Exit Polls Research Articles

Estimating voter migration in Canada using generalized maximum entropy

As voters switch political preferences from election to election, understanding the magnitude of voter flows among parties and transitions between voters and non-voters is an essential element of political analysis. As exit polls are uncommon in Canada, voter migration can also be estimated using suitable statistical techniques. Backing out micro-level voter migration probabilities from macro-level election data is a problem of ‘ecological inference.’ This paper uses the method of generalized maximum entropy (GME) to estimate voter migration patterns for the two most recent Canadian federal elections (2004 and 2006) and two most recent provincial elections in British Columbia (2001 and 2005). The estimation results answer important questions about voter behaviour in Canada. These results will be of interest to political scientists, historians, and politicians, as well as econometric practitioners who wish to estimate voter migration.

Tyranny of the White Majority? Minority Representation in American Democracy

Critics have long feared that America's winner-take-all system, it's clear white majority, and it's sharp racial divisions would combine to undermine the interests of racial and ethnic minorities. Unfortunately, we have few means of assessing minority representation in American democracy. Descriptive representation misses most of the potential influence of minorities and studies of substantive representation generally have to focus on a particular locale or policy with little hope of a more general assessment. I introduce a new measure of representation. For any election, I simply count up how many voters from each racial and ethnic group vote for the side that wins. Using this new measure and data from the series of Voter News Service exit polls and a sample of mayoral elections, I find that across the range of American elections, blacks are consistently more likely than other groups to end up losers. I then show why blacks are uniquely disadvantaged.

Income and Vote Choice in the 2000 Mexican Presidential Election

Using multilevel modeling of state-level economic data and individual-level exit poll data from the 2000 Mexican presidential election, we find that income has a stronger effect in predicting the vote for the conservative party in poorer states than in richer states - a pattern that has also been found in recent U.S. elections. In addition (and unlike in the U.S.), richer states on average tend to support the conservative party at higher rates than poorer states. Our findings are consistent with the 2006 Mexican election, which showed a profound divide between rich and poor states. Income is an important predictor of the vote both at the individual and the state levels.

RACE, GOD, AND GUNS: UNION VOTING IN THE 2004 PRESIDENTIAL ELECTION

The conservative political preferences of many working class Americans have been the subject of much academic and popular analysis in recent years. This article investigates the voting behavior of union household residents in the 2004 presidential election. The source for this information is national and state exit polls from the 2004 election. There has been much debate about whether white working class support of Republicans is rooted in conservative cultural values. Despite ardent opposition by the Bush administration to the goals of organized labor, 46 percent of white voters who resided in union households voted Republican in the 2004 presidential election. The impact of race, religion, and gun ownership on the voting choice of labor households is investigated in an effort to provide an understanding of conservative voting by so many households affiliated with an interest group that is at odds with the Republicans.

Reaching disenfranchised youth and mobile populations in Ghana through voluntary counselling and testing services for HIV

This paper documents the evaluation of a 20-month project to provide voluntary counselling and testing (VCT) to a mobile population of youth surrounding the Agbogbloshie market in Accra, Ghana. The specific objectives of the evaluation were to determine: 1) to what extent targets for providing VCT services to the specified population were reached; 2) how HIV prevalence among clients compared to that of the general population; 3) to what extent former clients self-reported behaviour change; and 4) whether useful lessons could be drawn regarding fees, hours, and location of services, as well as use of peer educators to increase use of VCT services among the target population. Various methodologies, including questionnaires, focus group discussions, a review of the service statistics and an exit poll of clients were used to evaluate the project. The service statistics demonstrated that the project exceeded the life-of-project target for number of clients by nearly 40%. Prevalence for the VCT client population (aged 15–25) was higher than for the general population (aged 15–24), although the gender differentials were similar. Focus group data suggested that clients may have adopted behaviour changes as a result of VCT. Finally, focus group discussions and VCT service trends showed that the high number of clients was largely influenced by three factors: services being free, location and hours of services being convenient to the target population, and use of peer educators to promote the services. In addition, the evaluation highlighted the importance of the counselling component of VCT, even as counselling can get short-changed at the expense of HIV testing when large numbers of clients are involved. The evaluation stressed the need to appropriately remunerate peer educators for their work whenever possible. Finally, VCT programmes continue to face challenges such as: HIV stigma as a barrier to people coming to be counselled and tested; insufficient availability of medication, support and services for HIVpositive clients; and difficulty of ensuring the sustainability of VCT programmes.

