Watching the Whole World: The Media Framing of Foreign Countries in US News and its Antecedents

ABSTRACT

For over five decades, researchers explored the volume of coverage countries receive in other countries’ media, the factors that shape countries’ newsworthy, and sentiment in their coverage. However, a quantitative systematic analysis of how countries are covered, specifically, the framing of coverage, has been lacking. We first utilize machine learning for the descriptive inductive identification of frames in large-scale content (N = 105,991 news articles) to examine how US news outlets covered the 55 countries included in the 6th Wave of the World Value Survey, over a year (Study 1). We then examine the factors that predicted the prominence of frames at the country level (Study 2). Study 1 identified three frames—conflict, economic, and human-interest, which correspond with, but also different from, previous framing frameworks. Study 2 found that factors connected to prominence, relatedness, and conflict, predicted the use of specific frames. Prominence (Specifically GDP) increased the use of the economic frame and decreased the conflict one. Relatedness, with emphasis on trade, cultural proximity and geographic distance increased economic framing and decreased conflict framing. Lastly, Conflict-related variables, mainly military expenditure, increased the salience of the conflict frame.

KEYWORDS:

• Framing
• nation branding
• foreign affairs coverage
• Political Communication
• topic modeling
• network analysis

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Although this method can be used to assess the prevalence and framing of every country in the world, the sample of countries examined in this study was limited by the availability of data for the independent variables, especially the cultural proximity measure data. However, our sample included countries from various continents. It included 11 of European countries, 6 of the countries in South America, 10 countries from Africa, 23 countries from Asia (including the Middle East), 3 countries from North America, and 2 countries from Oceania. Thus, while the sample is not complete, it does include various continents, with different countries from China, to Brazil, or Rwanda.

2 After the topic model was estimated, these documents were returned to the dataset to accurately assess the frame prevalence of all countries mentioned in the article.

3 Full data for this analysis as well as the full script for the procedure has been made available to researchers at the following Github link: https://github.com/DrorWalt/JournalismStudies2022

4 All in USD. West Bank and Gaza Strip are combined for Palestine for lack of comprehensive category in datasets.

5 Trade overall was not available from WITS for 2017 for Haiti, Iraq, Palestine, Taiwan, Trinidad & Tobago and Yemen. These were collected from the Observatory of Economic Complexity.

6 Palestine’s geographic proximity was calculated using Google Maps from Washington DC to Ramallah.

7 SIPRI data is designated “uncertain” for Algeria, Argentina, Egypt and Iraq and as “estimated” for China, Kyrgyzstan, Lebanon and Ukraine. Yemen military expenditures were obtained from World Bank for the year 2013, the most recent data available. For Uzbekistan it was collected from World Bank for 2018 as the next most recently available was from 2003. Because Hong Kong military expenditures are funded by China, expenditure was estimated at NA and was dropped from the relevant regression models. Libya military expenditures were unavailable from SIPRI and were collected from a Reuters report. Palestine military expenditures were obtained from Wikipedia (based on Haaretz report in 2013 by Amira Haas: https://www.haaretz.com/.premium-pa-budget-reflects-donor-dependence-1.5236235). Iraq boots on the group data is an estimate from a December 2017 US Department of Defense report.

8 2017 data was used for some countries when 2018 was not available. These countries include: Kazakhstan, New Zealand, Thailand, Ukraine, and Uzbekistan. Additionally, 2017 data from the World Health Organization was used for some countries for which UNODC data was unavailable. These countries include Egypt, Iraq, Kuwait, Libya, Qatar, Rwanda, Tunisia, Yemen, and Zimbabwe. Peru, Taiwan and Trinidad data from 2018 was pulled from Statista as it was not available through the UNODC or WHO. For the Natural Disasters Index, Hong Kong and Taiwan are tied to China’s score. Palestine is tied to Israel’s score.