The President, the Press, and the War: A Tale of Two Framing Agendas

Abstract

The alignment between media and presidential framing following 9/11, as well as surrounding the Iraq war, has been criticized as an instance of “when the press fails.” We explore this idea further by comparing presidential and newspaper framing in the case of 9/11 and the subsequent “war on terror.” We argue that high president/press framing alignment after 9/11, and again during the start of the Iraq war, was largely driven by institutional incentives. Thus, “failure” of the press should be expected in these cases, as in the case of other crisis events that yield a strong rally response. Because the media and the president operate under different incentives, they exhibit different framing behaviors—and different framing dynamics. The result is that, in general, the framing messages of these two institutions sometimes align, especially at critical moments, but more often differ. And in the case of major crises like 9/11 and Iraq, we should see a distinct pattern in president/press framing alignment over time—namely, high initial alignment followed by steep decay—as incentives lead the president to “stay the course” while leading news outlets to shift their framing in line with elite and public opinion. We test this idea by applying a new measure of framing alignment to over 3,400 news stories and 500 presidential papers about 9/11 and the war. We find support for our theoretical expectations, showing that, despite their immediate similarities in the cases of 9/11 and Iraq, the president and the press exhibited increasingly divergent framing behaviors over time.

[Supplementary material is available for this article. Go to the publisher's online edition of Political Communication for the following free supplemental resource(s): crisis definition, detailed data and coding descriptions, summary statistics, a sample alignment calculation, and additional figures illustrating frame correlation and media tone]

Keywords:

• framing
• media
• president
• war
• attention dynamics

Acknowledgments

The authors wish to thank Frank Baumgartner, Larry Berman, Rosalee Clawson, Jane Cramer, Matt Eshbaugh-Soha, Regula Hänggli, Kent Jennings, Kyle Joyce, Regina Lawrence, Patrick Merle, Jeff Peake, Walt Stone, Rens Vliegenthart, Gadi Wolfsfeld, and anonymous reviewers for helpful comments on previous versions; John Woolley and Gerhard Peters for generous access to the Presidential Papers archive at the University of California, Santa Barbara; and the many student coders on this project, especially Scott Huffard. Support for the New York Times data collection and coding was provided by National Science Foundation Grant SES-0617492.

Notes

1. We use the term “war on terror” for simplicity, recognizing that the phrase is itself an example of framing (see Reese & Lewis, 2009, for a discussion of how this label was coined by the administration and adopted and internalized by the press and public; see also Snow, Rochford, Worden, & Benford, 1986).

2. We use the term “the media” (and, interchangeably, “the press”) again for simplicity, recognizing that while our study focuses on newspaper coverage, “the media” is multifaceted.

3. See our online Supporting Information document for discussion of how we define a crisis.

4. Snow et al. (1986) use the term “framing alignment” to refer to the linkage of individual and social movement organization interpretive orientations (p. 464), whereas we use it to refer to similarities in how different actors (or institutions) frame a given policy issue.

5. For emerging research on the importance of tone, see Eshbaugh-Soha and Peake (2010).

6. Consider also the Patriot Act as a (highly imperfect) natural experiment. With widespread bipartisan support and scant media criticism (Abdolian & Takooshian, 2003), it easily passed the House and the Senate and was signed into law barely 6 weeks after 9/11. Over time, however, this same piece of legislation came to be seen in a much different light and, by 2011, it was renewed for only an additional 3 months (Lengell, 2011).

7. Although beyond the scope of this article, questions of strategy naturally arise when comparing the framing behaviors of these two institutions. The president wants the press to communicate his or her chosen frames, and the press has a complicated relationship with elected officials (Bennett et al., 2007).

8. Presidency scholars typically also include making good policy in the list of motivations driving presidential decisions (Light, 1999), but we focus here on reelection and historical legacy, viewing good policy as a means to these two ends.

9. Although a changing media environment does mean this attention is dwindling; see Baum and Kernell (1999), Lowry (1997), Young and Perkins (2005), and Cohen (2008).

10. We believe this general pattern holds, but in the notoriously idiosyncratic field of presidential studies (Barilleaux, 1984) we do expect variation in framing behavior across presidents (Rozell, 1995) and personal governing styles (Kumar, 2007).

11. Although work on indexing suggests that elite support may be a necessary condition, we expect that in a crisis scenario characterized by either strong public support or strong elite support—but, for whatever reason, not both—either of these rally markers would be enough to prompt high president/press alignment. But then, we find it hard to imagine a major crisis that would elicit such different responses from elites and citizens.

12. The NYT and WSJ coverage correlated highly, in terms of both amount of attention and frames used. Because of this high correlation and to maximize the sample size, we combine the NYT articles and WSJ abstracts (which we call stories, collectively) in our analysis.

13. See our online Supporting Information document for additional details on our study, including discussion of our data sources and collection procedures, Presidential Papers coding examples, correlations between our NYT and WSJ data sets, and a step-by-step demonstration of how we calculated president/press framing alignment.

14. In fact, in another project (Boydstun & Glazier, n.d.) we find similar patterns in media coverage for Hurricane Katrina, providing additional support for the idea that the crisis framing cycle is generalizable.

15. Our complete codebook—available upon request—includes more than 200 specific frames encompassed within the 12 frame dimensions. Because our analysis here deals exclusively with data at the dimension level, we use the terms “frames” and “frame dimensions” interchangeably.

16. Since we hypothesize that the press will be more dynamic than the president in frame change over time and more varied in the use of different frames overall, limiting each news story to one primary frame (as compared to multiple frame codes where applicable in the case of the Presidential Papers) offers a stricter test of these hypotheses.

17. For the NYT and WSJ stories, intercoder coding into the 12 dimensions specified correlated at an average of 0.96; intercoder reliability for presidential speeches was 0.92.

18. A random sample of our data shows that President Bush used a negative tone only 1% of the time.

19. As shown in Supporting Information Figures SI2 and SI3, the data are similar when using a percentage-based version of this net positive tone measure, calculated by subtracting the percentage of stories in each month that were negative from the percentage that were positive.

20. Excellent studies of framing competition and deliberation include Chong and Druckman (2007b) Entman (2003), Jerit (2008), Porto (2007), and Simon and Xenos (2000).