07/27/2010 11:31 am ET | Updated May 25, 2011

Analysis of Civilian Casualties in WikiLeaks Afghan File Reveals Media Bias

The release of 91,000 classified military documents relating to Afghanistan by the organization known as WikiLeaks offers the opportunity for a controlled experiment in an analysis of media bias. This was a suggestion by the Nieman Journalism Lab immediately following the documents release. Three mainstream media organizations (The New York Times, The Guardian, and Der Spiegel) were given the same amount of time to analyze these documents prior to their public release on July 25th and all three published their accounts on the same day. Therefore, any emphasis or de-emphasis in how the material was presented can be used to test hypotheses about the mainstream media through a process known as content analysis. This involves both assessing the meaning of a given text as well as measuring how frequent a word or phrase shows up in a specific context.

The hypothesis I seek to test is that different levels of access to American officials influenced how media outlets framed their respective analyses. A first glance at the material presented in the two English-language sources, The New York Times and The Guardian newspapers, reveals dramatically different approaches that each took in reporting on these leaked documents. In The Times, for example, the first headline on their Afghanistan War Logs page reads, "Pakistan Spy Service Aids Insurgents" and three of their four featured reports on July 25th either emphasize the security and military implications of Pakistan's involvement or focus on US military strategy in executing the war. The New York Times provided no article focusing on civilian casualties in the war and mention them only as small points in their summary of individual documents. In contrast, The Guardian offered two prominent articles detailing the thousands of civilians whose deaths were documented in these files--not including those who died at the hands of Task Force 373, the shadowy special forces unit engaged in assassination raids.

However, while a qualitative assessment is extremely useful, finding a way to measure media bias can be even more revealing. The word "civilian" in an active war zone is most often associated with casualties and would be expected to have the vast majority of uses in that context. I therefore counted every reference to civilian casualties that were mentioned in the featured articles in both newspapers--four articles in The Times and five in The Guardian. Of the twenty times the word "civilian" is used in The Times only nine uses are in reference to casualties resulting from combat operations (four of these are clustered in a single section midway down the page and two were at the hands of Afghan soldiers or police). The Guardian's coverage used the word "civilian" 41 times in their primary coverage and 37 of these uses referred specifically to civilian casualties (two cases occurred in each newspaper concerning hypothetical casualties and these have not been included). The difference between The Times and The Guardian is dramatic and represents a ratio of 2:1.

This means that either The Guardian overemphasized civilian casualties because they are biased against coalition troops or that The New York Times sought to underemphasize these same casualties because of their political leanings. Both newspapers have taken editorial stances critical of the war and, in the case of The Guardian, eleven of their thirty-seven references to civilian casualties were found in their article analyzing the deaths that resulted from Taliban IEDs. This wouldn't suggest that The Guardian has a clear bias against coalition troops. Furthermore, the fact that bothWikileaks founder Julian Assange and the office of Afghanistan President Hamid Karzai have stated that the most significant revelations in these documents concern the scale of civilian casualties, the lack of reporting on this issue from The Times would appear to be a glaring omission in the paper of record.

The clear implication is that The New York Times, whether intentionally or unintentionally, avoided what these leaked documents reveal about the impact on Afghan civilians in the conduct of this war. The disparity in reporting is even more dramatic when shown as a percentage of the overall word count. The Times published just over 11,100 words in their four featured articles, while The Guardian had 6,600 words in five articles. When the reporting on civilian casualties is compared after taking into account the overall word count the results couldn't be more clear. The New York Time's analysis focused seven times less attention on the civilian cost of the war than did The Guardian's.

However, the scale of civilian deaths is much higher than either The Times or The Guardian newspapers indicated. In The Guardian's article entitled "Logs Reveal Grim Toll on Civilians" they state that the documents show "144 entries in the logs recording...hundreds of casualties." However, this was only in the so-called "blue on white" events (those cases where US and NATO forces acknowledged firing on civilians). Further analysis of the data show that these are only a small percentage of the overall impact on the Afghan population. In the category labeled High Severity there are 1,539 pages including 50 military reports on each page. A search for CIV KIA (military code for civilians killed in action) among the first 5,000 reports brings a total of 796 hits. In other words, an average of one in six reports contains evidence of a civilian death, and most involve more than one.

While many of the civilian deaths analyzed in the High Severity section involve the explosion of IEDs set by the Taliban and other rebel groups, others are the direct result of American combat operations. In one example from 2007, eight civilians were killed and four were injured when American forces returned fire after reportedly being attacked with an RPG in Kandahar province. In another report from 2009, seven civilians were killed and one was wounded in a rural village near Nad-E'ali when American troops engaged and killed four alleged militants (though the report says that these enemy deaths were unconfirmed). Both of these reports are listed under the classification Enemy Action, but from the details provided it's impossible to know who was ultimately responsible for the civilian deaths. By emphasizing only the known cases of coalition troops killing civilians The Guardian was actually underreporting the real impact.

What the WikiLeaks material reveals most clearly is the devastating toll this war has had on Afghan civilians. That The New York Times chose not to emphasize this fact suggests a political motive to avoid discussing the human impact of the war. This is consistent with the hypothesis that a close association between journalists and American political, economic, and military officials would influence reporters in the direction of those same officials. We would expect the same if The Guardian were reporting on an issue central to English politics. Relationships form between journalists and political officials over time and the desire for access can result in a breakdown between an otherwise antagonistic press and representatives of political power. When this occurs compromises are made in order to maintain the same level of access. This hypothesis is supported by a second comparison of the word "official." The Times contains 44 mentions of the word, with 27 referring to American or Afghan government officers (and a further 8 referencing Pakistani officials). The Guardian uses the word 8 times and only 4 in reference to American or Afghan representatives, as would be expected for England's more limited role in the conflict.

The general conclusion is that The New York Times had both the official access and the motivation for a conflict of interest to have occurred whereas The Guardian did not, a factor that influenced how news of this military leak was framed. This conclusion is consistent with previous studies of how The Times has covered issues of American foreign policy (pdf). In their "Note To Readers" The Times' editors explained how they were involved in talks with the White House prior to making a decision over what they would publish. How much self-censorship The New York Times ultimately volunteered will be revealed in the years to come as journalists and scholars analyze this data in our effort to better understand the conflict.