Studying Nonresponse in Mexican Exit Polls

Studying Nonresponse in Mexican Exit Polls

Low enrolment in Ugandan Community Health Insurance Schemes: underlying causes and policy implications

BackgroundDespite the promotion of Community Health Insurance (CHI) in Uganda in the second half of the 90's, mainly under the impetus of external aid organisations, overall membership has remained low. Today, some 30,000 persons are enrolled in about a dozen different schemes located in Central and Southern Uganda. Moreover, most of these schemes were created some 10 years ago but since then, only one or two new schemes have been launched. The dynamic of CHI has apparently come to a halt.MethodsA case study evaluation was carried out on two selected CHI schemes: the Ishaka and the Save for Health Uganda (SHU) schemes. The objective of this evaluation was to explore the reasons for the limited success of CHI. The evaluation involved review of the schemes' records, key informant interviews and exit polls with both insured and non-insured patients.ResultsOur research points to a series of not mutually exclusive explanations for this under-achievement at both the demand and the supply side of health care delivery. On the demand side, the following elements have been identified: lack of basic information on the scheme's design and operation, limited understanding of the principles underlying CHI, limited community involvement and lack of trust in the management of the schemes, and, last but not least, problems in people's ability to pay the insurance premiums. On the supply-side, we have identified the following explanations: limited interest and knowledge of health care providers and managers of CHI, and the absence of a coherent policy framework for the development of CHI.ConclusionThe policy implications of this study refer to the need for the government to provide the necessary legislative, technical and regulative support to CHI development. The main policy challenge however is the need to reconcile the government of Uganda's interest in promoting CHI with the current policy of abolition of user fees in public facilities.

Patient satisfaction with health services in Bangladesh

Concern over the quality of health care services in Bangladesh has led to loss of faith in public and private hospitals, low utilization of public health facilities, and increasing outflow of Bangladeshi patients to hospitals in neighbouring countries. Under the circumstances, assessment of the country's quality of health care service has become imperative, in which the patient's voice must begin to play a greater role. This study attempts to identify the determinants of patient satisfaction with public, private and foreign hospitals. A survey was conducted involving inpatients in public and private hospitals in Dhaka City and patients who have experienced hospital services in a foreign country. Their views were obtained through exit polls using probability and non-probability (for foreign hospital patients) sampling procedures. Regression models were derived to identify key factors influencing patient satisfaction in the different types of hospitals. Doctors' service orientation, a composite of 13 measures, is the most important factor explaining patient satisfaction. Policy implications are discussed.

Moral Values and Vote Choice in the 2004 U.S. Presidential Election

Scholars and journalists have emphasized the growing importance of cultural issues and a “values divide” in shaping recent political behavior of the American electorate. Discussion of a values‐based cleavage has been especially evident since exit polls for the 2004 presidential election revealed that “moral values” topped the list of issues that respondents cited as the “most important” issue in the election. While there has been considerable debate and disagreement over the influence of moral values in 2004, no analysis has performed a crucial test of the moral values hypothesis, namely developing a multivariate model to examine the effect of moral values on vote choice relative to other issue preferences and demographic variables. This article develops such a model using data from the 2004 American National Election Study. Findings suggest that moral values exerted an important effect on vote choice in the 2004 presidential election, even when other predictors of vote choice were included in a multivariate model.

BLOGS AND JOURNALISTIC AUTHORITY

This paper asserts that the role of blogs cannot be adequately understood without examining the established media context in which they appear. Blogs operate along side, in conjunction with, and in opposition to established vehicles for political information, which creates tension among journalists seeking to preserve their authority. As a site to observe the blog-traditional journalism relationship, this article examines the reaction by journalists and others to blogs’ role in US Election Day 2004 coverage. Much of the attention by journalists focuses on assessing the well-publicized decision by some blogs to release incomplete exit polls erroneously predicting a victory for Democratic candidate John Kerry. This discourse works to make sense of the status and credibility of blogs while simultaneously allowing journalists to negotiate their role as authoritative providers of political news. Ultimately, the discourse underlines the dynamism of news in a contemporary media environment marked by new forms of complexity and competitiveness.

