Barack Obama has George W. Bush beat at one thing: travel to battleground states. As The Wall Street Journal reported:
When President Barack Obama went to Scranton, Pa., Wednesday to promote his jobs package, he logged his 56th event in a presidential battleground state this year, putting him well ahead of President George W. Bush’s record-breaking swing-state travel in 2003.
It’s not surprising that presidents conveniently find themselves in battleground states. But does this travel make any difference? Political science studies of the effects of presidential “candidate appearances” don’t provide a straightforward answer.
A famous example of this sort of campaigning was Harry Truman’s “whistle-stop” tour in 1948. Truman’s job approval ratings were then under 40 percent and he was the clear underdog in the race. So he crisscrossed the country on a train, hoping to hammer home his criticisms of the “do-nothing Congress.”
The political scientist Thomas Holbrook has found that Truman’s targeting strategy wasn’t much different than Mr. Obama’s: he devoted particular attention to states where the election was closely contested. More striking was how much harder Truman campaigned than his opponent, Thomas Dewey: 248 campaign stops versus Dewey’s 40.
Did this help Truman win? It appears so. Mr. Holbrook finds that for every 4 additional stops that Truman made in a state, he earned another 1 percent of vote share. Moreover, Mr. Holbrook conducts a simulation in which Truman’s travel was reduced to the level of Dewey’s. The simulation suggests that Truman would have lost California, Illinois, and Ohio as a result, enough to have given Dewey the victory that the Chicago Tribune famously assumed.
There is also modest evidence of political benefits from candidate travel in presidential primary campaigns. The political scientists Lynn Vavreck, Constantine Spiliotes, and Linda Fowler found that New Hampshire voters who reported meeting one of the Republican candidates in the 1996 primary had a more favorable impression of them holding a number of other factors equal. And Dan Balz’s and Haynes Johnson’s post-mortem of the 2008 election reports that both Barack Obama and John Edwards made many more campaign stops in Iowa than Hillary Clinton, which could help explain her weak showing.
But there is some evidence that these effects are weakening, especially in the general election. An article by the political scientist Daron Shaw finds that, in the 1988-96 presidential elections, an additional 3 appearances by the Republican candidate in a state, assuming the Democrat did not respond in kind, would shift an average of 64 electoral votes toward the Republican. The comparable effect of appearances by the Democratic candidate was 87 electoral votes. However, in Mr. Shaw’s subsequent study of weekly tracking polls in battleground states during the 2000 and 2004 presidential elections, presidential or vice-presidential candidate appearances did not consistently move poll numbers and the effects were small — “on the order of tenths of a percentage point.”
A final example: the now-famous experiment conducted during Rick Perry’s 2006 gubernatorial campaign — which is examined in Sasha Issenberg’s forthcoming book. Thanks to a collaboration between his campaign and four political scientist “eggheads,” Mr. Perry’s candidate appearances, radio ads, and television ads were randomly assigned to different media markets within Texas. I’ve discussed the results of the advertising buys here. The results of Mr. Perry’s appearances, reports Mr. Issenberg in this interview, were more significant:
They [the eggheads] found that Perry’s presence in a city had an impact: his approval ratings went up, and contributions and volunteer signups increased after he did a public event.
What are we to make of this somewhat inconsistent set of results? In some ways, they parallel what we know about other forms of campaigning such as candidate advertising. In general, campaigning will be more effective when voters are less familiar with the candidates (as in many presidential primaries) or where one candidate is simply out-campaigning opponents — as Truman did against Dewey and as Mr. Perry was doing when the 2006 experiment was conducted. But when the candidates are relatively familiar to voters and devote roughly equivalent amounts of time to particular states, as in presidential general elections, the effects of candidate travel are smaller or even non-existent. It’s hard to get leverage when your opponent can match you, plane flight for plane flight and stump speech for stump speech.
The potential benefits of candidate travel may also be compromised by the media outlets in the places candidates visit. Candidates often expect local media to be more deferential and less cynical than national media, thereby providing them with positive coverage when they visit the Scrantons and Lansings and Green Bays of the country. This was, according to Issenberg, the lesson learned by Mr. Perry:
Perry just got much better press when he was in Lubbock or Beaumont than when newspapers and TV stations there relied on stories out of Austin, which tended to be much more cynical or negative.
But this may not be consistently true. Consider two studies of George W. Bush’s presidency by the political scientists Matthew Eshbaugh-Soha and Jeffrey Peake. When Mr. Bush traveled in 2005 to promote his plan to privatize Social Security, local coverage was typically more positive than that of the Washington Post. But media coverage of Bush in 2003 was not more favorable in local outlets than in The New York Times. This finding parallels Brendan Nyhan’s recent argument that local New Hampshire media often cover the presidential primary in ways similar to the national media, emphasizing the horse race and candidate strategy.
The lesson? Although it would seem foolhardy for Mr. Obama or his G.O.P. opponent simply to stay home in 2012, there may be no direct relationship between the frequent flier miles they earn and the votes they receive.