Using the Internet to Examine Patterns of Foreign Coverage

The first week of April 2003, several hundred people were killed in ethnic violence in the Ituri region of the Democratic Republic of the Congo (DRC). Given the magnitude of the event — up to a thousand civilians killed in a single incident — and the history of violence in the region, it made sense to expect media coverage. Shortly before the killings, the International Rescue Committee published a study suggesting that 3.3 million people had died as a result of conflicts in the DRC, making the ongoing violence in the region the deadliest war in the world since World War II.

But the events in Ituri went almost unreported. On April 7th, the first day American newspapers reported the killings, The New York Times ran a brief Associated Press story on the conflict, buried on page A6. Google News, a Web site that monitors 4,500 news sources, listed only 1,200 stories in the preceding month that mentioned Congo. By contrast, on the same day Google News showed 550,000 stories for Iraq, and The New York Times ran five Iraq stories on the front page, as well as a separate section, “A Nation at War.”

While it’s predictable that the U.S. invasion of Iraq would squeeze most other news off the front page of American newspapers, it’s only one of several reasons the conflict in Ituri received so little attention. In their seminal 1965 paper, “The Structure of Foreign News,” Johan Galtung and Mari Holmboe Ruge proposed 12 factors that influence the publication of international news. While Galtung and Ruge’s statistical analysis has been questioned, their propose factors are still widely used by media theorists to explain the inclusion and exclusion of international news stories.

Galtung and Ruge, writing almost 40 years before the Congo event, could have predicted the events in Ituri would have been ignored in the United States:

• The Democratic Republic of the Congo is a “non-elite” nation.
• No “elite people” were killed in Ituri.
• There’s little cultural proximity between the United States and the DRC.
• The conflict had little meaning for American readers.
• And the decade-long war in the region meant that further killings weren’t unexpected.

Their analysis doesn’t consider newsgathering factors — the difficulty of deploying reporters to northeastern Congo, language barriers, and the lack of communications infrastructure — all of which make it more difficult for reporters to cover the conflict in DRC, especially in contrast to the war in Iraq, which featured opportunities for reporters to be “embedded” within U.S. military units.

Global Attention Profiles

While Galtung and Ruge used 1,250 Norwegian newspaper clippings gathered over four years to propose their rules, the advent of Internet publishing gives us the opportunity to test some of their conclusions with hundreds of thousands of data points. Shortly after the incident in Ituri, I started collecting data from the Web sites of U.S. and British newspapers, news services, and television networks for a project I called Global Attention Profiles. My intention was to create daily maps of news stories to demonstrate graphically where Western media attention was focused. As the project progressed, I began to look for correlations to economic and political factors to explain the distribution of news.

My main conclusion: Andy Warhol was wrong — we won’t all get 15 minutes of fame.

Media attention, measured by the number of stories that mention a country by name, is correlated only loosely to a nation’s population. It’s correlated much more strongly to economic factors, especially to a nation’s wealth, as measured by gross domestic product.If this were true, populous nations like China, Indonesia and Brazil would be better represented in the Western media. Media attention, measured by the number of stories that mention a country by name, is correlated only loosely to a nation’s population. It’s correlated much more strongly to economic factors, especially to a nation’s wealth, as measured by gross domestic product. For example, while Nigeria and Japan have roughly equal populations, Japan’s economy is about 100 times the size of Nigeria’s — and there are roughly seven times as many mentions of “Japan” as there are of “Nigeria” in the average American newspaper on any given day. All the American news sources I tracked showed this pattern; the lone source to show a different pattern was the BBC, which showed a strong bias towards news in former British colonies, including populous and poor nations like Nigeria, India and Pakistan.

Correlation is not causation, and it’s unlikely that news directors check a nation’s current account balance before sending a TV crew to cover a story. But, consciously or not, the people who decide what becomes news are far more likely to cover a story if it involves people from wealthy nations. (Indeed, the less developed nations best covered during the year of my study — Iraq and Afghanistan — are nations that Americans invaded and occupied.)

While it’s tempting to accuse news organizations of dereliction in failing to cover events in the developing world, blame might fall equally on market forces and the preferences of media consumers. Confronted with the inequity of media attention, many editors and news directors will readily own up to the disparity and go on to explain that they’re the good guys, encouraging coverage of developing nations: If their customers had their way, there would be even less international news and almost no news from poorer nations. Given the need for publications to maintain an audience to sell ads to, perhaps we’re lucky that there’s any coverage of the developing world.

