Ipsos Presidential Poll 2004: What We Will Look For In The Exit Polls

Between October 28-31, Ipsos Public Affairs conducted a poll for its own information, interviewing a random selection of 2,501 adults nationwide (excluding Alaska and Hawaii), yielding 2,167 registered voters and 1,618 likely voters. Here are some key findings.

Our models range from the broadest possible (Registered Voters), to somewhat restrictive (LV1), to very restrictive (Likely Voters). With this final poll, we allocate, to the fullest extent possible, the undecided 4% in our final poll before the 2004 Presidential Election. Here are the results:

That means either side could win!

UNALLOCATED POLL RESULTS Anybody But Bush, or Is It Bush Over Anybody? The most important difference between Bush and Kerry voters arises from their fundamental motivation. Only half of Kerry voters are driven primarily by their strong positive support for Kerry--an equal number of Kerry voters concede that theirs is primarily a vote against Bush.

• Kerry gets almost all the anti-Bush vote (49% disapprove of the way Bush is handling the job of President), but Kerry fails to win all the votes of those who feel the country's direction is off on the wrong track (52%). That is the difference between an easy Kerry victory and the incredibly close race we face.
• Kerry's failure is a failure of promise--he does not promise anything new, just a more effective prosecution of the war in Iraq and the war on terrorism. Health care, jobs and other domestic issues have fallen off the agenda. Overall, 67% say protecting the country is the more important issue and 28% say creating jobs is the more important issue in deciding how to vote.

The Sense of the Country's Direction Drives the Issues Agenda People who feel the country is on the right track and want to maintain the status quo say the most important voting issues are terrorism and moral values.

People who feel the country is on the wrong track and want change say the most important voting issues are economy/jobs and Iraq.

People who feel the country is on the right track are unanimous in their assessment that right now, the more important priority is protecting the country, not jobs creation.

People who feel the country is on the wrong track are divided evenly between the roughly half of them who say the more important issue is creating jobs and the other half who say the more important topic is protecting the country.

Naturally, those who choose terrorism as the top voting issue also choose protecting the country as the priority over jobs creation. Those who choose moral values as the top voting issue also put protection over jobs creation. People looking for a strong, honest and trustworthy leader also make protection the priority over jobs.

Voters looking for a candidate who brings change are as likely to choose job creation as they are to choose protection as the higher priority.

Kerry wins as the candidate better at creating jobs (although his margin is slipping back to where it was in Kerry's worst poll in early September.) Bush wins as the candidate better at protecting the country, but protection trumps jobs after a closing campaign that focused on foreign policy issues.

ALLOCATED VOTE RESULTS: In our final poll, 4% were undecided--some of them early voters who declined to say whom they supported, and others are voters having a devil of a time making up their minds. We employed a very simple allocation:

Test 1: If they have a party identification, we voted their preference, except if their scores on Test 2 indicated the opposite result ; in addition, any that we could not allocate based on party identification, we allocated based on their assessments of Bush's job performance and whether Bush or Kerry would do better on specific issues;

Test 2: If party identification did not allocate them, we assigned them if they showed a consistent preference on five Bush performance ratings and 4 Bush vs. Kerry performance expectations. By consistent, they had to always choose Bush (or Bush job approval) or undecided to be allocated to Bush, or they had to always choose Kerry (or Bush job disapproval) or undecided to be allocated Kerry.

We started out with 4.3% undecided, and this allocation test reduced that number to 0.3%, yielding our final allocated measure of the vote:

Bush/Cheney Kerry/Edwards Nader Other Undecided

50% 48 1 1 0

Four people we interviewed could not be allocated. We left them in our allocated vote model.

The next eight pages show the allocated vote by subgroups:

[+] click to enlarge [+] click to enlarge [+] click to enlarge [+] click to enlarge [+] click to enlarge [+] click to enlarge [+] click to enlarge [+] click to enlarge

Here Are the Key Groups to Watch A few key constituencies hold the balance of the election in their hands:

Decided in final week/weekend (10%): KE 45%, BC 40%, Nader 6%, Other 9%. The final movement is to Kerry, however there were many Nader-Camejo and other candidates winning support in that same time frame. These aren't the same as new voters. They are disadvantaged in terms of work, schooling, and finances.

Looking across the columns by when the decision was made,

BC tied KE among those deciding more than 3 months ago (KE 49%, BC 51%), BC won those deciding in the past 3 months (KE 44%, BC 54%), but then... BC tied KE among those who decided in the past month (KE 50%, BC 48%), and BC lost those who were not decided when we interviewed or had just decided.

