ABSTRACT
This article examines whether media portrayals of Islam and Muslims are overwhelmingly negative, whether they have evolved over time, and what factors most consistently affect their tone. We analyse every fourth New York Times headline about Islam or Muslims between 1985 and 2013. We find that headlines have not been predominantly negative. In addition, New York Times headlines about Islam and Muslims have become more positive over the long term, even after the 9/11 attacks. Most counter-intuitively, we find that terrorist attacks have had a systematic positive effect on headline tone. During the first four weeks after each Islamist terrorist attack on an American target, the tone of New York Times headlines became significantly more positive compared to the four weeks prior to the attack. However, over the subsequent few months, coverage reverted back toward the tone that prevailed during the weeks before the event.
Acknowledgments
The authors thank Bert Johnson and Maurits van der Veen for helpful feedback on this project. The lead author also thanks Middlebury College for grants supporting data collection and analysis.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. See http://www.usatoday.com/story/money/business/2014/10/28/aam-circulation-data-september/18057983/, accessed 11 March 2015.
2. See http://www.jsonline.com/news/opinion/newspapers-more-relevant-than-ever-before-b99112179z1-226211941.html, accessed 11 March 2015.
3. It is regarded as a left-centre newspaper and has a higher proportion of international stories than most newspapers. Moreover, studies of media coverage of Muslims and minorities in multiple US newspapers have demonstrated distinctions across outlets (Nacos and Torres-Reyna Citation2007; Trevino, Kanso, and Nelson Citation2010).
4. Some researchers view victim frames as ambiguous or even negative, but we follow Benson (Citation2013, 7), who convincingly argues that victim frames convey a positive valence because they are likely to generate sympathy.
5. Psychological research on media effects shows that multiple variables influence individuals’ reactions to media (Valkenburg and Peter Citation2013). Our study is based on the premise that headlines with a tone that is identifiable by multiple researchers are likely to evoke a similar reaction in the majority of readers.
6. The number of headlines in our annual samples containing either a positive or negative tone ranges between ten in 1988 and sixty-one in 2006, with most years (1992–1997; 2001–2013) containing more than twenty-five toned headlines.
7. Two other indirect measures of tolerance of difference confirm that the patterns are more complex than linear. Openness to immigration increased between 1995 and 1999, but then decreased in 2006. Being ‘very proud’ of American nationality has decreased since the mid-1990s, but was similarly strong in the 1981 and 1995 surveys.
8. This strategy focuses on the events likely to have had the clearest media effect, setting aside failed terrorist attempts and terrorist plots.
9. p=.028 in chi-square tests of positive, negative, and no toned headlines comparing the two time periods.
10. p=.025 in chi-square tests of positive, negative, and no toned headlines comparing the two time periods.
11. p=.85 in chi-square tests of positive, negative, and no toned headlines comparing the two time periods, which is not statistically significant at any level.
12. Among headlines with a positive tone, those with a victim frame increased by a factor of 5.4 while those with a beneficial frame went up 3.3 times. Chi-square tests show that the change in positive headlines is statistically significant at the p<.05 level (p=.011) while the change in negative headlines approaches but does not obtain statistical significance at the p<.05 level (p=.0501).
13. This overarching finding is consistent with Ibrahim's (Citation2010) observation that American network news contained more positive coverage of American Muslims than foreign Muslims in the weeks following the 9/11 attacks.