Forecasts Predict Bush Victory

Release Date: September 7, 2004 This content is archived.


BUFFALO, N.Y. -- George W. Bush has a very good chance of winning a second term in the White House, according to "trial-heat-and-economy" and "convention bump" forecasts produced by James E. Campbell, professor of political science at the University at Buffalo.

Campbell's trial-heat-and-economy forecast predicts that Bush will receive 53.8 percent of the major party vote. The model has never been more than 3.7 percentage points off of the actual vote. The forecasting model uses the second-quarter growth rate in the gross domestic product and results of the trial-heat (preference) poll released by Gallup on Labor Day.

Because the Republican National Convention was held so close to Labor Day, Campbell for the first time produced a second forecast to take into account Bush's post-convention improvement in the polls. This forecast model uses the same GDP data, plus Bush's standing in the polls before and after the two parties' conventions. Campbell's convention bump forecast predicts that Bush will receive 52.8 percent of the popular vote for the major party candidates.

Campbell's trial-heat-and-economy forecasting model has correctly predicted the popular vote winner in the three elections since it was first used in 1992. Campbell's forecasts are the last to be released of the eight major forecasting models this year. Seven of the eight models forecast that Bush will receive a plurality of the popular vote in November, and the eighth forecast predicts a dead-heat.

"The trial-heat-and-economy forecast model takes into account the historical context of the Gallup poll conducted two months before the election and the contemporary context of the campaign as measured by the state of the economy," Campbell says. "The combination of these two factors, taken along with the advantage of incumbency, produces a forecast that historically has been quite accurate.

"The economic growth in the second-quarter this year was only so-so by election year standards," he adds. "The reason that the model points to Bush is based more on incumbency and his lead in the post-convention polls."

The second quarter GDP growth rate was 2.8 percent. The Gallup Poll released on Labor Day indicated that 52 percent of likely voters expressed a preference for George W. Bush, 45 percent expressed a preference for John Kerry, and 3 percent favored a third candidate or were undecided. Converted to two-party preferences, President Bush had 53.6 percent of the two-party split. The convention bump equation used a pre-convention poll split of 49 percent for Bush, a net convention bump of 4.7 percent, and the second-quarter GDP growth rate.

"Forecasts are based on the fundamental factors that shape public opinion during a campaign, but there's always the possibility of an 'October surprise,' such as a terrorism event, a major gaffe by a candidate, or something else unanticipated that would shift a large number of votes," Campbell says.

"Every campaign has some unpredictable element to it. There is no way, for instance, that anyone could have predicted that Al Gore would not emphasize the Clinton economic record in the 2000 election. Nevertheless, election surprises usually make only a percentage point or two difference."

Since 1992 Campbell's forecasts have been within 2.3 percentage points of the popular vote, on average. Following the 1992 presidential election, The Washington Post called Campbell the "newly crowned king of election prognosticators," after his forecast came within a half percentage point of the popular vote.

Campbell's forecasts will be published in the October issue of PS: Political Science and Politics, a publication of the American Political Science Association.

Media Contact Information

John Della Contrada
Vice President for University Communications
521 Capen Hall
Buffalo, NY 14260
Tel: 716-645-4094 (mobile: 716-361-3006)
Twitter: @UBNewsSource