Gallup Wednesday morning hosted an invite-only briefing for political journalists to review the metrics -- survey and non-survey based -- that matter most to predicting presidential elections.
The discussion was moderated by Gallup Editor-in-Chief Frank Newport, with analysis from Gallup Senior Editor Lydia Saad, Temple University Professor of Political Science Chris Wlezien, Gallup Managing Editor Jeffrey M. Jones, and Yahoo! Research Economist David Rothschild.
Their presentations included a review of the predictive power of each of the following, in terms of how they have related to the outcomes of past presidential elections. Here are some of the key learnings:
- The Trial Heat ballot becomes more predictive of the election outcome as the campaign progresses, especially in the last 60 days before Election Day, with the predictive validity increasing as actual voting nears. This is the key metric to watch in the home stretch of the general election campaign.
- Presidential Job Approval is very predictive of an incumbent's re-election chances, especially in the months before the party nominating conventions. Approval ratings of 50% or higher seem to assure re-election, with George W. Bush’s final Gallup approval rating of 48% providing one exception. (Ford lost with 45% approval.) Presidential approval ratings, historically, can change during the course of an election campaign and are more likely to decline than increase during the presidential election year.
- Positive Intensity Scores are a good measure of a candidate's potential before they are well-known, and is an excellent indicator of a candidate’s ability to generate strong enthusiasm from voters.
- Satisfaction With the Way Things Are Going in The Country is a broad measure that captures the climate of the election. The percentage “satisfied” to ensure incumbent re-election may be somewhere between 22% (level at which George H.W. Bush lost) and 39% (level at which Bill Clinton won).
- Economic Confidence summarizes how voters feel about the economy, with the Gallup Economic Index ranging from -100 to +100. A president can win with a less than robust economy, but a Gallup Economic Index score near zero or better may be necessary to win.
- Most Important Problem is Gallup's longest-term indicator of public attitudes toward the economy, measured in terms of Americans’ open-ended responses. Successful incumbents generally have had 35% or fewer Americans mentioning either the economy, jobs, inflation, or the federal budget in the final months prior to the election, with Reagan, at 51%, the sole exception.
- Economic Indicators, such as the unemployment rate, inflation, GDP growth, etc., help to explain why a president's approval ratings are high or low. Regardless of the absolute values, strong positive movement on these metrics in the election year can help an incumbent's re-election chances.
- Expectations of Who Will Win -- that is, answers to the question, "Who do you think will win the upcoming election?" have had a good degree of success predicting election outcomes. Research shows that expectations can be more accurate than the trial heat in certain situations.
- Prediction Markets -- where wagers are placed based on who those betting think will win or lose -- have tracked closely to political election outcomes from late-19th and early-20th century elections (prior to Gallup’s first representative polls) through the 2008 election. Prediction markets' predictive advantage over traditional trial heat polling decreases as Election Day approaches.
- Social Media sentiments are a new phenomenon and it is too soon to place value of their ability to augment traditional election measures. Social media posts are not representative of the population as a whole, and are difficult to categorize. The results should be considered more descriptive than scientifically indicative.
And keep an eye out for great balanced reporting on all of the above from top tier news organizations, most of which were represented at our briefing. Labels: economy, election 2012, most important problem, politics, presidential job approval, satisfaction, unemployment