Spotlight on: South Sudan

As we mark 10 years of conflict in South Sudan, Every Casualty Counts’ Prof Michael Spagat interviews Dr Sophia Dawkins. Here they explore body counts, the importance of context and the need to honour lives as more than data points.


When I graduated from my masters at the Fletcher School, Tufts University, I was recruited by a peacebuilding and humanitarian negotiations organisation on the basis of my Sudan experience. It was 2011, and South Sudan was about to become an independent country. The new state faced a host of unresolved conflict issues. My job for the next four years was to support consensus-building among elites around those issues. I met state governors, spent many days in the South Sudanese National Legislature, and worked with women parliamentarians. Then, in December 2013, the army fragmented and civil war erupted across South Sudan. One of my most emotive memories is of traveling to the rebel military headquarters behind the front lines to prepare officials for humanitarian negotiations. 

I became deeply troubled by upticks in civilian killings and human rights abuses during moments of apparent elite agreement.

Throughout this work, I moved fluidly between conflict conditions on the ground and elite discussions in Juba, Nairobi and Addis Ababa. I found that what I assessed as a mediation success in the heat of a political moment didn’t necessarily translate to improvements for ordinary people on the ground. I became deeply troubled by upticks in civilian killings and human rights abuses during moments of apparent elite agreement. This pattern repeated itself again and again, and I couldn’t explain it. This is what drove me to apply to PhD programs.  

Eventually, I arrived at the Political Science department at Yale University bent on studying patterns of civilian insecurity during peace negotiations.  But when I opened statistical software on my computer to view large datasets on civilian killings, I felt completely frozen with horror by the picture they drew. I saw thousands of points on my screen in black and white, which I knew represented individual people.  I became transfixed by these individuals’ stories. I wanted to know how these people had lived, died and then arrived as data points on my computer screen. 

This obsession, to this day, shapes how I do research.


South Sudan covers a territory that lies below the tenth degree parallel in East Africa. This is where desert gives way to a verdant landscape and the Sudd – the largest marsh on Earth. Throughout history, colonisers frequently stopped at the tenth degree parallel, finding the terrain too difficult to navigate. For example, the British and Egyptians who ruled Sudan by condominium from 1899-1955 built road networks that ended where South Sudan now begins. To simplify a vastly more complicated reality: These inequalities underpinned periods of civil war that swept Sudan following its independence from colonial rule in 1956. 

At the turn of the millennium, the civil war became particularly violent and involved dozens of armed groups across what is now South Sudan and Darfur. This posed huge challenges to Western powers in managing the War on Terror. Osama bin Laden kept a mansion in the Sudanese capital, Khartoum. The US pushed for a peace agreement, fast. To prevent the peace process from falling apart, the mediators brought the Government of Sudan and only one of the southern rebel groups, the Sudan People’s Liberation Movement / Army (SPLM), to the table. While the SPLM was the largest rebel group, it by no means harboured a majority of the armed opposition. 

The eventual peace agreement provided for a secession referendum. In 2011, southern Sudanese voted and South Sudan become an independent country under the rule of the SPLM. However, this political process did not accommodate the interests of the many factions that still existed in South Sudan. 

Politics became a tinder box when the SPLM government precipitated an economy crisis. At the time, 98% of South Sudanese government revenue came from oil, pumped from southern wells through a pipeline that crossed Sudan. The two governments couldn’t agree a transit fee. When Sudan started holding back barrels of oil at port Sudan, South Sudan played a game of chicken by shutting off its oil wells. South Sudanese government revenue vanished and army pay became unsteady. 

These problems came to a disastrous head in December 2013. A shootout between soldiers at the army headquarters swept through the capital, Juba, and then through the rest of the country. Over a decade, the political disputes and patterns of violence that define this conflict have shape shifted, swinging between escalation and de-escalation.  

The government and rebel groups have signed several peace agreements premised on a power sharing structure that never seems quite congruent with the configuration of political interests on the ground in South Sudan. At present, South Sudanese are preparing for their first post-independence election in December 2024, meant to bring an end to a transition period out of civil war.   

“…body counts reveal little if divorced from circumstance, it is impossible to construct theories about violence with numbers decoupled from the human experiences that lie beneath.”  


As Carolyn Nordstrom has put it: ‘In war, there is always something wrong with the facts that one is given.’ This is especially true where bureaucracies are extremely weak. In South Sudan, there is no reliable birth or death registration system. It falls to peacekeeping officials, human rights investigators, public health NGOs and journalists to record information on violent events. All these actors have varying access to locations where violence happens. Journalists often gather information by phone from the capital, Juba, and human rights investigators necessarily arrive on the scene some time after an event has happened. All of these actors counting the dead are entangled, because they often go to the same limited sources to verify information. Journalists will call human rights officers and vice versa, with everyone calling the same person on the ground, such as a local government official.

