Political news in the United States is no longer dominated by the nightly newscasts of three broadcast networks committed to a common set of values like journalistic objectivity. Contemporary cable and satellite television brings, among other options, a variety of partisan public affairs programs into American homes, primarily on the conservative Fox News Channel and progressive MSNBC. The Internet provides readers an even broader array of political commentary. Does this media fragmentation deepen the political and social divides of the American public? In his thoughtful and thorough book, Matthew Levendusky finds that it does, or at least can.

Levendusky first demonstrates that Fox News and MSNBC provide meaningfully different content. Most casual observers expect that this is the case, but Levendusky provides an accessible, rigorous comparison of how the two channels covered stories like the 2011 debt ceiling debate and President Obama’s support for same-sex marriage recognition in May 2012. He finds these channels focused on only one side of these and other issues, offered extensive criticism of opposition views, and provide clear cues on which political candidates their viewers should support.

Using a series of experiments that expose study participants to segments from Fox News and MSNBC shows and Public Broadcasting System’s News Hour as a neutral source, he finds a variety of partisan news effects. People exposed to like-minded programs (e.g., a Democrat shown a clip from MSNBC) become more extreme in their political views (p. 75), more hostile to the political party they oppose (p. 98), less open to bipartisan compromise (p. 104), and more skeptical about the legitimacy of their own party’s electoral defeat (p. 131). He also shows the extremity affected by a like-minded political program endures at least two days (p. 85) and is increased with exposure to a second like-minded program (p. 90).

Levendusky’s experiments are designed with care. He uses both a pre-test/post-test experimental design to show how individuals are themselves affected by an experimental treatment, and post-test-only studies to address the concern that the pre-test itself could influence his findings. He uses multiple sources of subjects for his studies, including a university student sample and adult participants recruited online.

He also tries to clarify which segments of the partisan media audience are more affected by it. He finds that people who prefer to watch like-minded partisan media, given the choice among different types of news, are most influenced by it. He argues that this is due to motivated reasoning among these viewers, seeking biased information and using it to further reinforce their biased perspectives (e.g., p. 83). Consequently, Levendusky argues that partisan media polarizes not by making moderates more extreme but by pushing extremists farther away from the political center.

He complements these experimental findings with an analysis of self-reported partisan media use by respondents to the 2008 National Annenberg Election Study. He finds, for example, that people who said they view specific partisan talk shows were 5 percent more likely than non-viewers to vote for their political party’s candidate (p. 118). Levendusky is clear to note a major limitation of this research design: because the media-use data is self-reported rather than observed in an experiment, he is unable to void the analysis of all potential self-selection effects, which could undermine his findings. This is a limitation inherent to virtually all media-effects research using public-opinion survey data. However, he is careful to use both statistical matching procedures that allow him to compare politically and demographically similar survey participants who differ primarily on whether or not they watch partisan media and a fixed-effects model to investigate how individuals change across waves of the survey.

The deleterious effects of partisan media do not end with their direct effects on people who watch them. Levendusky offers a preliminary demonstration of “agenda bleed” (p. 150), the influence of partisan media on the stories covered by mainstream news. He shows that nonpartisan news sources such as CNN and The Washington Post began to cover controversies surrounding former White House Special Advisor Van Jones after Fox News hosts increased their attention to Jones, his personal history, and political views (p. 47). Jones ultimately resigned as Obama’s green-jobs advisor under mounting Congressional pressure.

Levendusky concludes that partisan media play an important role in contemporary U.S. politics: “they may increase the difficulty of governing in contemporary America” (p. 151). Others are more skeptical of the partisan media as a cause of mass polarization (e.g., Arceneaux and Johnson 2013, Ladd 2012). The influence of Fox News and MSNBC is limited both by the choices viewers make to watch other types of programs, as well as by the characteristics of people who choose to watch partisan news. Partisan news viewers are generally more interested in politics, have more knowledge of it and, we argue, more stable opinions.

Nonetheless, I agree with many of Levendusky’s normative points and am quite satisfied by much of his empirical work. The rhetorical and reportorial indulgences of partisan cable news are disquieting. Partisan media at least mechanically polarize viewers by allowing them to sort themselves into partisan audiences (e.g., Stroud, 2011). Further, partisan media do affect people who are directly exposed to them, even as the broader effect of partisan media on the US population is diluted by the narrowness of the audiences for these shows. Further, we agree that partisan media appear to have an indirect effect on politics by influencing the agenda of mainstream news.

My main theoretical disagreement with Levendusky lies with the persuadability of regular viewers of partisan media. He argues, soundly, that these viewers are more persuadable because of their directional goals in information processing—they want reinforcement for their views. However, regular viewers of partisan media are also likely to have firmly held pre-existing opinions and more knowledge of politics, making them generally less persuadable than less regular viewers. Future research will need to sort out the relative importance of these competing characteristics of the people most likely to watch partisan news.

Levendusky’s research designs imply useful directions for future work on the effects of changing media in the United States (see also Feldman, Stroud, Bimber, and Wojcieszak 2013). His experiments offer viewers a choice only among different types of news. This is a reasonable starting point; after all, he is interested in political communication. However, it optimistically assumes that everyone is, on some level, a news consumer. One alternative approach would allow people to express preferences for entertainment programs in addition to news choices (e.g., Arceneaux and Johnson 2013). Levendusky also focuses on “moderately salient issues” (p. 68), which he anticipates are more prone to media effects than better-known issues (p. 139). In studies of media selectivity, the choices offered audience members are likely to affect what researchers infer about media effects. Similarly, partisan news viewers may be more influenced on some issues than others.

These points notwithstanding, Levendusky’s How Partisan Media Polarize America is an excellent addition to a growing research literature on the changing media environment in the US and the effects of this new mass media. Students of media effects, contemporary partisan politics, and political polarization will find it of great value.