Methodological and Ethical Reflections on Researching With Children of Accused Witches in Ghanaian Witches’ Camps

Abstract

Motivated by calls for qualitative researchers to document their field experiences and challenges (especially while researching vulnerable participants), this paper reflects on the methodological and ethical challenges we encountered in a qualitative study with children living in Ghanaian witches’ camps — communities where women accused of witchcraft live in isolation. The empirical research upon which this article is based is a wider qualitative study which explored the “children of witches” rights and access to education in witches’ camps. Drawing on our experiences as Ghanaian researchers, we examine five methodological and ethical tensions: (a) seeking parental consent, (b) acquiring children’s assent, (c) negotiating payment and compensation, (d) navigating researcher–child power dynamics, and (e) minimising participant stigma and exclusion. Overall, these reflections underscore the challenges that arise when seeking parental consent within marginalised contexts such as witches’ camps, as well as parental requests for payments before they consent to their children’s participation. Further, we discuss the strategies used to minimise the adult-child researcher power dynamics, such as adopting child-sensitive techniques like storytelling, which enabled children to express themselves more freely. We also describe how community-based strategies — particularly sports and storytelling activities — helped us build rapport and reduce participant risk. Our reflections are based on fieldnotes, debriefings, and community engagement processes during the study. We argue that ethical and methodological decision-making in such contexts must be reflexive, adapting to local realities while protecting participant welfare. We aim to contribute to the ongoing debates and discussions about ethical and context-sensitive research involving vulnerable children in Global South contexts.

Introduction

Reflecting on the methodological and ethical challenges she faced while researching survivors of rape and sexual violence during the Bosnia-Herzegovina war, Clark (2017, p. 1) rightly notes that “researchers rarely write about the challenges that they confront and navigate while undertaking fieldwork. This is a missed opportunity for us to learn from each other.” Accordingly, in this paper, our aim is to contribute to filling this methodological gap in the literature on qualitative research methodology. We intend to do this by reflecting and discussing the key methodological and ethical challenges we encountered during fieldwork while researching “children of witches” in the Gnani, Kukuo and Kpatinga witches’ camps in northern Ghana. Considering the sensitive nature of the topic (participants’ lived experiences of witchcraft accusations), as well as the vulnerable nature of our participants (women accused of witchcraft and their children), we often wondered, before and during fieldwork, if other authors had documented similar challenges. Surprisingly, we could not find published material on the fieldwork difficulties faced in research of this kind. Driven by this gap in the methodological literature, we documented our everyday struggles in the field, aiming to help future researchers when they investigate such sensitive topics.

Research with children in marginalised contexts often presents ethical and methodological dilemmas that are difficult to anticipate. This paper reflects on such challenges in the context of a study conducted with children living in witches’ camps in northern Ghana. The study explored how living in these camps affected children’s education and well-being. While the substantive findings of that study are reported elsewhere, this paper focuses on the ethical and methodological issues encountered during the research process. Specifically, we reflect on five key challenges that emerged during fieldwork: (i) seeking parental consent in a context of marginalisation; (ii) acquiring children’s assent in culturally appropriate and meaningful ways; (iii) negotiating the ethics of payment or compensation in a witches’ camp setting; (iv) managing power asymmetries between adult researchers and child participants; and (v) developing strategies to deal with stigma and exclusion. These reflections are grounded in our positionality as Ghanaian researchers with cultural and linguistic ties to the field sites. Our intention in this paper is to contribute to ongoing debates about how methodology and ethics intertwine in research involving children in vulnerable settings. Throughout the data collection process, both researchers took detailed field notes. We recorded and documented all field observations, as well as our reflections on those events each day. This strategy proved very useful in this specific reflective piece, as it allowed us to recall the various challenges we experienced. The paper is organised into two main sections: the first provides a brief review of the literature to situate our study, and the second examines ethical and methodological dilemmas we faced, as well as the coping strategies employed to deal with them. We conclude by identifying implications for future research in similar contexts.

A Brief Contextual Overview of Witchcraft and Witches’ Camps in Ghana

Defined as the ability of certain people to intentionally cause harm to others through supernatural powers (Gershman, 2022), witchcraft has a long and diverse history (Adolfsson et al., 2024; Hutton, 2017). Within witchcraft cosmology, humankind is said to be “embedded in a cosmic universe of invisible forces, spirits, demons, and the souls of the departed” (Leistner, 2014, p. 55). Thus, the belief in witchcraft essentially involves engaging with the metaphysical —the fear of supernatural forces and spirits. Witchcraft beliefs and practices are widespread in Africa (Owusu, 2023) and constitute an integral part of African cultures (Essien, 2010).

