Media Bias - Scholarly Treatment of Media Bias in The United States and United Kingdom

Scholarly Treatment of Media Bias in The United States and United Kingdom

Media bias is studied at schools of journalism, university departments (including Media studies, Cultural studies and Peace studies) and by independent watchdog groups from various parts of the political spectrum. In the United States, many of these studies focus on issues of a conservative/liberal balance in the media. Other focuses include international differences in reporting, as well as bias in reporting of particular issues such as economic class or environmental interests.

One high profile academic survey of American journalists is The Media Elite. The survey found that most journalists were liberal Democratic voters.

Martin Harrison's TV News: Whose Bias? (1985) criticized the methodology of the Glasgow Media Group, arguing that the GMG identified bias selectively, via their own preconceptions about what phrases qualify as biased descriptions. For example, the GMG sees the word "idle" to describe striking workers as pejorative, despite the word being used by strikers themselves.

Herman and Chomsky (1988) proposed a propaganda model hypothesizing systematic biases of U.S. media from structural economic causes. They hypothesize media ownership by corporations, funding from advertising, the use of official sources, efforts to discredit independent media ("flak"), and "anti-communist" ideology as the filters that bias news in favor of U.S. corporate interests.

Many of the positions in the preceding study are supported by a 2002 study by Jim A. Kuypers: Press Bias and Politics: How the Media Frame Controversial Issues. In this study of 116 mainstream US papers (including The New York Times, the Washington Post, Los Angeles Times, and the San Francisco Chronicle), Kuypers found that the mainstream print press in America operate within a narrow range of liberal beliefs. Those who expressed points of view further to the left were generally ignored, whereas those who expressed moderate or conservative points of view were often actively denigrated or labeled as holding a minority point of view. In short, if a political leader, regardless of party, spoke within the press-supported range of acceptable discourse, he or she would receive positive press coverage. If a politician, again regardless of party, were to speak outside of this range, he or she would receive negative press or be ignored. Kuypers also found that the liberal points of view expressed in editorial and opinion pages were found in hard news coverage of the same issues. Although focusing primarily on the issues of race and homosexuality, Kuypers found that the press injected opinion into its news coverage of other issues such as welfare reform, environmental protection, and gun control; in all cases favoring a liberal point of view.

Studies reporting perceptions of liberal bias in the media are not limited to studies of print media. A joint study by the Joan Shorenstein Center on Press, Politics and Public Policy at Harvard University and the Project for Excellence in Journalism found that people see liberal media bias in television news media such as CNN. Although both CNN and Fox were perceived in the study as being left of center, CNN was perceived as being more liberal than Fox. Moreover, the study's findings concerning CNN's perceived liberal bias are echoed in other studies. There is also a growing economics literature on mass media bias, both on the theoretical and the empirical side. On the theoretical side the focus is on understanding to what extent the political positioning of mass media outlets is mainly driven by demand or supply factors. This literature is surveyed by Andrea Prat of the London School of Economics and David Stromberg of Stockholm Yniversity.

According to Dan Sutter of the University of Oklahoma, a systematic liberal bias in the U.S. media could depend on the fact that owners and/or journalists typically lean to the left.

Along the same lines, David Baron of Stanford GSB presents a game-theoretic model of mass media behaviour in which, given that the pool of journalists systematically leans towards the left or the right, mass media outlets maximise their profits by providing content that is biased in the same direction. They can do so, because it is cheaper to hire journalists that write stories which are consistent with their political position. A concurrent theory would be that supply and demand would cause media to attain a neutral balance because consumers would of course gravitate towards the media they agreed with. This argument fails in considering the imbalance in self-reported political allegiances by journalists themselves, that distort any market analogy as regards offer: (...) Indeed, in 1982, 85 percent of Columbia Graduate School of Journalism students identified themselves as liberal, versus 11 percent conservative" (Lichter, Rothman, and Lichter 1986: 48), quoted in Sutter, 2001.

This same argument would have news outlets in equal numbers increasing profits of a more balanced media far more than the slight increase in costs to hire unbiased journalists, notwithstanding the extreme rarity of self-reported conservative journalists (Sutton, 2001).

As mentioned above, Tim Groseclose of UCLA and Jeff Milyo of the University of Missouri at Columbia use think tank quotes, in order to estimate the relative position of mass media outlets in the political spectrum. The idea is to trace out which think tanks are quoted by various mass media outlets within news stories, and to match these think tanks with the political position of members of the U.S. Congress who quote them in a non-negative way. Using this procedure, Groseclose and Milyo obtain the stark result that all sampled news providers -except Fox News' Special Report and the Washington Times- are located to the left of the average Congress member, i.e. there are signs of a liberal bias in the US news media. However, the news media also show a remarkable degree of centrism, just because all outlets but one are located –from an ideological point of view- between the average Democrat and average Republican in Congress.

The methods Groseclose and Milyo used to calculate this bias have been criticized by Mark Liberman, a professor of Computer Science at the University of Pennsylvania. Liberman concludes by saying he thinks "that many if not most of the complaints directed against G&M are motivated in part by ideological disagreement – just as much of the praise for their work is motivated by ideological agreement. It would be nice if there were a less politically fraught body of data on which such modeling exercises could be explored."

Sendhil Mullainathan and Andrei Shleifer of Harvard University construct a behavioural model, which is built around the assumption that readers and viewers hold beliefs that they would like to see confirmed by news providers. When news customers share common beliefs, profit-maximizing media outlets find it optimal to select and/or frame stories in order to pander to those beliefs. On the other hand, when beliefs are heterogeneous, news providers differentiate their offer and segment the market, by providing news stories that are slanted towards the two extreme positions in the spectrum of beliefs.

