Snowball sampling, also known as chain referral sampling or respondent-driven sampling, is a common way for researchers to recruit hard to reach populations into qualitative research. The method involves identifying a small number of initial subjects, interviewing them, and then asking them to suggest additional potential participants. The process repeats itself until the sample reaches its desired size. It sounds relatively straight forward but doing it correctly can be surprisingly difficult. The researcher must actively and creatively develop and manage the sample’s initial recruitment, progress and closure.

In the case of post-conflict societies, it can also be ethically treacherous. The local population are in most cases hesitant, distrustful and suspicious when it comes to sharing information with non-local researchers. Chain-based sampling methods need to be adapted in culturally and experientially sensitive ways in order to keep the participants safe and the research on ethically solid ground (see Philipp Schulz’ piece in this series). One successful example of this is the way I gained access to former female combatants and wartime sexual violence survivors in Sierra Leone.

Starting a snowball after a war

The Sierra Leonean civil war (1991-2002) saw many different rebel groups capture girls and women and subject them to sexual violence, forced marriage, and forced labour. Some of those captured also took on active roles within the rebel groups. Once the war ended, these women and (now grown up) girls found it difficult to reintegrate. They faced discrimination and stigma in their communities, and institutional support for reintegration was lacking. It was extremely difficult for them to socially and economically survive in the city and villages. For many the situation has changed little in the intervening years.

I visited Sierra Leone in Autumn 2018 to speak with some of these women. I knew that finding them would be difficult. Seventeen years of recovery and reintegration had already passed. Their history compelled them to keep a low profile. And, even when I found ways to contact them, re-stigmatisation, re-traumatisation and research fatigue would make many of them want to distance themselves from my project.

As a Ugandan woman visiting Sierra Leone for the first time, the success of my project was dependent on managing to recruit an interpreter and transcriber to help with the research. Relying on three interpreters and three transcribers, I conducted interviews in a local language, audiotaped and translated to English.

Research is trust

Success also depended on the development of meaningful rapport and trust between the participants and myself that would inspire them to encourage other women to come forward. Before I left for Sierra Leone, I nursed a feeling that compared to a white western researcher, I would probably have an easier time accessing participants for the study. I believed as an African woman, raised, educated and living in Africa, I would easily blend in and people would look at me and assume I was a Sierra Leonean. Indeed, people often approached and asked me something in Creole. Realising that I did not speak the language, they asked where in Africa I came from and what I was doing in their country.

My identity as a black African woman made my access to these women possible. Because of my accent and physical appearance some people thought I was a Kenyan or Rwandan working with an NGO. I did not mind being presented in this way. But to these women, I introduced myself as a Ugandan researcher working with a University in South Africa. I relied on this racial and continental connection, solidarity and proximity to access and negotiate my interviews. Throughout my study, I felt the women were my fellow black people and so did not feel a huge break between them and me. And while I would be naive to assume that the relationships I formed with them were ever equal, or that I had become an ‘insider’, my identity and character as an African woman aided in forming relationships of trust that were essential to achieving the goals of the research.

Relying on my identity, I accessed participants like Jane (not her real name) who are hard to reach in the city. Now 38, Jane had been captured at the age of thirteen and served four years in captivity as a child combatant. I asked her to introduce me and to encourage her friend to participate in the study. The woman agreed. “This is my first interview”, she said when we first met.

At the end of the interview, I asked her if she could identify women with similar experiences who might be willing to speak with me. She encouraged three of her former comrades to share their stories as well. She also travelled with me to her home district to help me map her journey with the rebels from her village to Freetown. While we were there she also encouraged other women to participate.

Unlike Jane, Musu wore her past on her sleeve and testified at the Special Court. Captured and gang-raped by rebels, this 57-year-old woman is now an activist survivor of wartime sexual violence. She also heads the ‘War Affected Women Association’ – a group of about a hundred former female combatants and sexual violence survivors in one of the resettlement camps in Freetown. She agreed to an interview, and when we were done I asked her if she could encourage other women in the association to take part in the study. Some came forward. But, after learning that most of them had been previously interviewed, I instead relied on this group to establish links and networks outside the camp where I could meet women who had not been previously interviewed.

At the end of the interviews, I asked my participants if they could identify women with similar experiences. Many women encouraged former comrades to share their stories as well. These sorts of chains played out many times over until I had reached my desired sample size. They worked because I was able to develop rapport and trust through my interactions with these women, to the extent that by the end our interviews they were willing to help forward my research.

Aware that asking already interviewed women to ‘recruit’ others would be against research ethics, I asked them to ‘encourage’ women in their networks to come forward. In most cases, participants gave multiple referrals. I met all referrals but only those with experiences in line with the aim and objectives of the study were interviewed. At times this aspect of the chain referral proved a problem, particularly in the villages. Although I always explained my selection criteria, some women favoured their closest friends and kin as referrals. In this situation, turning down some women became a daunting experience for them and me. To minimise potential resentment towards the interviewees in the village, I took the time to acknowledge everyone’s experience.

The women were not keen on formal consent protocols. But I observed how some of them took advantage of some aspects of oral consent, particularly confidentiality. For example, before I commenced with the interviews, the women in the resettlement camp rejected my female interpreter and asked her to leave the room: “we want to talk to someone whom we trust” one of them insisted. I sensed that such responses spoke to their larger experience of stigma and security. That rejecting a non-captive Sierra Leonean woman was to conceal/control information about their past history. That the reaction highlighted the friendship, bond and solidarity shared among these women in post-conflict societies. To resolve this dilemma, I asked them to select someone in the camp they trusted to serve as my interpreter.

Outside the resettlement camp, I also carefully handled issues of confidentiality and intimidation. I was particularly concerned with issues of stigmatisation and victimisation by husbands/family/community for participating. So to ensure their security, I asked the women to choose a safe and convenient interview place and time. Women, especially those who wanted anonymity, asked to be interviewed in places far from their homes. In most cases, they preferred the person who made the referral. To facilitate these movements, I gave all the participants money for their transport and compensation for their time.

As an African researcher, I am mindful how difficult it is for people to talk about issues related to sexual violence both in times of peace and conflict. People hardly talk about such taboo subjects within families, let alone to outsiders. So throughout the study, I acknowledged the task of asking these women to talk about their war experiences for the first time. Thus I was concerned with “[o]nce having opened the trauma, they must return to an often demanding and unsympathetic environment, without a support system to help deal with the flood of strong emotions that accompany or follow such discussion” (Bell 2001, 185). To minimise this risk, I asked the women to have someone they trusted in the interview room to offer them some form of psychosocial support during and after the interview. Many chose their elder sisters or friends. Although I left my contact details with them, after two days, I visited the women to find out if the interview had affected them.

Conclusion

Reflecting on my experience in Sierra Leone, my identity as a black African researcher made my access to former female combatants and sexual violence survivors possible. White westerners continue to dominate the study of wartime sexual violence in Africa. So there is an urgent need to address the diversity of experiences with regards to research methods, positionality, ethics, reflexivity and field dilemma for black African researchers. Such researchers with kinship, race and continental connections to their research communities should write about their field experiences. This will balance the debate about identities and how they relate to research methods, participants and knowledge production.