Monday, December 19, 2005

UCLA Study Finds That the MSM Tilts Left

Frequently asserted but rarely substantiated, the question of whether or not American media outlets operate under the influence of pervasive liberal bias is an old blogosphere staple. Today I learned of recent research that claims to show that most news organizations actually do lean to the left of the American mainstream. The study's authors, Tim Groseclose at UCLA and Jeffrey Milyo at the University of Missouri at Columbia, developed an innovative method of quantifying bias in the news: comparing the think tank citation patterns of lawmakers to those of media outlets. They started with ideology ratings provided by the liberal interest group Americans for Democratic Action, which assigns each senator and House member a score between 0 and 100 (with higher numbers indicating increased liberalism) based on their legislative record. Groseclose and Milyo then counted the number of times each Congress member cited a given think tank (such as the Brookings Institution or the American Enterprise Institute) and did the same for each news source. Outlets and Congresspeople with similar citation patterns were assumed to share the same ADA score; e.g. since both the New York Times and Joe Lieberman were found to cite from very similar bodies of sources, the theory dictates that the paper and the senator would be comparably liberal.

Groseclose and Milyo's methodology is certainly non-intuitive, but how valid is it? The idea that the ideological tilt of politicians and news providers alike can be divined from the sources they cite seems logical, since liberal politicians tend to cite liberal think tanks and vice versa for conservatives. But I have two comments: one, Groseclose and Milyo's theory predicts that major news organizations should discuss the ideas of liberal think tanks disproportionately in their articles. This is a hypothesis that could be investigated without much difficulty, since they conveniently rate the top 20 most-frequently-cited policy organizations at the end of this older paper on the same topic. Secondly, the authors fail to account for what could be a gaping hole in their theory—the possibility that many political news articles may have been left out of the analysis because they cited no think tanks at all. It's possible that for some news organizations (particularly TV news), the ratio of stories-with-citations to total stories might be too small to properly extrapolate conclusions about the former to the latter.

I have no idea though; I only skimmed the earlier paper, and all I have is a press release for the new research. Interesting stuff anyway.