It is important to note that South Sudanese have their own painstaking, albeit informal, processes for recording the dead. It is common for chiefs to list the dead in their communities after a violent event. These lists may be in biro on a scrap of paper. Sometimes, chiefs will journey to a UN peacekeeping base to present the lists to officials. Like any source, the lists are open to inaccuracies and biases, and they are inconsistently kept. However, I suspect they may be amongst the most accurate records of the civilian dead that exist on the ground. 


Let me be specific about levels of information.  Policy makers often draw on large datasets about the dead and events where people have died in conflicts. The researchers that compile these large datasets draw on media and human rights reports, while also consulting regional experts. 

This data-generation process begins on the ground in the aftermath of a violent event. A journalist or a human rights investigator will try to make sense of what has happened. A typical statement they might write after investigating a site would be something like: ‘We spoke to X who said that they observed that 10 people died in this village.’  That report may get published, or it may not. It may get stuck in a bureaucracy, left in a drawer, or published and launched through social media. 

If the report becomes public, then report-based conflict data projects do a very important job of deploying highly trained researchers to interpret the report and code its contents into a dataset. These researchers often adhere to rigid methodological rules in this process. There is no doubt in my mind that these data are amongst the best the public can access about violent events and the civilian dead.  However, biases remain.  

Violence that is too violent to report goes unreported. Soldiers and their commanders implicated in an atrocity can restrict physical access to a site so that investigators never get there. They can ensure people don’t answer phones if journalists try to call them.  Insecurity can restrict UN investigative teams to their bases for long periods after a conflict event. Belligerents can change the observable facts on the ground before external monitors see anything. I have heard of cases where soldiers have strategically concealed some bodies, while putting others on display. 

For example, in April 2014, opposition soldiers massacred hundreds of civilians taking shelter inside a mosque in Bentiu, South Sudan. A UN human rights report records accounts of the rebel group using the few survivors to load bodies onto a pickup truck. Accounts allege that rebel soldiers stripped these survivors of their identity documents, drove them with the bodies deep onto the bush, and forced them to hide the dead in a drainage ditch.

An exceptionally resourceful UN investigator pieced together this information over months through repeatedly visiting the town, interviewing former guards at the mosque gate, and building trust to get past witnesses’ fear of speaking. This is an example of a large-scale atrocity that was deeply researched, with an investigation headed by one of the most skilled human rights officers in the world.  It makes me consider the many smaller events with less-skilled reporting that go unrecorded.

Casualty recording is one of many domains where the personal is political and the political is personal.

Another issue is self-fulfilling narrative bias. The philosopher Philip Kitcher has argued that humans will never be able to research everything and know all truth. We make normative choices when we decide what to examine. This is especially true for casualty recording. South Sudan is a resource-intensive context for conducting investigations. Human rights and media organisations must be selective about where they send researchers and how long personnel stay on the ground. Agency mandates, advocacy agendas and public interests will shape those decisions. I have spoken to newswire journalists who described their hand-to-mouth existence in South Sudan’s capital, Juba, and how their ability to afford food depended on selling stories. Reporters told me how it was impossible to get fees for some types of report. Any old story of killings or famine wouldn’t do. One journalist described how they struggled for many years to sell stories about cattle conflicts because they were such a complicated component of the broader civil war.  Therefore, journalists had to seek out stories they could weave into a narrative an international media house felt would resonate with a global public.  

In sum, many conflict datasets draw on published reports. Those published reports are the products of a process shaped by what reporting agencies think a conflict is about, where organisations invest resources, and where belligerents will allow people to go.  


I wish it were otherwise – but my paper’s analysis still applies. Not much has changed in the infrastructure for casualty recording in South Sudan. However, many agencies recognise the problems I have laid out and are developing initiatives to record the dead more systematically. The UN has been developing internal data management systems that aim to record ground-level information more efficiently. The UN has also commissioned deep human rights investigations and been bold in publishing findings that identify perpetrators. Human rights organisations are also making more use of satellite data in combination with AI and machine learning applications. This has limits in identifying individuals, but promise in tracking demographic changes in inaccessible locations.


Every society makes decisions about who is remembered and who is forgotten.  Every person who dies in an untimely way holds normative weight, regardless of whether a society includes them in a body count.  Furthermore, perpetrators kill for many different purposes and in many different ways – sometimes to project a message, sometimes to propagate a practice. I interviewed an aid worker who told me about discovering a dead man in a displaced people’s camp. The killers had castrated the man and ritually hung up his corpse to transmit fear through his community. Taken in isolation, the appearance of this man as a single data point on my computer screen reveals little about the consequences of his death – of the terror his community felt observing his body, and how that terror shaped people’s daily decisions. 

Casualty recording is one of many domains where the personal is political and the political is personal. Both dignified memorialization and effective humanitarian policy require deep investigation of how distinct experiences of killing and dying ripple across society. 

Dr. Sophia Dawkins is a scholar and humanitarian practitioner who has followed South Sudan for 12 years. 

Share this article