The ‘liberal’ media: bias or customer preferences?

Abstract

Allegations of liberal bias against the national news media in the US remain very contentious. Conservative critics of the media argue biased news stems from liberal-leaning journalists. Another explanation of alleged bias could be that liberals have a greater demand for news than moderates or conservatives, so the media simply tailors its product to their customers. I test whether liberals in the US have a greater demand for news by investigating the determinants of newspaper circulation across states and metropolitan areas. I find mixed evidence for liberal demand: circulation is higher in more liberal metropolitan areas but is independent of local political preferences across states.

Acknowledgements

I would like to thank Marc Poitras and an anonymous referee for comments.

Notes

¹ Other studies documenting liberal bias include Irvine (1984), Lichter et al. (1986), Bozell and Baker (1990), Baker (1994), Pines with Lamer (1994) and Graham (1996).

² Many media scholars would argue instead that advertisers are the primary ‘consumers’ of news. Advertisers bundle their messages with news or entertainment to increase the audience for their ads. Advertisers need not bundle their messages with news; they could deliver them directly to consumers via advertising circulars, billboards or infomercials. If they see news as a means to attract audience, advertisers would desire news that would secure the largest audience. Hence, the preferences of consumers comprise the relevant standard against which to evaluate the news market.

³ In the US almost all newspapers enjoy local monopolies; less than a dozen of cities have competing daily newspapers. Consequently, audience maximizing papers should cater to the median of the local preference distribution (Endersby, 2000). Critics on the left who allege the news media of a pro-business, pro-establishment bias blame advertisers for censoring potential anti-business messages. If the news media are judged relative to the preferences of news consumers, the liberal critique of the news also alleged a form of new market failure.

⁴ Except where noted, the source for the state variables is the Statistical Abstract of the United States and the source for the MSA dataset is the 1990 Census.

⁵ I could not find a measure of the average price of newspapers in a state, and since the newsstand prices of papers fall in a relatively small range (and each state has a sample of several papers), I did not attempt to construct one. Thompson (1989) in any event found that price did not significantly affect circulation of British and Irish papers.

⁶ The circulation of the USA Today, a paper with a national circulation, is subtracted from Virginia's total.