Matthew Gentzkow and Jesse Shapiro of Chicago GSB present another demand-driven theory of mass media bias. If readers and viewers have a priori views on the current state of affairs and are uncertain about the quality of the information about it being provided by media outlets, then the latter have an incentive to slant stories towards their customers' prior beliefs, in order to build and keep a reputation for high-quality journalism. The reason for this is that rational agents would tend to believe that pieces of information that go against their prior beliefs in fact originate from low-quality news providers.

Given that different groups in society have different beliefs, priorities, and interests, to which group would the media tailor its bias? David Stromberg constructs a demand-driven model where media bias arises because different audiences have different effects on media profits. Advertisers pay more for affluent audiences and media may tailor content to attract this audience, perhaps producing a right-wing bias. On the other hand, urban audiences are more profitable to newspapers because of lower delivery costs. Newspapers may for this reason tailor their content to attract the profitable predominantly liberal urban audiences. Finally, because of the increasing returns to scale in news production, small groups such as minorities are less profitable. This biases media content against the interest of minorities.

Jimmy Chan of Shanghai University and Wing Suen of the University of Hong Kong develop a model where media bias arises because the media cannot tell “the whole truth” but are restricted to simple messages, such as political endorsements. In this setting, media bias arises because biased media are more informative; people with a certain political bias prefer media with a similar bias because they can more trust their advice on what actions to take.

The economics empirical literature on mass media bias mainly focuses on the United States.

Steve Ansolabehere, Rebecca Lessem and Jim Snyder of the Massachusetts Institute of Technology analyze the political orientation of endorsements by U.S. newspapers. They find an upward trend in the average propensity to endorse a candidate, and in particular an incumbent one. There are also some changes in the average ideological slant of endorsements: while in the 40s and in the 50s there was a clear advantage to Republican candidates, this advantage continuously eroded in subsequent decades, to the extent that in the 90s the authors find a slight Democratic lead in the average endorsement choice.

John Lott and Kevin Hassett of the American Enterprise Institute study the coverage of economic news by looking at a panel of 389 U.S. newspapers from 1991 to 2004, and from 1985 to 2004 for a subsample comprising the top 10 newspapers and the Associated Press. For each release of official data about a set of economic indicators, the authors analyze how newspapers decide to report on them, as reflected by the tone of the related headlines. The idea is to check whether newspapers display some kind of partisan bias, by giving more positive or negative coverage to the same economic figure, as a function of the political affiliation of the incumbent President. Controlling for the economic data being released, the authors find that there are between 9.6 and 14.7 percent fewer positive stories when the incumbent President is a Republican.

Riccardo Puglisi of the Massachusetts Institute of Technology looks at the editorial choices of the New York Times from 1946 to 1997. He finds that the Times displays Democratic partisanship, with some watchdog aspects. This is the case, because during presidential campaigns the Times systematically gives more coverage to Democratic topics of civil rights, health care, labor and social welfare, but only when the incumbent president is a Republican. These topics are classified as Democratic ones, because Gallup polls show that on average U.S. citizens think that Democratic candidates would be better at handling problems related to them. According to Puglisi, in the post-1960 period the Times displays a more symmetric type of watchdog behaviour, just because during presidential campaigns it also gives more coverage to the typically Republican issue of Defense when the incumbent President is a Democrat, and less so when the incumbent is a Republican.

Alan Gerber and Dean Karlan of Yale University use an experimental approach to examine not whether the media are biased, but whether the media influence political decisions and attitudes. They conduct a randomized control trial just prior to the November 2005 gubernatorial election in Virginia and randomly assign individuals in Northern Virginia to (a) a treatment group that receives a free subscription to the Washington Post, (b) a treatment group that receives a free subscription to the Washington Times, or (c) a control group. They find that those who are assigned to the Washington Post treatment group are eight percentage points more likely to vote for the Democrat in the elections. The report also found that "exposure to either newspaper was weakly linked to a movement away from the Bush administration and Republicans."

Another unaffiliated group, Media Study Group, established seven categories of poor journalistic practice: for example, the journalist stating personal opinion in a report, asserting incorrect facts, applying unequal space or treatment to two sides of a controversial issue; then analyzed The Age Newspaper (Melbourne Australia) for the frequency of infraction of this code of practice. The resultant instances were then analyzed statistically with respect to the frequency they supported one or other side of the two-sided controversial issue under consideration. The goal of this group was to establish a quantitative methodology for the study of bias.

A self-described liberal media watchdog group, Fairness and Accuracy in Reporting (FAIR), in consultation with the Survey and Evaluation Research Laboratory at Virginia Commonwealth University, sponsored an academic study in which journalists were asked a range of questions about how they did their work and about how they viewed the quality of media coverage in the broad area of politics and economic policy. “They were asked for their opinions and views about a range of recent policy issues and debates. Finally, they were asked for demographic and identifying information, including their political orientation”. They then compared to the same or similar questions posed with “the public” based on Gallup, and Pew Trust polls. Their study concluded that a majority of journalists, although relatively liberal on social policies, were significantly to the right of the public on economic, labor, health care and foreign policy issues.

This study continues: “we learn much more about the political orientation of news content by looking at sourcing patterns rather than journalists' personal views. As this survey shows, it is government officials and business representatives to whom journalists "nearly always" turn when covering economic policy. Labor representatives and consumer advocates were at the bottom of the list. This is consistent with earlier research on sources. For example, analysts from the centrist Brookings Institution and conservative think tanks such as the Heritage Foundation and the American Enterprise Institute are those most quoted in mainstream news accounts; liberal think tanks are often invisible. When it comes to sources, ‘liberal bias’ is nowhere to be found.”