At times, when we had spent some time in a particular location and word had spread about our presence, people approached us for interviews. Surprisingly, these were often very poor or vulnerable people, typically women and frequently widows, who wanted someone to listen to their story, just for once. These were some of the most amazing conversations we had.
We felt strongly that the value of our research did not justify the risk of retraumatizing people, especially as we had nothing to offer in terms of services or support. As a result, we did not collect much information on painful private matters. For example, I am sure that household conflicts are one of the most serious problems faced by many Burundians, especially women and children. The stunning outside pressures Burundians have been subjected to – the ravages of war, the insecurities of banditry and theft, the grinding pressure of misery, the land scarcity and the fear for tomorrow – often get mediated into deep and long-lasting intra-household conflicts, pitting husbands against wives, brothers against sisters, children of one marriage against children of another. These conflicts are very painful and debilitating: they take up enormous energy, and create fear and pain in the lives of people who are already under considerable stress. Yet our conversations rarely addressed these issues. There were many allusions, but few direct discussions. Similarly, some ex-combatants and women hinted at personal war trauma, but few elaborated. So, clearly, our research did not get at some of the most painful and private things in people’s lives. This was the price I was willing to pay in order not to push people into sensitive and painful territory, but it is a price that must be acknowledged.
Finally, for research like this, the quality of the translators is crucial. Kirundi is a language of allusion and proverbs: information is conveyed between the lines, hinted at, but rarely expressed directly. The challenge is also social: the translator is the front-line person who interacts with the interviewees, making the connection, maintaining the social aspects of the relation, putting people at ease.
I was fortunate to be able to hire excellent translators, with considerable experience. In rural areas, I worked with two young women, Etionette and Alice, who had lengthy experience in promoting rural dialogues. Their capacity to interact with people, put them at ease, listen with empathy and respect was crucial to the success of this research.
Midway through the research project, Alice left and I added Innocent, who was initially our driver. Every time we returned to the car at the end of the day, we found him hanging out with local youth, kicking a ball, chatting. After his ‘promotion,’ he ended up doing excellent interviews in Nyanza-Lac and Bujumbura. In Bujumbura city, I added a few men who had experience working with urban youth – Adrien, the co-founder, about a decade earlier, of a major urban youth organization that had brought gang members together and played a significant role in calming violence; and Lionel, who had worked for years with the most respected youth organization in Burundi, the Centre Jeunes Kamenge.
Sampling
The categories We wanted to interview a cross-section of people who had been refugees, internally displaced, and those who stayed at home during the war. This proved to be a much more elusive task than expected, as it is very hard to put people into categories that make real-life sense. The category of refugee should be clear: everyone who has not crossed the border is by definition not one. But things are more complicated. Take Pierre, a Tutsi and a driver now. Every time things became too hot in Bujumbura, he went on a boat on Lake Tanganyika and stayed there, moving around until things were calmer. One time, he stayed away for a year, traveling as far as Zambia. He never set foot in a refugee camp, never got a return package. Do we consider him a refugee or not?
Things are as hard with internally displaced people (IDPs). Easiest to recognize as IDPs are those people who currently still live in displacement camps, typically close to the communal centers. But many Burundians – mainly Hutu – did not flee to these formal sites, but rather fled away from the town centers, from the police. They are called ‘dispersed’ instead of displaced. This was hardly a permanent condition: after a few days or weeks of sleeping in the forest, they would return home, and flee again at another dangerous moment. So, is someone who lives for nine years in a camp in the same category as someone who on six different occasions fled into the forest for a week? Or is it a different category? How many absences from home make you a displaced person? According to the most recent statistics, 52 percent of all Burundians have fled their homes at least once during the war (MINIPLAN 2006: 31).
And many people fall equally into more than one category. Take the story of Innocent, a thirty-three-year-old farmer and boutique owner:
When the war started I had to flee to a camp for displaced people. There, too, things became very hot, and in 1995 I had to flee to Tanzania. In the refugee camp I lived in awful conditions and had lost all hope. I managed to arrange that I was transferred to a camp for 1972 refugees who had a right to plots of land. I, too, eventually managed to get one. I cultivated it and sold my production. I also did some artisanal jobs and sold the products and began acquiring a small capital to do a commerce of dried fish. When returning two years ago, my money was stolen but I did not abandon the métier. I still had my bike and I borrowed a bit of money and started a little trade of bread and peanuts. This allowed me to live well in the displacement camp with my wife (who had stayed there when I fled to Tanzania: we had not seen each other for seven years and did not know if the other was still alive, and I was so pleased when I found her again and she had waited for me). When the fields started producing well, I sold part of the land to increase my capital, so now I own a boutique and I pay others to help my wife cultivate the lands. I now live in my own house.
Innocent started as an IDP, became a refugee, then got himself into another refugee camp, then returned to a displacement camp again, and then moved into his own home. What category do we put him in?1 Note also the stunning dynamism displayed by this man. This is what survival is all about.
