When journalists report on human rights abuses, which region do they report on most? Africa, due to the Rwandan genocide, Darfur, or the al Qaeda-linked militants in Mali? The Middle East, as a result of Egypt, Syria, and Gaza?
The correct answer — at least for Newsweek, the Economist, and The New York Times, from 1981-2000 — is Latin America.
Controlling statistically for government repression, income, population, political regime, and other factors, violations in Latin America received a whopping 42 to 82 percent more media attention than similar abuses elsewhere in the world. This reporting pattern, moreover, contradicts other studies demonstrating that when it comes to general foreign affairs, the international media is far less interested in Latin America than in other world regions.
We discovered this “Latin Bias” while analyzing our data on all Economist and Newsweek stories with the keywords “human rights.” We loaded a battery of additional factors into our computers, ran dozens of statistical tests to verify our results, and cross-checked our findings on New York Times data gathered by others.
The results of our recent labors are online in International Studies Quarterly, a leading scholarly journal; the print version is due out in March 2013.
To be sure, there is no Latin Bias until we use statistical controls. Based on raw counts alone, the number of articles devoted to abuses in Asia outstripped those devoted to Africa or the Middle East and were often much higher than Latin America.
But when we control for the influence of other factors, including government repression, population size, per capita income, and more, Latin American abuses gained more attention than those occurring elsewhere.
To investigate further, we questioned more than a dozen veteran foreign correspondents, editors, and bureau chiefs at the Economist, Newsweek, Washington Post, Financial Times, Foreign Affairs, Die Zeit, Ottawa Citizen, and others.
Collectively, the journalists offered a range of possible explanations. We located new data to represent these arguments, and ran new statistical tests to see if their explanations were correct.
Many of the journalists we spoke to, for example, mentioned the fact that US strategic and economic interests have long loomed large in Latin America, so we first looked for the effects of Washington’s influence on news coverage.
“Human rights abuses are more frequently covered when their continuation appears to depend at least in part on US foreign policy,” one journalist explained, and in the 1980s, “the wars in Central America created a direct link between human rights in the region and US policy.” This prompted journalists to investigate whether ”US foreign policy [was] aiding and abetting human rights abuse.” As another put it, the ”extent to which Washington opposed abuses—or became complicit in them—developed into a major part of the story.”
To test this possibility, we introduced a battery of variables into our statistical models, the goal of which was to measure the intensity of US policy interest. These included geographic distance between a nation’s capital and Washington, DC, international trade flows, voting practices at the United Nations, and US aid.
When we controlled for the statistical effects of geographic distance and a country’s propensity to vote with the US at the UN, the Latin Bias still endured. Despite our best statistical efforts, the Economist and Newsweek still reported 58 to 78 percent more on abuses in Latin America than anywhere else in the world.
Other journalists noted that Latin America experienced democratization earlier than other world regions, providing “a channel of protest that found a wider audience, one that was perhaps unavailable to Asians and Africans [at the time], where human rights abuses were surely just as serious, if not worse.” Democratization, moreover, is often a compelling media drama, as the Arab Spring recently demonstrated. The struggle for political freedom makes for great copy, accompanied as it is by tales of heroism, villainy, and intrigue.
Although we had already controlled for political regime type in our initial statistical models, we went a step further, identifying periods of particularly dramatic political change. We found, however, no statistically significant “democratization effect.” No matter which statistical method we tried, there was no evidence linking more intense political shifts with greater levels of human rights media reporting.