Comment: Data Troubles: Explaining Discrepancies between Official Votes and Exit Polls in the 2004 Presidential Election

Comment: Data Troubles: Explaining Discrepancies between Official Votes and Exit Polls in the 2004 Presidential Election

Explaining Discrepancies between Official Votes and Exit Polls in the 2004 Presidential Election

Explaining Discrepancies between Official Votes and Exit Polls in the 2004 Presidential Election

Geopolitics and the Sight of the Nation: Elizabeth Bowen's The Last September

Geopolitics and the Sight of the Nation: Elizabeth Bowen's The Last September

Rural Voters in Presidential Elections, 1992-2004

The ``Red" versus ``Blue" state debate has reached a fever pitch in popular commentary, but scholars have contributed very little to the discussion by way of examining rural voting behavior. With the use of national exit poll data, this study attempts to fill this considerable void, with a detailed analysis of rural voters in the 1992-2004 presidential elections. In 1992 and 1996 the rural vote was split between the parties, but in 2000 the rural vote shifted decidedly in favor of Republican George W. Bush and it stayed with the incumbent in 2004. This research on the voting behavior of rural voters in recent presidential elections documents and evaluates the many differences between rural and non-rural voters, and accounts for several of the factors leading to an increase in rural Republican voting in 2000 and 2004. The conventional wisdom that rural voters are more likely to be so-called values voters is true and this translates into greater Republican support. Further, on virtually every survey item in which their non-rural counterparts share the same survey response, rural voters are consistently more Republican in their presidential vote choice. Dissatisfaction with President Clinton—termed Clinton fatigue—was much more pronounced among rural voters and this was a major reason for the strong rural shift in favor of the Republican Party in 2000.

One of Our Own: Black Female Candidates and the Voters Who Support Them

This article examines the role of race and gender in candidate evaluations. Unlike previous research, we argue that the role of race and gender in electoral politics must be examined simultaneously because of their mutually reinforcing relationship. To do so, we explore the connection between the race and gender of voters and their propensity to support black female candidates. Using precinct‐level data, experimental data, and national exit poll data from two congressional election years, we demonstrate that black women are the strongest supporters of black female candidates. We also find that support for black female candidates is contingent on their background and political experience. Black female candidates with significant experience in politics can attract both black and white voters, regardless of gender.

Voter Dealignment or Campaign Effects? Accounting for Political Preferences in Ontario

Elections in Canada are rarely subject to the delays in voting counts present in the United States. For Americans, the time required to count hundreds of millions of ballots has made exit polls an attractive option to political commentators hoping to fill the void between the close of polls and the confirmation of results. Often face-to-face surveys with individuals leaving the ballot box, exit polls not only allow for a quick prediction of election results, but also allow students of voting to say something meaningful about the reasons behind ballot decisions. The speed with which election results are usually announced in Canada, and the sheer cost of mounting an exit poll for such a comparatively small voting population, has meant that exit polls are not a feature of Canadian politics. For the 2003 Ontario election, however, the Laurier Institute for the Study of Public Opinion and Policy (LISPOP), at Wilfrid Laurier University, mounted an exit poll in one of the 103 Ontario ridings to determine what lessons, if any, could be drawn from the experience. The LISPOP exercise, the first American-style exit poll conducted in Canada, provides unprecedented data on voter motivations only minutes after individuals cast their ballots. This article provides an analysis of the results of this poll. The LISPOP exit poll had two purposes. First, the project provided an opportunity to determine the feasibility of exit polls in a Canadian context, and we have published an analysis of the methodological lessons of the experiment (Brown et al. 2006). Second, the poll provides much-needed information about the provincial voting habits of Canadians. At the federal level, reports of the Canadian Election Study (CES) allow us to determine why individuals vote and why they back the parties that they do. Such studies allow us to make links among an individual's attitudes, demographic background, and political behavior. Students relying on federal electoral studies to understand sub-state voter behavior face two limitations, however. First, the act of voting and participation in the survey are separated by several days, a period in which media coverage often highlights the main themes of the campaign, and provides justifications for why voters backed the eventual winner. It is possible that in such circumstances voters could begin to blend their own reasons for voting with those highlighted by the national or local media. The timing of data collection thus presents a challenge to the reliability of the survey. The second limitation to federal election studies has nothing to do with methodology and more to do with the gaps that remain in our knowledge. Although operating since 1965 at the federal level, electoral studies at the provincial level are rare. Obviously, federal electoral studies are not meant to probe provincial political behavior. We still do not know whether the factors that lead individuals to vote one way at the federal level are also at play in provincial voting decisions. As a result, this paper addresses three research questions. First, are the predictors of support for political parties similar to those we find at the federal level? Obviously the nature of partisan competition is different in the three-party system of Ontario than in Canadian elections, where the partisan spectrum is more crowded. We are less interested to discover that predictors work in the same direction than that they are equally relevant. Second, in the context of the Ontario election, are we able to identify predictors of behavior that are consistent with past studies of voter behavior? Research has suggested that gender was a significant factor in determining support for the Conservative Party in previous provincial elections. We are interested to see whether voting behavior in the 2003 election can be explained by this and other previously identified predictors. Last, we know that the 2003 election produced a change of government, an event made possible by the existence of vote switching within the electorate. …