In short, the easy categorization we had in mind proved to be very hard to apply to people’s lives with any degree of realism. It may do the job adequately when it comes to writing project documents or producing aggregate statistics, but it totally fails to capture the reality of people’s lives. Under conditions of protracted war, if one takes a dynamic, rather than a snapshot, view of things, almost nobody’s real life falls within a single category (and the usual statistics we see are meaningless). If we were to focus on the more socially relevant level of households, rather than individuals (as we had so far), this would become even more the case.
The places I chose the rural towns we worked in because they had high concentrations of the different categories of people we wanted to interview. Hence, we worked in Ruhororo because it has the country’s largest IDP camp; Nyanza-Lac, because it has the highest number of returned refugees; and Busiga because it was one of the calmest communes during the war, and consequently most households stayed home throughout. Within each of these locations, we chose two to three collines (literally: hills, the lowest geographic marker for Burundians, consisting of about 150–300 households) and randomly walked through them, day after day, for two weeks or so. We spent about six to seven weeks in each commune.
In each province, we asked the governor and then the communal administrator for permission to work. Most of these people were very open toward us: they gave us suggestions, introduced us to people, and generally cared about our subject. Once in the colline (or the urban neighborhood), we first met the chef de colline to explain what we were doing; often interviewing that person right away as well. Thereafter, we essentially followed a random sampling strategy. We simply walked along minor roads and mere paths and met people who were walking, in front of their house, at the local health center, in the field, etc. Usually, the first day, we were continuously encircled by tens of people, making private interviews pretty hard. By the next day, these numbers would already have decreased – only the kids still run
ning after us. By day three or four, we had become part of the decor: the novelty had worn off. Each day, we would comb a different section of the colline, eventually covering most of it.
TOTAL: 388 interviews, sampled by
Age and gender
• 272 men, 116 women
• 245 young (below 30 years old) and 143 adult
Relevant categories
• 82 IDPs, 45 repatriated refugees, 89 stay at home households
• 217 rural people and 171 urban dwellers (incl. 61 migrants)
• Over-sampling on former combatants: 63 child soldiers and adult ex-soldiers, incl. 17 deserters
• Four economic categories: indigent (80), very poor (143), poor (86) and rich (22)
TABLE 2.1 Rural sample by category
Busiga Ruhororo Nyanza-Lac Total rural
Ex-combatants 6 102 12 28
Repatriated refugees 3 1 35 39
Displaced people 0 37 16 53
People who stayed home 20 303 2 52
Economic migrants 0 0 6 6
Unknown 371 0 3 40
total 66 78 74 218
Notes: 1. The reason we end up with this large number of ‘unknowns’ is that we started by doing many focus groups in Busiga: under these circumstances, we could not ask everyone their personal trajectory. But there is no doubt that these are overwhelmingly people who stayed at home. 2. Most of these ex-combatants lived in the IDP camp as well. 3. Overwhelmingly these are people from Banda colline.
TABLE 2.2 Urban sample by category
Kamenge Musaga Bwiza Other urban Total urban
Ex-combatants 20 9 3 3 35
Repatriated refugees 3 1 1 1 6
Displaced people 16 8 4 1 29
People who stayed home 6 9 8 15 38
Economic migrants 8 33 8 9 58
Unknown 2 2 0 1 5
total 55 62 24 30 171
The urban communes of Kamenge and Musaga were chosen because they reflect the realities of life for the large majority of ordinary urban people. One is almost totally Hutu and the other Tutsi. We also conducted some interviews in Bwiza, a slightly better-off, mixed neighborhood in the center of the city; and with a number of well-off people spread throughout the city – so we had a decent sample of urban people who were non-poor as well.
Age and gender As this research was particularly interested in youth and masculinity, our aim was to talk to two-thirds men, and, within each category, two-thirds young people (below thirty). Contrary to our initial expectation, it was as easy to talk to women as to men. Even in mixed small groups, women talked easily and with confidence. Young women were probably the hardest to reach, although if one could get them alone, they spoke with great ease as well. We ended up very close to our aims. In general, almost half of our sample was composed of young men, although this was more the case in the city than in the countryside. About one quarter was composed of adult men (but as many as twenty-four of those were exactly thirty years old!), and a bit more than one quarter of women, mainly young women.
In Burundi, a stunning 73 percent of the population is below thirty years of age; 46 percent is below fifteen (MINIPLAN 2006: Table 2.7c). Our research focused principally on the fifteen-to-twenty-nine-year-olds – 27 percent of society – the new generation, who grew up during the war, committed most of the violence and suffered most from it, and who will be the builders of the future of their country.
Youths spoke to us with great ease as well, and their analysis was often extremely sophisticated and nuanced. We reached our goal, as two-thirds of our interviewees were below thirty years old, their average age being twenty-two. The average age of the adults interviewed was forty. The average education level of the entire group was sixth grade. Half of all our interviewees were unmarried – a sure sign of being a ‘youth’ in Burundian culture, where marriage is the hallmark of adulthood.
