|
Article Excerpt
"We [Democrats] stayed off FOX for a long time because your news department is, in fact, biased ... there are some things in the news department that have really been shockingly biased, and I think that's wrong. And I'll just say so right up front."
--Howard Dean, appearing on Fox News Sunday, May 4, 2008
"I think Fox News has come on the scene and identified itself as 'fair and balanced.' We try to do that every day. I think others, instead of trying to get more fair and balanced, probably are offended by that or worried about it ... What they're trying to do is say that Fox News is mixing opinion and fact. That's just simply not true ... Bias can be a lot of different ways--story selection, story placement, story emphasis ... I looked at other people's polls, national polls, and most of the people thought the news was either biased or boring or both. And they generally thought it was biased in one direction."
--Fox News Chairman Roger Ailes, interview with C-SPAN's Brian Lamb, 2004
For many decades, Republicans have repeated a mantra alleging that the media as a whole are biased against their party and their politicians. This attitude toward the media was perhaps best exemplified by a popular 1992 Republican bumper sticker that said, "Annoy the Media: Re-elect George Bush."
However, with the rise of new media such as conservative Web sites, talk radio, and especially the Fox News Channel, Republicans have seen Democrats begin to embrace and extend their complaints of bias in the news. (1) In the run-up to the 2008 elections, those complaints only increased in volume and ferocity. In early 2007, these complaints actually affected the conduct of the campaign when liberal activists pressured the Nevada Democratic Party to cancel a Fox-sponsored Democratic candidate debate. In launching the successful campaign to drop Fox as a debate sponsor, liberal blogger Chris Bowers of MyDD.com argued that, "... instead of giving [Fox] a golden opportunity to further distort the image of Democratic presidential candidates, and instead of providing them with credibility for all of their past and future attacks against Democrats, it would be best if the Nevada Democratic Party chose a different media partner to broadcast this debate" (Bowers 2007). (2)
In this study, I will attempt to empirically measure whether Fox News has, in fact, systematically skewed its news over the past decade and compare its news choices to those of the network evening newscasts. Specifically, I will be examining whether Fox's Special Report, ABC's World News (Tonight), the CBS Evening News, and the NBC Nightly News presented biased portrayals of public opinion regarding the president in their coverage.
Empirically Examining Media Bias
Claims of media bias raised by politicians from either party should be regarded as exceptionally suspect for several reasons. First, politicians might prefer that a news source be perceived as biased against them, even if the source is actually unbiased. As Matthew Baum and I demonstrate elsewhere (Baum and Groeling forthcoming), when members of the public perceive the news to be biased against a candidate or party, harmful messages from that outlet are discounted, while favorable messages are seen as particularly credible. Similarly, partisans might strategically choose to allege bias--even in the absence of such bias--in an attempt to "work the ref"--that is, vociferously protest a close call in an attempt to have the next one go your way. And, because of well-documented cognitive biases--such as confirmation and disconfirmation biases, selective perception, anchoring, attention bias, the clustering illusion, and selective perception, among others--partisans might sincerely perceive news as being biased against their preferred stance, even when it is actually unbiased (see Hastorf and Cantril 1954; Dalton et al. 1998; Baum and Groeling n.d.) The possibility that perceptions of bias rest in the eye of the beholder is not lost on journalists, who readily turn to that explanation to blunt charges of favoritism. In so doing, they typically echo longtime CBS anchor Walter Cronkite's aphorism that, "Our job is only to hold up the mirror--to tell and show the public what has happened. Then it is the job of the people to decide whether they have faith in their leaders or governments" (quoted in Alan and Lane 2003, 139-40).
