Fall Campaigns Have Predictable Effects on Presidential Elections, UB Political Scientist Says

BUFFALO, N.Y. -- Contrary to long-held beliefs of the media and political scientists, U.S. presidential campaigns have systematic, predictable and significant effects on the outcomes of elections, an authority on the electoral process at the University at Buffalo has found.

The theory of predictable campaigns developed by James E. Campbell, professor of political science in the UB College of Arts and Sciences, is based on a decade of research on 130 years of electoral history and the history of public-opinion polling in elections over the past 50 years.

Results of the research by Campbell is the topic of his new book, "The American Campaign: U.S. Presidential Campaigns and the National Vote," published recently by Texas A&M University Press.

In a related book that also was published this year, "Before the Vote: Forecasting National Elections" (Sage Publications), Campbell and co-editor James C. Garand invited seven political scientists who had developed models for forecasting the results of presidential elections to revisit their forecasts of the 1996 election and present contingent forecasts for this year's general election.

Those 1996 forecasts had been published in American Politics Quarterly in October 1996.

"It was a risky proposal because predicting anything from the weather to the stock market always involves a degree of uncertainty," Campbell points out. "But several election-forecasting models have had a good track record in recent years and 1996 was no exception."

As a group, Campbell observes, the forecasts of the 1996 presidential election made two to three months before the election were more accurate than the final polls conducted right before the election. In fact, Michael Lewis-Beck of the University of Iowa and Charles Tien of Hunter College were right on target, predicting Clinton's share of the vote within one-tenth of a percentage point.

Does the accuracy of the forecasts made before the fall campaign mean that the campaign has no impact on the election? "Not so," says Campbell.

The reason that presidential elections can be forecast so accurately, he explains in "The American Campaign," is that the effects of general-election campaigns are largely predictable.

"Journalists and political scientists tend to have different misconceptions of the role campaigns play in the outcome of presidential elections," he notes.

"Political scientists tend to look at campaigns as rituals or spectacles that have minimal effects on the outcome of presidential elections," he points out, while journalists believe that election results turn on the unforeseeable and unpredictable twists and turns of the campaign, such as the inspirational speech, the regrettable gaffe or the brilliant strategic move.

Campbell says the effects of campaigns are predictable because their impact takes place between a large group of solid Democratic voters and a large group of solid Republican voters who have decided well before the election for whom they'll cast their ballots.

Campaigns are about winning over those in the middle -- those voters who are affected by their initial impressions of the major party candidates, their attachment to a political party, the state of the election-year economy and whether or not an incumbent is running.

The candidate ahead in the polls after the parties' conventions usually wins the election, but with a much smaller margin than that shown in the polls at the campaign's beginning.

Once the voters know the candidate match-up, many reach a quick verdict. But the intense competition of the fall campaign also tends to narrow the front-runner's lead.

In his research, Campbell found one of the most important reasons for this "narrowing effect" is that the trailing candidate in the pre-campaign polls often starts the campaign with some disarray within the party but that many party members drift back to that candidate over the course of the campaign.

He also found that the campaign tilts toward a certain candidate depending on how well the economy is doing in the election year and whether an incumbent is running.

"If the economy is strong and an incumbent is running, voters evaluate campaign information with a tilt toward re-electing the candidate of the party currently in the White House."

Campbell says it's too early to tell who the next president will be without an incumbent in the race and with both parties nearly equal in their numbers of identifiers.

"I would anticipate a particularly close contest," he believes.

It would all depend on how the voters between now and August evaluate the Democratic and Republican candidates, how amicably the Bush-McCain and Gore-Bradley primary fights are settled and how the economy does in the next few months.

The election may well depend on how well and quickly Republicans can regroup after they choose their nominee.

"On the one hand," Campbell notes, "the Republican battle is fierce but, on the other, Republicans have been out of the White House for eight years and may pull together in the end."