Professors Find Bias in Associated Press College Football Poll

Joanna Norris, Assistant Director

Department of Media Relations and Events

(904) 620-2102

A new study by four University of North Florida professors finds that voters in the college football media poll are biased in favor of teams and conferences from their home state and toward selected Bowl Championship Series (BCS) conferences.

The UNF Coggin College of Business professors evaluated the weekly ballots submitted by 65 Associated Press (AP) voters in 2007. Drs. Jay Coleman, Andres Gallo, Paul Mason and Jeff Steagall found evidence of geographic bias toward teams and conferences in the voter’s state—and that such geographic bias is state-related and not distance-related.

The professors also found evidence of excessive favor toward teams in the SEC, Big 12, and PAC-10 (and against teams in the Big East, and to a lesser extent the Big 10), as well as toward teams that played on television, particularly on prominent^{networks. Additionally, they found that voter ballots were too heavily related tosimplistic performance indicators such as the number of losses andlosing in the preceding week. The business professors did not find evidence of a so-called East-coast bias by AP voters.}

“The media often heavily criticize the so-called computer rankings—particularly when they differ from the media’s judgment—as well as the coaches’ poll, with its presumed self-serving biases,” said Coleman. “And both of these are used by the BCS, which amplifies the rhetoric. However, our findings indicate that there is ample evidence that the media perspectives are hardly unbiased either.”

Although followers of college football have long surmised that many such biases exist, no previous study has ever formally evaluated whether the group noted above—particularly those involving geographic bias—are indeed true. However, the AP’s public release of individual voter ballots, which started in 2006, finally provided enough detail in the voting data to allow specific analysis of individual voter behavior.

“Our results have significant managerial ramifications for the selection and distribution of voters by the AP, and whether the champions so designated would have been the same without such bias. To the extent that similar biases may have existed in other seasons, it also calls into question the BCS’s previous use of the AP poll,” stated Coleman.

The study by the UNF professors will soon be published in the Journal of Sports Economics. Go to jse.sagepub.com/cgi/content/abstract/1527002509346823v1 to read an abstract of the article.

-UNF-