A Department of Homeland Security Center of Excellence led by the University of Maryland

Textual Analysis of Electronic Media Coverage of Homeland Security-Related Risks

Textual Analysis of Electronic Media Coverage of Homeland Security-Related Risks


Project Details


As citizens increasingly turn to electronic media (i.e., the internet) as a source of information, the need exists to characterize the dynamics of electronic media coverage in shaping public perceptions of homeland-security-related risks. This study analyzes the dynamics of electronic media coverage of homeland security-related risks. This study uncovers patterns of homeland-security-related risk communication in electronic media following 9/11 in order to assess how internet-based electronic news media frame homeland security-related risks.

Primary Findings: 

This research employed quantitative and qualitative textual analysis of terrorism-related news coverage across electronic (online news and blogs), print, and television media. First, our review of extant research concerning media coverage of terrorism found a need for comparative, multi-modal studies of terrorism reporting and raised questions concerning whether electronic media news patterns resemble those of traditional media sources. Using centering resonance analysis (CRA), a unique data-mining algorithm, coupled with qualitative thematic analysis, we analyzed online, television, and blog discourse of the 2007 Boston Bomb Scare and online and blog discourse surrounding the 2005 London Bombings. Our study of media coverage of the 2007 Boston Bomb Scare found few discernible differences across media channels during the two week life-cycle of news coverage, which focused mainly on fact reporting. However, citizen message boards and blogs framed the incident in terms of 9/11, revealing how suspected acts of terrorism are interpreted in relation to that iconic event. A second study of news discourse across US and UK online newspapers and blogs following the "7/7" London bombings of 2005 examined how images of "resilience" worked to solidify or undermine a sense of national unity and to shape public responses to terrorism. These studies thus reveal how electronic media serve as critical sites for assessing and participating in public discussions about terrorism, and for discerning the contested and contradictory meanings of homeland security. Recommendations for policymakers implied by these studies reinforce the need to effectively use new information technologies as a means of providing homeland security information before, during, and after times of crisis.


These studies blended quantitative and qualitative textual analysis of multi-modal media reports about terrorism and homeland security-related risks. Quantitatively, they employed centering resonance analysis (CRA; for more information, see Corman, Kuhn, McPhee & Dooley, 2002), a novel data-mining algorithm that searches for noun centers in large discursive systems. CRA assists in the study of large, complex discourses by systematically analyzing word associations, organizing them into a network of related words, and representing intentional, discursive acts of speakers. Grounded in theories of communication coherence or "centering theory," CRA identifies "conversational centers," which are typically found in the form of noun phrases that provide coherence across sets of utterances. In the case of the Boston Bomb Scare, we analyzed discursive patterns across hundreds of pages of text captured from Boston-area online message boards (BostonMetBlogs, Boing Boing, Slashdot, and Universal Hub), 21 national television news transcripts (ABC, NBC, CBS, MSNBC, Fox, and CNN), and 76 articles from national (n = 21) and local online news sources (n = 55). National sources included The New York Times, Washington Post, LA Times, USA Today, and the Wall Street Journal, while local coverage encompassed the Boston Herald and Boston Globe. We examined reports occurring on or between January 31, 2007, through February 14, 2007, by which time media coverage had waned. In the case of the 7/7 London attacks, we used CRA to analyze more than 370 pages of text captured from popular blogs that follow counterterrorism issues: DailyKos (left-leaning US, n=8), Instapundit (right-leaning US, n=38), CounterTerrorismBlog (US blog devoted specifically to terrorism, n=44), and HarrysPlace (right/center-leaning UK blog, n=48), and 693 articles from leading US (n = 381) and UK online newspapers (n = 312). Online US newspaper sources included The New York Times (n=173), Washington Post (n=114), and Los Angeles Times (n=94), while UK sources included FT.com (n=60) and The Guardian (n=252), from July 7, 2005, through December 31, 2005. By the end of our collection period, news sources were no longer consistently reporting on the 7/7 attacks. We augmented the CRA analysis with in-depth qualitative thematic content analysis to add nuance to our interpretation and analysis.


Project Period: 
September 2006 to May 2009