A recent Genocide Intervention Network (GI-NET) analysis based on open-source incident reports estimates that roughly 1200 people have been killed in direct, reliably and publicly reported fatalities in Darfur in the first nine months of 2008. As Alex de Waal has noted here and elsewhere, this turns out to closely match data from UNAMID, counting 1500 direct fatalities for the entirety of 2008. Since then, the statement that there have been 1500 deaths in Darfur in 2008 has been put forth on this blog, on Democracy Now, and in Andrew Natsios’ piece in Foreign Affairs.

There are two major problems, however, with how the number of 1500 deaths is likely to be understood publicly. First, critical caveats are being left behind, leading to a perception that this number is a comprehensive summation of total violence in Darfur. It is not. Second, some of the most useful information behind this number is not being correctly or sufficiently discussed, particularly regarding who the perpetrators and victims are. At bottom, the figures obtained by collating existing, reliable reports are useful and should be kept in mind – but they only add accurately to our knowledge if it is made clear what they do and do not account for.

Part 1. We must take seriously the caveats about what this number means and what it doesn’t.

GI-NET sought to aggregate existing open-source data on recent violence in Darfur in order to arrive at the best possible picture of violence. While we realized this would be necessarily incomplete, the fact is that we read and process reports as they come across our desk anyway – better to carefully aggregate them, remove duplicates, and do a careful accounting of what information is available than to sit back and rely on the human mind as a rough aggregator. We also knew that the types of violence we are able to clearly track in an incident database — direct, reported violence and particularly fatalities – would offer only a very narrow window into the total harm being perpetrated on civilians in Darfur.

In our report, we took great care to stress the caveat that the approximately 1200 hundred deaths between January and September of 2008 accounts only for direct, reported fatalities. This represents the number killed by bullets, bombs, rocket-propelled grenades, or knives, and was only included if it was clearly reported by reliable sources. Our statement of direct mortality in Darfur does not include what may easily be the most significant threats to civilians in Darfur today, including the following:

“¢ Indirect mortality due to displacement and other intentional acts by those who continue to commit crimes with impunity. Across conflicts, the death toll due to indirect effects is many times higher than the number killed directly by bombs and bullets. In his Foreign Affairs article, Andrew Natsios suggested this component of fatality is no longer significant because aid groups are now in place. While the impact of displacement and injury on fatality is no doubt mitigated by the presence of aid groups, particularly relative to early 2003 and 2004, some degree of indirect fatality surely continues to occur. When people are displaced in large numbers, not everybody makes it to succor, or survives as long as they otherwise would once they do reach it, especially the young, sick, or elderly. We do not include this component of violence in our analysis, only because we do not have the data needed to come up with even the beginnings of an estimate for indirect mortality during 2008. Effective research on indirect mortality requires carefully sampled mortality surveys which do not yet exist for this period.

“¢ Second, as our analysis relies entirely on information available in the public sphere, these numbers cannot be interpreted as an exhaustive list of all deaths in Darfur in 2008. There are surely unreported direct fatalities. In addition, there are reports of fatalities which are simply too unreliable to use, such as statements issued by the warring parties directly.

“¢ Finally, deaths in Darfur, whether direct or indirect, do not represent the entire scope of ongoing violence. In fact, one problem with our report, and with providing a number for direct mortality, is that it serves to emphasize direct mortality as a key metric, which may answer the wrong question altogether. One must also consider the massive scale of non-fatal harms inflicted upon the population of the region. These include the widespread destruction of livelihoods, pervasive rape, assault and other crimes. These appear, at least in part, deliberately intended to coerce civilian populations to leave their homes or deny support to the movements. Over 2.7 million remain displaced within Darfur – suggesting that threats of violence remain too high for many to return home permanently. Whether one believes the term “genocide” applies to Darfur or not, there has been a growing recognition that some modern atrocities, particularly those that occur in the face of international condemnation, take a form resembling Article 2(c) of the Genocide Convention: “Deliberately inflicting on the group conditions of life calculated to bring about its physical destruction in whole or in part.” Physicians for Human Rights and the Feinstein International Famine Center have done excellent work on the applicability of this strategy in Darfur’s conflict. Likewise, GI-NET’s own protection projects partly involve efforts to help improve the UN’s ability to protect civilians from mostly non-fatal harms, such as rape and assault that women face when they leave the camp to collect firewood. These types of harms are difficult to quantify to begin with, but they are made much more so in Darfur where reporting is staunched by a government that expels organizations for releasing such information. We should not fall into the trap of believing that these harms do not occur just because they are not as neatly quantified, especially when this is largely the result of an intentional effort to restrict such information.

