat different from that in 2005. Both concern potential biases that might arise because some voters are beyond the reach of the exercise. The first such group comprises those who opt to vote by post. The second consists of those who refuse to participate or who for operational reasons are not successfully contacted by the poll.

  Postal voters. The exit poll was conducted by contacting a random sample of voters as they left the polling station. Thus those who voted by post, a method of voting that since 2001 anyone can choose to use, lay outside the sampling frame. Given the logic of the research design this is not a problem if, within constituencies and polling districts, the proportion of postal voters is stable over time5 and the change in party support among postal voters is similar to that among in-person voters. Between 2001 and 2005 there had been a substantial increase – from 4% to 12.4% – in the proportion of people eligible to vote by post as voters opted to take advantage of the new regulations (Electoral Commission, 2005). As a result, in 2005 an attempt had been made, using information gleaned from opinion polls, to ascertain whether the change in party support amongst postal voters in 2005 and that amongst such voters in 2001 was different from that in the exit poll, though in the event it made less than half a percentage point difference to the overall estimated swing between Labour and the Conservatives (Curtice & Firth, 2008). This time around, however, it seemed from the evidence available that the growth in postal voting since the previous election would be nothing like on the same scale and thus the decision was made not to repeat the exercise.6 The implied assumption that change in vote choice would be similar for in-person and postal voters might have appeared somewhat heroic but there was no specific reason to anticipate that in practice it would not hold.

  Refusals and non-contacts.

  Sign up to read

  Learn more about book