TABLE 2.3 Sample by age
Young men (15–29) Adult men (≥30) Young women Adult women (≥30) Total
Busiga 18 14 19 9 60
Ruhororo 34 20 12 11 77
Nyanza-Lac 27 30 9 5 71
Bujumbura 88 40 34 7 169
Total % 47 26 19 8
Total %, rural 38 30 19 11
Total %, urban 52 24 20 4
We used a simple tool to rank people’s households in income categories. In rural areas, it was based on information about the type and number of animals they owned, whether they had off-farm employment,2 whether they hired people to work for them, and the quality of the house (roof, walls, furniture). In urban areas, we used the last two criteria again and also included the nature of their work (salaried, informal but constant, informal and occasional) as well as their regular mode of transportation (from foot to car). This allowed us to rank people in four categories, from indigent to well off (only 2 percent in the countryside but 20 percent in the city in our sample).
We constantly monitored the interviewee profiles we had obtained and tried to correct if we were far removed from our ideal proportions. The most frequent, hard-to-address imbalance was economic: it was difficult to interview the lowest economic category of (destitute or indigent) people. This difficulty of talking to the poorest is not a surprise. Their houses are not on the road or the path, and they do not stand in the front circle of those who come to greet you. Sometimes the differences were stunning. In the Ruhororo camp, for example, all it required was to move away from the first row of houses along the road: by the third row, one was in a different world, where the houses were all, without exception, shacks with grass roofs, and everyone dressed in rags. People were much more unwilling to talk to us: they hardly looked at us, their long-standing marginality making them feel unworthy of attention, doubt they had anything to say; the low quality of their houses making them feel ashamed to invite an outsider in. In those places, it could take us as much as half an hour to get a person to talk to us. We have interviews there where the person we spoke to never looked us in the eyes, or sat with his back to us – so many signs of internalized inferiority. We did try to do as many of these interviews as possible, but it was a challenge to maintain the energy.
We had early on decided to over-sample two categories: child soldiers and adult ex-combatants. As this research was very interested in understanding conflict dynamics, interviewing this population of overwhelmingly young men was important, and so we interviewed a total of sixty-three of them.3
Finally, within all three of the rural locales, we chose to work in at least two collines: one in the center of town, easily accessible by road, typically close to the communal offices; and one remote, hard to reach. We did this because we were interested in following up on Marc Sommers’ (and, of course, before him, Robert Chambers’) argument that ‘It always rains in the same place’ (2005), i.e. that life in the centers of communes is a lot better than in the peripheries, for it is there that all the infrastructures and programs are concentrated.
Our study confirmed that geographical maldistribution within communes is prevalent. In Busiga, for example, one zone out of three contains the communal offices, three elementary schools, the oldest lycée, a health center, and all of this along a moderately good road that itself is the only decent way in or out of town. In Ruhororo, schools, health centers, and the only two functioning secondary schools in the commune were all in the same zone as well. In Nyanza-Lac the situation was less dramatic. There were secondary schools, as well as health centers and primary schools, in all five zones of the commune.
Interestingly, in both Busiga and Ruhororo, the collines where the future felt palpably better were the remote ones. While there were frequent complaints about missing infrastructures, the economic prospects were better and the sense of future more vibrant in these remote collines. The most dramatic case of this is of course the IDP camp of Ruhororo, located close to the commune headquarters, in possession of a health center and a school, and lying on one of the country’s major roads. Yet some of the unhappiest people we met, with the most blocked sense of f
uture, were found there. Similarly, in Busiga, people in the remote colline of Kimagara seemed more positive in their outlook on life than those in the other collines. In Nyanza-Lac, the remote colline of Buheka was full of people who looked forward to the future, whereas in Kabongo – a former port, close to the center and connected by a good road – the feelings were much gloomier, with many people complaining about the future, and a lot of angry young men. I have no explanation as to why this is so: I think the reasons are idiosyncratic, but they do suggest that the relation between infrastructure neglect and political conflict or grievance is not as straightforward as Sommers suggests.
Location
As stated, Busiga, in the province of Ngozi, was chosen because the war did not hit it hard. In 1993, there were almost no local pogroms there. Busiga is also less poor than the average for the province: according to CARE data, after the provincial capital town, it has the second-lowest proportion of ‘vulnerable households.’ Coffee is very productive there, and its population has a reputation of being independent, engaged in trading and smuggling with neighboring Rwanda. Still, it is a poor place. On average, most of the year, most of its inhabitants eat at most two meals a day, of which neither is balanced and only one contains sufficient calories: five months a year, they are below even that low level (Louvain-Développement 2004). The communal administrator told us that population density is 615 persons per square kilometer – a very high number. The economy is almost entirely dependent on agriculture, principally food crops and coffee and bananas as a cash crop.
Life After Violence Page 4