Voter madness

Whether politicians win or lose may come down to how local athletes play the game. When local football and basketball teams were victorious, voters were more pleased with elected officials, a study appearing online July 6 in the Proceedings of the National Academy of Sciences finds. The capricious link between sports teams and politicians’ performance is a clear example of how irrelevant events can shape important judgments.  

The idea that emotions from unrelated events spill over into other areas isn’t new, says study coauthor Neil Malhotra of the Stanford Graduate School of Business. Lab studies have found that in the afterglow of a free gift, people rate their cars and televisions more highly, for instance.

“There is a lot of evidence of the predictable irrationality of human beings,” Malhotra says. “The question is, does this stuff actually happen in the real world?”

So Malhotra and his colleagues tallied up the wins and losses of 62 Division I college football teams from 1964 through 2008 and found how voters in that team’s home county behaved. The study excluded the University of Connecticut and University of South Florida, which are relative newbies to Division I status, and excluded the University of Southern California and UCLA because they share Los Angeles county.

A local football team’s win in the 10 days before an election garnered the incumbent senator, governor or president (or his or her political party) an extra 1.61 percentage points of the vote, the researchers found. They found no effect for games played earlier than two weeks before the election, suggesting that the game must be fresh in the voter’s mind to have an effect.

The find is “a pretty arresting result,” Malhotra says. The extra points for incumbents from counties with winning teams means that voters are basing their judgments on “their mood and feelings rather than analyzing the data,” he says.

Political scientist Herb Weisberg of Ohio State University in Columbus says that while the study is statistically sound and based on interesting logic, it didn’t adjust for the overall political leanings of a county. “They find that the vote in a county is 1 or 2 percent more favorable to the incumbent’s party when a local team wins, but what if the whole state is 3 or 4 percent more favorable to the incumbent’s party when that team wins?”

In a second analysis, the researchers surveyed over 3,000 people at three times during the 2009 NCAA college basketball tournament. Respondents were asked to name their favorite team and then were asked to rate the performance of President Obama. On average, people whose favorite teams had just won a March Madness game rated the president 2.3 percentage points higher than did those whose teams had recently lost.

The researchers also found that the importance of these irrelevant events was shattered when they were pointed out. When the respondents were explicitly told about the results of the basketball game before they were asked to judge the president’s job performance, the effect disappeared completely, Malhotra and his colleagues found. “Making people more aware of these biases is how to counteract them,” Malhotra says.

Pointing out subtle effects like these and learning how to eliminate them may ultimately help people process information in a more reasoned manner, Malhotra says. “Just because we’re looking at college football doesn’t mean [the research is] trivial.”

The same principle at work in the new study could help explain other phenomena, too, Weisberg says. For instance, the results could explain why a good economy leads people to vote for the incumbent.

“But what are the limits of this logic?” Weisberg asks. “Should victories by pro football teams also affect voting in the area in which the team plays? What about victories by high school football teams? Would Obama have gone up in the public opinion polls if the U.S. had won the World Cup?”