It’s difficult to test this theory without extremely detailed data about what news stories readers and viewers view or skip. But Weblogs give us a way to guess at reader interest: If a Weblogger mentions a country in her post, she’s likely expressing an interest in that nation. If we found a pattern of Weblogger interest in developing nations — proportionally more mentions of Africa than in the mainstream media, for instance — we might conclude that editors are underestimating their readers.

Alas, we don’t see this pattern. Looking at data from Weblog search engine BlogPulse, we see roughly the same correlation between wealth and mentions as we do in media aggregator sites like Google News or Altavista News and a slightly tighter correlation to national wealth than in single media sources like The New York Times or The Washington Post. Comparing on a country by country basis, Weblogs are more likely to name travel destinations (Caribbean Islands, some Central American and Southeast Asian nations) and far less likely to mention African, Eastern European, and Central Asian nations than mainstream media sources. Disparities aside, the statistics suggest that mentions of nations in blogposts are strongly correlated to their appearance in the mainstream media.

Consequences of News Decision-Making

If readers aren’t interested in international news and it’s expensive for news networks to generate, does it matter that the media doesn’t cover violence in Ituri?

It matters a great deal to Iturians. Governments are less likely to send peacekeepers to work to stop the conflict from spreading if they don’t read about it in the news. And citizens can’t pressure their governments to intervene without awareness of the situation. The huge aid packages coming to Iraq and Afghanistan suggest a relationship between media attention and foreign aid. In the wake of these conflicts, international aid workers have expressed concern that aid to neglected, “unpopular” conflicts will suffer as a result. In more peaceful times, attention makes it more likely that a country will become a trading partner or receive foreign investment.

A recent report by the Center for Global Development, ‘On the Brink: Weak States and U.S. Security,’ suggests that roughly 50 failed and failing states need to be both closely watched and aided so that they don’t find themselves participants in terrorism and global crime. All but three of the states mentioned in the report are systematically undercovered by mainstream media.Wealthy nations have a good reason to care about news in undercovered nations — their security may depend on it. The events of September 11th were carried out by a network that bases itself in weak and failed states. For a brief interval after the attacks, Americans were deeply interested in the Central Asian states that hosted al-Qaeda operatives — this interest waned as global attention shifted from Afghanistan to Iraq. A recent report by the Center for Global Development, “On the Brink: Weak States and U.S. National Security,” suggests that roughly 50 failed and failing states need to be both closely watched and aided so that they don’t find themselves participants in terrorism and global crime. All but three of the states mentioned in the report are systematically undercovered by mainstream media. Like the U.S. intelligence community, the U.S. news media are better configured for a world where threats come from superpowers than from failed states.

It seems unlikely that commercial news organizations will refocus on the developing world without some form of external pressure. In 1980, Sean MacBride led a UNESCO committee that published a report, “Many Voices, One World,” which proposed legal and structural changes to news organizations to improve media coverage of the developing world. The report was opposed so vehemently by media organizations in the United States, United Kingdom, and Singapore that the three nations withdrew from UNESCO to protest implementation of the committee’s proposals. One could be forgiven for skepticism that CNN or Fox News will react any better to suggestions to globalize their coverage than newspapers did two decades ago.

The recent crisis in Darfur, Sudan points to one way concerned individuals and organizations can influence global news coverage. A network of NGO’s — most notably Human Rights Watch — which had monitored human rights situations in Sudan for years, provided extensive information on the Janjaweed militias to major newspapers, making it possible for them to write their first stories on the situation. In effect, they did the first round of investigative journalism that news organizations failed to do. After a major report by Human Rights Watch and strong statements from the United States and the United Nations, media attention to Sudan increased dramatically — it is now receiving the third-most media attention in sub-Saharan Africa (behind South Africa and Nigeria).

The attention paid to Darfur also points to the importance of caring. A global community of evangelical Christians has closely monitored the Khartoum government for years, accusing it of systematic persecution of a Christian minority. This community was deeply interested in seeing that stories came out of Sudan and was able to provide feedback to editors letting them know that they cared about the situation. To encourage news organizations to report on forgotten stories, readers and viewers will have to demonstrate that they care about these issues. But for viewers to care, they will likely need to know a great deal more about these nations. Is this a Catch-22? Or could it present an opportunity for new, participatory media like Weblogs to draw attention to situations and stories that a small group of individuals care about?

I’ll be counting news stories and let you know.

Ethan Zuckerman is a research fellow at the Berkman Center for Internet and Society at Harvard Law School. His Weblog can be found at blogs.law.harvard.edu/ethan.