The tiny fraction of voters who are still undecided (from 5 to 8 percent in most polls) may determine who is the next president, as Mitt Romney and President Obama remain effectively tied in polls. But who will they support?
Here are three theories on how the last swing voters will break.
1. The incumbent theory
Political scientists traditionally have argued that most undecided voters will back the challenger, arguing their indecision is in effect an acknowledgement they don’t like the incumbent. Often, particularly in races where the challenger is much less well-known than the incumbent, the challenger’s biggest task is just making sure voters know enough about his biography and personal history to achieve basic comfort with voters.
“In the ’80s and ’90s, most incumbents were much better-known than their challengers. So when a voter said he was “undecided,” what he really meant is that he knew the incumbent and did not particularly want to vote for him or her … In Senate races between 1994 and 2004, on average 70 percent of undecideds broke to the challenger, 30 percent to the incumbent,” wrote Democratic pollster Mark Mellman in a recent piece.
Obviously, voters usually know the name of the two people running for president. But as Christian Heinze wrote recently for The Hill, the last two presidential elections with an incumbent and a challenger showed evidence that voters deciding at the end of the campaign favored the newcomer. In 2004, John Kerry won the majority of voters who decided how they vote in the last week before the election, while George Bush led among voters who had decided before that. (The vast majority of voters had decided before the last week, so this disparity did not help Kerry that much.) In 1996, the plurality of voters who decided in the last week backed Bob Dole, even though Bill Clinton easily won the election.
If the race is effectively tied, as many polls show, a majority of undecided voters breaking to Romney could give him a victory.
2. The “it depends” theory
But Mellman argues that the incumbent theory largely held true in an era in which challengers were much less known. Breaking down data from the 2008 and 2010 elections, he found the incumbent candidate won more of the undecided voters than the challenger in most of the key U.S. Senate races.
Extended to the presidential race, his argument is that undecided voters know a lot about Obama, but also Romney. Their indecision shows reservations about both candidates and gives little guide to how they will vote on Election Day.
3. The “staying home” theory
There is a popular view that undecided voters are people who are saying they will vote, but actually won’t. But in fact, most elections include a fraction of people who tell exit pollsters they decided who to back in the last week. In 2004, about 11 percent of voters they said decided in the last week, as did about 10 percent in 2008.
So there is almost certainly a bloc of people who will vote on November 6 who don’t yet know who they are backing.
Follow Perry Bacon Jr. on Twitter at @perrybaconjr