If the final round of polls from all sources showed a general tightening up of a race that most pollsters showed BC to hold nominal, statistically insignificant leads before the final weekend to a race that was a flatfooted tie or + 1%, this final movement explains why Kerry was able to close that final gap before Election Day.

Early voters (22%) KE 45%, BC 53%, Nader 1%. The typical early voter is a retiree. They tend to be older, retired, and live in the south and west and in urban areas. Early voters also include a number of evangelicals, and they tend to be male and gun owners.

New voters (12%) KE 49%, BC 49%, Nader 2%. Note that we define "new voters" simply as those who do not report having voted in 2000, but are likely to vote on Tuesday. They may have voted in previous elections. New voters appear to be drawn from groups which we tend to think of as people who stay away from the polls. Six in ten have not voted before at all, and not all of them were too young to vote in 2000. They are employed, but have little financial security. Politically they are moderate, and include a higher than average number of women and minorities.

Attend church At Least Once a week (40%) KE 37%, BC 61%, Nader 1%. These are conservative households, and a majority are evangelical or born-again Christians. Most are married and they are somewhat more likely to have children in the household. They tend to live in the South. Eight in ten are homeowners. But this is not a homogenous group, it also includes a number of retirees and blacks.

Do not attend church weekly (58%) KE 56%, BC 42%, Nader 1%. Non-churchgoers tend to be single, have some college education, politically moderate, and reside in the Northeast and West.

More Important Priority Creating Jobs (28%) KE 86%, BC 11%, Nader 2%. One of the most educated groups, nearly half have a college degree. They are not particularly high income, however, and a third have experienced job loss. They tend to live in the Midwest.

Protecting US (67%) KE 31%, BC 67%, Nader 1%. Conservatives in many ways--politically conservative, religiously observant, evangelicals, shareholders, gun owners, Southerners.

Most important issue:

Jobs voters (24% economy/jobs is top voting issue): KE 77%, BC 20%, Nader 1%. Midwestern, above-average education, some experience with job loss.

Terrorism voters (21% top voting issue): KE 14%, BC 85%, Nader 1%. Southern white males, many of them veterans and gun owners, often well educated, homeowners and shareholders. They are also voting for a strong leader.

Moral values voters (20%): KE 15%, BC 84%, Nader 0%. Traditional, religious and evangelical households in the Midwest and South.

Iraq voters (17%): KE 74%, BC 25%, Nader 0%. Northeasterners, often retired.

For more information on this press release, please contact: Janice Bell Director, Ipsos Public Affairs Washington, D.C. 202.463.7300

About Ipsos Public Affairs Ipsos Public Affairs, headquartered in Washington D.C., is a non-partisan, objective, survey-based research company made up of campaign and political polling veterans as well as seasoned research professionals. The company conducts strategic research initiatives for a diverse number of American and international organizations, based not only on public opinion research but often elite stakeholder, corporate, and media opinion research. It has offices in New York City, Chicago, San Francisco, and Washington, with affiliates around the world. Ipsos Public Affairs conducts national and international public opinion polling on behalf of The Associated Press, the world's oldest and largest news organization, and conducts the young voters poll for Newsweek.com. Ipsos Public Affairs is an Ipsos company, a leading global survey-based market research group.

To learn more, visit: www.ipsos-na.com/news/pa

About Ipsos Ipsos is a leading global survey-based market research company, owned and managed by research professionals. Ipsos helps interpret, simulate, and anticipate the needs and reactions of consumers, customers, and citizens around the world. Member companies assess market potential and interpret market trends. They develop and build brands. They help clients build long-term relationships with their customers. They test advertising and study audience responses to various media. They measure public opinion around the globe. Ipsos member companies offer expertise in advertising, customer loyalty, marketing, media, and public affairs research, as well as forecasting and modeling and consulting. Ipsos has a full line of custom, syndicated, omnibus, panel, and online research products and services, guided by industry experts and bolstered by advanced analytics and methodologies. The company was founded in 1975 and has been publicly traded since 1999. In 2003, Ipsos generated global revenues of $644.2 million U.S. To learn more, visit: www.ipsos.com Ipsos is listed on the Euronext Paris Premier Marchй, and is part of the SBF 120 and Next Prime Indices as well as eligible to the Deferred Settlement System (SRD). Euroclear code 7329, Reuters ISOS.LN, Bloomberg IPS FP.