Establishing the presence or absence of partisan bias in news content has proven difficult. Self-described media watchdog groups such as Media Matters, the Media Research Center (MRC), the Center for Media and Public Affairs (CMPA), and Fairness and Accuracy In Reporting (FAIR) claim to objectively analyze media content, yet they routinely disagree on the incidence, severity, and direction of bias in the media. (3)
Most scholarly attempts to assess media bias are similarly inconclusive (e.g., Efron 1971; Patterson 2003; Sutter 2001). Among the principal difficulties in demonstrating the presence or absence of media bias is establishing a clear definition of what exactly constitutes bias. Several recent studies (Groseclose and Milyo 2005; Gentzkow and Shapiro 2006) have sought to empirically measure mainstream news media content against various standards and have done so with varying results. However, perceiving an ideological slant in media content is one thing; attributing such a slant to politically biased editorial judgment by the media is another. After all, as journalists are quick to point out, the observed patterns of coverage might simply represent a fair reflection of reality. For example, if one observes that 90% of the stories appearing on a TV news program were bad news for the president, that could reflect biased story selection by that program, or it may simply reflect the president only doing a good job 10% of the time. In such a case, this hypothetical 90% anti-presidential skew in media coverage would represent an accurate reflection of the president's performance.
Thus, two classes of problems confound bias research. The first concerns the inherent subjectivity of bias. As has been discussed above, identical stories are often perceived as having diametrically opposed biases depending on who is viewing them (Vallone, Ross, and Lepper 1985). While scholars do attempt to use rigorous standards and procedures for coding the content of the news, ultimately these procedures will confront the researcher with seemingly arbitrary choices, which subsequently expose the results to criticism from those who find them disagreeable. For example, if President Bush Is reported to be standing by his secretary of defense, who is being pressured to resign, that stance could be interpreted as positive (a man who sticks with his convictions and by his allies) or negative (a man who ignores reality or is foolishly stubborn). In aggregating such codes across multiple stories and outlets, researchers are also forced to combine disparate units. For example, does a negative story showing incompetent disaster response by a trusted official count the same as a positive story about throwing out the first pitch at a baseball game?
The second problem concerns selection bias. By examining only those news stories that are actually reported, one cannot determine whether any observed distribution of favorable and unfavorable presidential news accurately portrays real-world conditions or instead reflects bias on the part of journalists (Hofstetter 1976; Groeling and Kernell 1998; Niven 2002; Baum and Groeling forthcoming; Baum and Groeling n.d.). If so, research based exclusively on content analysis of reported news commits the fallacy of drawing inferences from data that has been selected on the dependent variable. The issue of selection bias presents scholars with a serious conundrum. How can one assess the representativeness of a news sample when the population comprises mostly stories that were never reported and thereby cannot be observed? To establish bias one must also somehow measure these nonevents. (4)
To address these concerns, I return to a method Sam Kernell and I first used a decade ago to study negative bias in network news (Groeling and Kernell 1998). In that research, we minimized these methodological and evidentiary problems by limiting our analysis to stories reporting the public's assessment of the president's job performance. Such polls represented a subset of presidential news that minimized subjective coding and for which we could observe the population of potential news stories as well as those actually reported. Containing quantitative information about the current state of public opinion, they were particularly well suited for formulating defensible definitions of good and bad presidential news: Decreases in the president's approval rating constituted bad news, and increases, good news. Beyond the direction of shifting public opinion, the percent approving identified the magnitude of the change.
Finally, with approval stories, we could easily identify and measure the population of potential news from which that reported by journalists was chosen. With each network's in-house public opinion surveys publicly archived, we could identify the full population of approval ratings--that is, potential news stories--and compare those selected for broadcast with those that were not. Figure 1 illustrates four general types of patterns one might expect to observe in those decisions. (5)
The first general pattern of poll story selection, Preference for Change, simply predicts that the greater the change in presidential approval a poll shows, the more likely that poll will be aired. Here, news value derives from a departure from the status quo, regardless of whether that departure is favorable or unfavorable to the president (and regardless of the president's party).
[FIGURE 1 OMITTED]
In contrast, the second graphic in Figure 1 illustrates the news pattern politicians often allege the "nattering nabobs of negativism" (i.e., journalists, in the eyes of former Vice President Spiro Agnew) favor: a strong preference for negative news. (6) As Bill Clinton alleged shortly into his first term in office, for journalists, "success and...
|