Was the 2004 presidential election stolen?: exit polls, election fraud, and the official count

Was the 2004 presidential election stolen?: exit polls, election fraud, and the official count

Estimating Voter Migration in Canada Using Generalized Maximum Entropy

As voters switch political preferences from election to election, understanding the magnitude of voter flows among parties and transitions between voters and non-voters is an essential element of political analysis. As exit polls are uncommon in Canada, voter migration can also be estimated using suitable statistical techniques. Backing out micro-level voter migration probabilities from macro-level election data is a problem of `ecological inference.' This paper uses the method of generalized maximum entropy (GME) to estimate voter migration patterns for the two most recent recent Canadian federal elections (2004 and 2006) and the most recent provincial election in British Columbia (2005). The estimation results answer important questions about voter behaviour in Canada. These results will be of interest to political scientists, historians, and politicians, as well as econometric practicioners who wish to estimate voter migration.

Breaking up Is Hard to Do: Explaining the 2002 San Fernando Valley Secession Vote

ABSTRACT:In November 2002 the voters of Los Angeles soundly rejected Proposition F which, if approved, would have authorized the secession of the San Fernando Valley in order to constitute a city independent of Los Angeles. This article attempts to delineate the interplay of social, economic, and political conditions related to support for secession by examining the results of a Los Angeles Times survey conducted one week prior to the vote. Because an exit poll was not taken following the vote, these data provide the most proximate evidence for the views of voters shortly before casting their ballots. The analysis of the data identifies distinctive social bases of support for secession in the San Fernando Valley that contrast with those outside the region. The results are interpreted in the context of Sonenshein’s discussion of the role of coalition formation in urban politics.

Controversies in Exit Polling: Implementing a Racially Stratified Homogenous Precinct Approach

In November 2000, exit poll interviews with voters in Florida indicated that Al Gore won the state. As a result, many television networks declared Gore the winner of Florida, a pivotal state to winning the presidency in 2000. Only a few hours later, the first vote tallies from the Florida Secretary of State's office revealed that George W. Bush was in fact leading in Florida. After 45 days of recounts and lawsuits, it was clear that the exit polls were wrong; Bush had won the state by the narrowest of margins. As a result of the flawed exit poll the media and pollsters scoured and reanalyzed the methodology used in 2000 to prepare and correct for the 2004 presidential election. The old system, Voter News Service (VNS) was scrapped entirely, and Edison-Mitofsky Research was chosen to implement a new and more accurate national exit poll in 2004 by a consortium of news organizations retained by the Associated Press called the National Election Pool (NEP). What happened? Exit poll results from Edison-Mitofsky showed John Kerry ahead in Ohio, Florida, and New Mexico—all states which he lost to Bush in 2004.Author names are listed alphabetically. The co-authors were also the co-principal investigators of the Loyola Marymount University 2005 Los Angeles Mayoral Exit Poll. Thanks to Salvador Paniagua and Haven Perez for their tremendous research assistance in implementing this project and to the more than 120 student researchers who participated in the exit polling and data entry. Robert Aguinaga and Antonio Gonzalez of the Southwest Voter Registration and Education Project also provided valuable assistance in implementing the poll. Mark Blumenthal, of mysterypollster.com was instrumental in tracking down exit poll archives.