Study points to bias in U.S. media

By David M. Smith

If you heard an American politician spewing such phrases as “tax relief for working Americans,” “the global war against Islamofascism,” and “the repeal of the death tax,” would you be able to determine his party identification? What if he instead emphasized “tax breaks for the wealthy,” “the quagmire in Iraq,” and the “estate tax?”

Furthermore, would you be able to identify a particular news outlet by the language it used? Is there a difference between the rhetoric on Fox News and CNN, for instance, and would you only watch a station that adheres to your views?

The latter question is part of the research problem that guided Jesse Shapiro, a Becker Fellow at the Becker Center on Chicago Price Theory, and Matthew Gentzkow, an assistant professor of economics in the U of C Graduate School of Business (GSB), in a recent paper they wrote on media bias.

Their paper, “What Drives Media Slant? Evidence from U.S. Daily Newspapers,” examines how consumer preference influences newspaper owners to tailor their rhetoric. Something so subjective and slippery as “media bias” may seem hard to quantify, especially when conservatives and liberals often criticize perceived biases in the same news outlet.

In their research, Shapiro and Gentzkow solved this conundrum by compiling a list of the 1,000 most commonly used phrases from both parties in the 2005 Congressional Record. By looking at the occurrence of these partisan phrases in newspapers, the researchers obtained a rough picture of what party a newspaper might belong to if it were a congressperson.

Their findings concerning particular newspapers tended to confirm common perceptions: The Wall Street Journal and The Washington Times bend toward the right, while The New York Times and the San Francisco Chronicle lean more toward the left.

What’s surprising about Shapiro and Gentzkow’s study is its analysis of the market forces that influence media bias. For instance, they found that newspapers with a Republican slant circulated more widely in zip codes that have high concentrations of registered Republican voters, which implies that any significant deviation from the party line would lead to a drop in demand.

Consumer preference therefore plays a bigger role in determining the partisan content of a publication than the political leanings of its owners. The findings, Shapiro said, show that “consumers care about slant.”