What do investors and college football pollsters have in common?

In a study of more than 20 years of college football polls, an economist at The Ohio State University and his colleagues found that poll voters reward teams that perform just how the voters expected they would.

"As a college football team, if you just meet expectations, you are rewarded by the poll voters. If you are expected to win in a blowout, the voters think you ought to win in a blowout," said Trevon Logan, an associate professor of economics at Ohio State.

The research found that poll voters don't take into account strength of schedule or margin of victory when deciding which college football teams are the best.

What is important is winning and meeting expectations, which in college football games means beating teams by the expected margin, Logan said. But going over that margin is not especially helpful.

This year, Florida State University climbed from #11 in the preseason Associated Press poll to #1 this past week. They did it by going undefeated and beating outmatched teams handily, as they were expected to do.

Ohio State, on the other hand, has suffered in the polls some because it has not met expectations, with some closer-than-expected victories on the way to an unbeaten season.

"But why is Florida State climbing in the polls when they are doing exactly what you expected them to do? In some ways, it doesn't make sense," Logan said.

The tendency of pollsters to reward teams that meet their expectations is an example of a cognitive bias called confirmatory bias.

"Confirmatory bias is the tendency of people to accept weak evidence if it affirms beliefs they already hold," Logan said.

Logan said Wall Street investors show the same type of bias. When analysts expect a company's profits to increase 3 percent in a quarter, and the company meets that target, it is often rewarded with higher stock prices. But if that company increases profits for the quarter -- but only by 2 percent -- stock prices may fall.

"The company may have increased profits, but that doesn't matter if analysts expected more than the company delivered," he said. "That's an example of how confirmatory bias can work against a company."

Logan said his research shows that college football poll voters, despite their expertise, are subject to the same cognitive biases as the rest of us.

"No matter how much they know or how much they learn, human voters will always be subject to human biases," he said.

Logan conducted the study with Rodney Andrews of the University of Texas at Dallas and Michael Sinkey at the University of West Georgia. Their results were published as a National Bureau of Economic Research working paper.