The distance bias in natural disaster reporting – empirical evidence for the United States

ABSTRACT

Whenever governments or international organizations provide aid in the aftermath of natural disasters, they typically justify this support by humanitarian motives. Previous empirical research found that media reports on natural disasters have a systematic impact on the amount of provided disaster aid. While this is unproblematic as long as media reports are unbiased and thus deliver an undistorted picture of the occurrence and severity of worldwide occurring disasters, systematic reporting biases would lead to distorted aid flows and perhaps other distortions like an insufficient perception of a region in international organizations. Based on data on three US news shows we show that disaster reporting is subject to a distance bias, e.g., the likelihood that a disaster is covered by the media depends on the distance between the country where the media are located and the country where the disasters occur. We also find evidence that besides the distance bias the state of economic development of a country and importance as export markets have a positive effect on the probability that US news shows are reporting on a natural disaster. As a result, international aid flows might be systematically biased and not distributed in line with the needs of the victims.

KEYWORDS:

• Media bias
• distance bias
• natural disasters
• disaster aid

JEL CLASSIFICATION:

• Q54
• Q56
• F35
• H84

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Such a distance bias in media reporting could contribute to explaining why governments of countries more likely offer post-disaster aid when they are geographically close to disaster-affected countries, a finding reported in a study of aid flows in the aftermath of hurricane Katrina in 2005 and the Wenchuan earthquake in 2008 by Wei, Zhao, and Marinova (2013).

² In communication science this sentiment or tone of coverage is called tonality (Haselmayer and Jenny (2017)).

³ As a consequence, Grimmer and Stewart (2013) conclude that in political text analysis there is no substitute for human reading.

⁴ http://www.emdat.be.

⁵ https://www.nato.int/.

⁶ https://www.andiamo.co.uk/resources/languages-spoken-in-each-country-of-the-world.

⁷ We also have to refrain from using the number of affected individuals rather than the number of events as for many events the number of affected individuals is not reported, leading to a critical shrinkage of the sample.

⁸ Our results do in parts confirm the research by Jones, Van Aelst, and Vliegenthart (2013). In their longitudinal analysis (1950–2006) the authors investigate the visibility of foreign nations in the two US media NYT and NBC and find evidence that distance, US troop deployment, GDP per capita, and population were good predictors of the visibility. However, with regards to trade, the results differ: Jones, Van Aelst, and Vliegenthart (2013) find that it is a good predictor of country coverage in the NYT but not on NBC, whereas we find evidence that it is a good predictor country coverage an TV news shows as well.