Given that direct, reported fatality is not a complete picture of these types of harms – especially when displacement, rape, and destruction of livelihoods are used as weapons of war, we realized there is some risk in publicizing a number for direct reported mortality at all. We hope, however, that at least the majority of analysts will understand it as a partial picture.

Part 2: Going beyond the number – what we know about these direct fatalities

The second major shortcoming with the use of this number so far has been that important information about how this number can be broken down has either been misrepresented, or has received little attention. In one case, this report was mis-cited as the source of a claim that half of these 1500 deaths in 2008 involved “Arabs killing Arabs” rather “Arabs killing Africans”. It is unclear, though, how such a claim can be attached to GI-NET’s analysis at all. It is true that the GI-NET analysis coded some of the violence in Darfur only as “tribal fighting,” and that some of the tribes involved in these attacks could readily be identified as “Arab”. Only 10% of civilian fatalities in our data were related to “tribal fighting”, though, so this could not be the source of this claim. Moreover, we did not generally identify the ethnicity of either the perpetrator or victim of attacks — in fact, the only time the word “˜African’ even appears in our database of over 140 events is in the description of a slain peacekeeper as “South African”. For us, and perhaps for many Darfur-watchers and activists, the relevant moral distinction is whether or not civilians are being killed or severely harmed on a large scale, and how deliberately and systematically.

This brings us to some of the useful information revealed by our analysis that is easily overlooked by use of this single number. Consider what the data say about who the victims are: the overwhelming majority of those killed in direct, reported violence appear to be civilians. By our estimate, about 76% of the direct fatalities tracked were among civilians. Next, consider who the perpetrators are. A shocking proportion of these direct civilian deaths — nearly 80% — died during attacks which clearly involved Janjaweed and/or Government forces attacking alone without evidence of concurrent fire by rebel groups. That is to say, these deaths are not simply the result of civilians being trapped in “cross-fire”, at least not in the literal sense. [ A more subtle point: we were initially concerned that many “Janjaweed” attacks may involve misidentified armed groups that are not, in fact, under the employ of the government. However, of all the civilians killed during attacks by the government and/or Janjaweed, only 10% were killed in attacks which identified the Janjaweed as the sole perpetrator. That is, groups identified as “Janajweed” during attacks that kill civilians are almost always working alongside government forces.]

Among civilians killed by rebel groups, about 49% — by far the largest share – were killed in attacks by JEM (many of these were killed during JEM’s attack on Omdurman).

All of this represents the best estimates we were able to determine with available, open-source data, only in January through September of 2008. As more data become available, these findings may be reinforced or challenged. More to come on this in the next few months.

When we released these data, we realized of course that certain findings like the direct, reported mortality count would garner more attention than others. We certainly realized that the attendant caveats must be taken seriously — an issue we spent almost a full page on in our report. Nevertheless, on balance we felt it best to provide information collated from available sources in order to increase our understanding of the present nature of the conflict in Darfur. We ask only that these data are used appropriately, with care taken to remember what they do and do not account for. Only then are they likely to be more helpful than harmful in designing and promoting effective means of achieving what we are all interested in – bringing about an end to atrocities against civilians in Darfur.

Chad Hazlett is Director of Protection at GI-NET; Joshua Kennedy is Education Associate. This statement is the work of the authors and does not reflect the positions or policies of the Genocide Intervention Network.