Any candidate who wins the popular vote by at least three percentage points is certain to win the electoral college, and any candidate who wins the popular vote by as much as a full percentage point is overwhelmingly likely to win the electoral college. So the best way to follow the election is to read the national polling averages.
A vice presidential pick should appeal to key groups in swing states.
The National Journal, for example, has said former Minnesota Gov. Tim Pawlenty makes sense as a running mate for Mitt Romney because "in an election that could be decided by Rustbelt battlegrounds, it couldn't hurt to have a guy capable of matching VP Biden's blue-collar appeal." Supposedly, Biden was picked in 2008 to help Obama with white working-class voters in Pennsylvania and Ohio.
The problem is that there's no evidence that vice presidential candidates have that kind of impact. The exception, research has found, is that a popular running mate might help by a couple of points in his or her home state. But even if a candidate knows what the swing states are, it's a lot harder to figure out where exactly the campaign could most use that boost.
Ignore the national economy and focus on swing-state economies.
Ever since political scientists showed that the economy is a major factor in presidential elections, they have struggled to determine what exactly that boils down to. If the local economy is the deciding factor, then it would make sense for the candidates to focus on how the economy is doing in, say, Dade County, Fla., or Hamilton County, Ohio.
It turns out, however, that impressions of the national economy are what really move votes. As one recent study of voting and the economy concluded: "Evidently, voters believe the president has little effect on their local economy, and they do not form their evaluation of the national economy based on surrounding conditions. ... People form their opinions of the national economy based on nonlocal factors, such as the national media."
Once a swing state, always a swing state.
It's true that some states will be competitive perpetually, but over time, some experience significant changes. West Virginia, for instance, went from being one of the strongest Democratic states in 1980 to being one of the strongest Republican states now. It's very hard to know in advance, certainly until the last few weeks of the campaign, what the key swing states — the ones that will truly determine the winner — will turn out to be.
The best illustration is to note which states have been closest to the national margin of victory. For example, when Obama won by seven percentage points in 2008, which state results most closely matched that number? Those states would have determined the winner if the Electoral College count were very narrow. The five states closest to the overall margin of victory in 2008 were Virginia, Iowa, New Hampshire, Florida, and Minnesota; in 2000, they were Oregon, Iowa, New Hampshire, New Mexico, and Florida; and in 2004, they were Pennsylvania, Nevada, Ohio, Michigan, and Minnesota. That's 11 states over three cycles, and completely different sets in 2000 and 2004.
Sure, we don't expect Wyoming or Vermont to be competitive. But we can't know right now whether the state that puts Romney or Obama over the top will be Colorado, Ohio, or any of a dozen other possibilities.
Republicans can't win without Ohio.
You'll hear plenty of similar pronouncements every election season. The Republicans have never won without Ohio; therefore, they can't win without Ohio. Or: There is a "blue wall" of states that the Democrats have captured consistently since 1992, so any poll showing a strong Republican tilt in one of those states indicates that Obama is doomed.
Forget these "rules." When Republicans won three consecutive presidential elections in the 1980s, pundits became convinced that the GOP had an Electoral College lock. That view lasted exactly as long as the party's national vote lead did; as soon as Bill Clinton took the national lead in 1992, it turned out that some of the Republican "lock" states were swingers after all.
Yes, one needs 270 electoral votes to win. But in most years, the Electoral College margin will be much larger than the popular one. And on the rare occasions when the popular vote is very close, it's not possible to guess in advance which states will really make a difference. Our elections are much more national than our obsession with swing states implies.
Jonathan Bernstein is a political scientist. He wrote this for the Washington Post.