If President Obama wins Franklin County, Ohio - a critical locale in a critical swing state - on Tuesday, he might want to write a thank-you note to Ohio State's football team. The Buckeyes defeated Penn State last week, thereby enhancing Obama's chances of taking central Ohio.

This isn't a guess. It's science - social science. Research shows that voters in a college team's home county tend to reward the incumbent presidential candidate after the local team wins two weeks before the election. The effect is often significant.

When it comes to predicting voter behavior, academics, pundits and reporters tend to train their attention on the big stuff - party affiliation, incumbent approval ratings, the state of the economy. All good. But like all human behaviors, voting can't be entirely reduced to an abstract set of numbers marching across a PowerPoint display. As an emerging body of research suggests, voting reflects the same whims, quirks and emotional crosscurrents that make humans such unpredictable creatures.

So, seemingly irrelevant things - where you vote, which team won on Saturday, what order the candidates' names appear on the ballot, etc. - all have small but measurable effects on voting outcomes, social scientists say. And in an election that is expected to be as close as this one, small things can turn into very big things.

Take football games. The performance of a bunch of muscled behemoths would seem about as important to a presidential vote as whether you burned your toast this morning. Which is why it was the perfect variable for a 2010 study entitled, "Irrelevant Events Affect Voters' Evaluations of Government Performance." Researchers at Loyola Marymount University in Los Angeles and Stanford University's business school sought to test whether an otherwise random and seemingly arbitrary event - the outcome of college football games - showed any correlation with the results of presidential, Senate and gubernatorial elections.

 And it did, consistently, in every election between 1964 and 2008. On average, the researchers found, a victory by a hometown team 10 days before the election resulted in incumbent candidates receiving an additional 1.61 percentage points of the vote in the team's county. A victory by an avidly followed, perpetual powerhouse team (like, say, Ohio State) had an even more significant effect, as much as 3.35 percentage points.

Why? The results suggest that the emotional state of voters is an important component in understanding their behavior, says Stanford professor Neil Malhotra, one of the study's authors. If they feel good, that can translate into how they vote. And if they feel good when they vote, they generally reward the incumbent, the embodiment of the status quo, he says.

To test this theory, Malhotra and his colleagues did a parallel experiment involving NCAA basketball fans. They asked people who identified themselves as fans of teams that had advanced to the NCAA championship tournament's later rounds in 2009 to rate Obama's job performance. Result: The further the respondents' teams advanced, the higher their approval of Obama. Another recent study found a relatively high correlation between local sports teams' success and mayoral reelection rates.

Assessing voters' emotional states can, of course, be a tricky business. A new study purporting to show a link between menstrual cycles and women's voting patterns has been questioned - well, trashed - by some researchers; a news article about it was recently removed from CNN.com amid an outcry. Nevertheless, one of the study's authors, Kristina Durante of the University of Texas at San Antonio, says she stands by her findings.

Other behavioral research has plied less-controversial territory. Studies about the effect of a candidate's location on the ballot, for example, stretch back for decades. Generally speaking, having your name listed first seems to convey a marginal but important advantage, especially in elections in which the candidates aren't well-known.

 Where you vote can also influence how you vote, too. People who voted at a poll located within a school in Arizona in 2000, for example, were more likely to support a state initiative to increase funding for education than people who voted elsewhere, according to a 2008 study. The effect wasn't huge - 55 percent of school-based voters supported the measure vs. 53.1 percent overall - but it was significant. It also persisted when the researchers zeroed in on where voters lived, their party identity or demographic factors such as age or sex.

The finding may have implications for Question 6, the Maryland ballot initiative to ratify or reject the state's same-sex marriage law. Just as a school can trigger positive or negative associations within a voter, those deciding Question 6 could be influenced by voting in a church or religious institution, says Jonah Berger, a marketing professor at the University of Pennsylvania who co-wrote the Arizona study.

"I'd be surprised if people weren't affected by it," he says. "My gut leads me to think they'd vote against it, but it's hard to say in aggregate. If the church reminds you about caring for others, about caring for mankind, it could go the other way."

In either case, says Thomas Holbrook, a political science professor at the University of Wisconsin at Milwaukee, emotion is a real and important factor in an election. Holbrook is one of many academics who model elections to predict their outcome. The models typically use variables such as economic growth and presidential-approval polls. But at their core, those aren't much different as measures of sentiment than football games.

"Elections," he says, "are basically about how satisfied people are."