The fragmentation of the mass media audience has led scholars to examine the growing partisanship of news consumers as less partisan individuals turn to more entertainment-oriented fare. An analysis of national survey data from 1998 to 2006 suggests that not only has the news audience grown somewhat more partisan over time, but among these partisans a distinct migration has occurred to sources more likely to be friendly to individual political beliefs or away from sources perceived as less friendly.
The fragmented media marketplace provides a host of communication channels from which individuals can learn of the day's events: some serious, some silly, and many laced with potent ideological and partisan content.1 In the last few decades, newspapers and broadcast television news programs faced growing competition from such sources as cable news networks, religious and talk radio programs, personality magazines, late-night talk shows, parodies such as The Daily Show with Jon Stewart, and of course the dizzying number of sites available on the Internet. In addition, the rapid diffusion of cable television and the Internet in U.S. homes has drawn precious audience away from news content to a huge buffet of entertainment programming.2
Changes in the mass media resemble those seen in day-to-day conversations. Scholars have noted a trend in the balkanization of interpersonal communication networks, with people more likely to speak about public affairs with individuals much like themselves, thus reducing their opportunities to hear the kind of divergent viewpoints deemed integral to a healthy democracy.3 The mainstream news media, once perceived as a remedy for these homogenous interpersonal networks, appear to be evolving into a similar information environment, one likely to lead to greater polarization among the U.S. electorate.4 Using national survey data gathered from 1998 to 2006, this study examines whether political partisans make up an increasingly larger share of the news audience as compared to less partisan respondents, and whether among these polarized partisans there is a fragmentation in the audience toward sources of news either likely to be more friendly to their beliefs or away from sources seen as unfriendly.
Audience Polarization and Selective Exposure
Two related but distinct theoretical approaches drive this research. The first is rooted in the classic studies of selective exposure and why people seek out or attempt to avoid certain types of information. The second is the approach characterizing more recent work in polarization which, in part, seeks to explain the perceived shift to a "red state-blue state" U.S. electorate and the likely consequences of such a split, both for democracy and the media.
In studies of the latter, scholars point to the technological shift from a low-choice to a high-choice environment that allows people to increasingly customize their media diet. In the late 1970s, over 90% of Americans watched the three major broadcast networks.5 The introduction of cable (and, later, satellite) television dramatically increased the choices available to an average consumer to well over 100 channels.6 While many praised this rise in consumer choice, others were more cautious. Turow warned that advertisers and media companies will create the "electronic equivalent of gated communities," while Sunstein cautioned against the demise of shared experiences and group polarization.7 In his analysis, Prior argued that news consumption patterns over time changed "not so much because people are different today, but rather because the media environment is different. People have not necessarily changed; they have merely changed the channel. And they would have done it sooner, had they been given the chance."8 The shift from a low-choice to high-choice environment had positive and negative effects for democracy.' For some, these new media choices resulted in improved political knowledge and involvement; for many, however, the opportunity to "change the channel" meant less inadvertent, chance exposure to news content, the result being less knowledge and, therefore, less likelihood of voting and other forms of political participation. …