In Ghana, witchcraft beliefs and practices are pervasive (Owusu, 2023), cutting across the geographical South and North. People from all backgrounds, social classes, levels of education, and religious affiliations believe in witchcraft (Adinkrah, 2017, 2019). The dominant attitude toward witchcraft in Ghana is fear because of the belief that witches have supernatural powers to cause misfortune such as accidents, drought, infertility, death, illness, or even the destruction of property (Adinkrah, 2019; Goloova-Mutebi, 2005; Hari, 2009). In southern Ghana, and especially among the Akans, Bayie, Obayifo is used to describe a witch (Adinkrah & Adhikari, 2014). In northern Ghana, especially among the Dagbamba, Sotali and So-tim respectively, denote witchcraft and witchcraft powers (Musah, 2020).

Women who are accused of witchcraft are subjected to intense persecution in many West African countries, and Ghana (Owusu, 2023). The accused witches suffer violent retribution, such as being stoned, burned, stripped naked, lynched or banished out of town (Adinkrah, 2019; Leff, 2014; Roxburgh, 2016; Yakubu et al., 2024). For example, a 90-year-old woman, Akua Denteh, was lynched over a witchcraft accusation in Ghana’s Savannah Region on July 23, 2020 (Forgor, 2023). Again, in May 2023, 70-year-old Sharu Mohammed and 40-year-old Safura Ibrahim were lynched in Zakpalsi on witchcraft accusations (Gbcghanaonline, 2023). Thus, when faced with persecution and to escape anticipated harm (Mgbako & Glenn, 2011), accused witches either flee to witches’ camps or are banished to such camps. Witches’ camps are informal settlements where persons accused of witchcraft relocate to avoid being lynched or assaulted by their families and communities (Agyapong, 2021). Despite campaigns to close down the witch camps due to stigmatisation, 600 accused witches and 300 children live in witches’ camps in Ghana (Mantey, 2012; Marshall, 2023). It must be noted that movements to the witches’ camps were often a form of temporary sojourns but have, over the years, morphed into permanent habitations for alleged witches (Richter et al., 2017).

Children’s Role in Witchcraft in Ghana

Globally, the inclusion of children in witchcraft literature varies. In some cases, children are seen as “practitioners” of witchcraft and are hence accused of witchcraft. In other instances, children are classified as either direct or indirect victims of witchcraft. There are different ways by which children are said to acquire witchcraft powers. For instance, it is believed by the Akans of Ghana that witchcraft powers could be given to an infant right from birth or even in the womb (Adinkrah, 2011). Children are gradually becoming victims of witchcraft. Sinopoli (2021) and Musah (2020) found that children of women accused of witchcraft automatically move to the witches’ camps with their parents to help them there. Thus, children become victims once their mothers or female parents are accused of being witches (Sanou, 2020). The argument supporting this practice is that the alleged witch, together with her family, ought to be shunned (Mkhonto & Hanssen, 2018). Accordingly, the “children of witches” cannot be allowed to stay in the communities while their parents are banished to witches’ camps. More importantly, there is a widespread belief that witchcraft is either inherited from parents or can be learned from parents (Nyabwari & Kagema, 2014; Owusu, 2023). Consequently, “children of witches” are either witches themselves through inheritance or by proximity to their parents.

The Study Context and Approach

This reflective piece is drawn from a wider qualitative study on children’s rights and access to education in Ghana’s witches’ camps. It was specifically based on the ethnographic engagements with children and their parents on the one hand, and selected community elders on the other hand. Grounded in children’s rights to education, the objective of the wider study was to explore the extent to which children’s rights to education were being respected in Ghana’s witches’ camps. Banishing alleged witches into camps has severe implications for their children’s rights, welfare and education. The threat is amplified since accused witches move to the witches’ camps with their children. However, despite the potential impact of this arrangement on children’s welfare, the discourses and empirical studies on witchcraft in Ghana and other African countries have mainly focused on the relationship between witchcraft and modernity and the effects of witchcraft on politics (Adolfsson et al., 2024; Ashforth, 2005). At the individual level, the extant witchcraft discourses have focused on analysing the impact of witchcraft on human development through the lens of older women (Adolfsson et al., 2024; Ashforth, 2005).

The study was methodologically guided by the micro-level approach to human rights. This approach contextualises human rights within the framework of local or micro-level experiences of human rights. To collect data for the wider study, the immersion of one’s self into the world of research participants was very critical. Hence, the deployment of the ethnographic approach, an approach Merry (2017, p. 142) refers to as focusing on “micro-social spaces in context”. This ensured that participants were studied within their natural settings. The approach helped in “analysing a micro-social situation within its larger social context” and also “provides a valuable methodology for examining how international human rights work in local situations” (Merry, 2017, p. 142). Data for the wider study were collected from three witches’ camps: Kukuo, Gnani and Kaptinga. Based on purpose, one-to-one in-depth interviews were conducted with 30 children aged 12 to 17 years (girls n = 17; boys n = 13).

Since the focus of the wider study was on exploring how children’s rights to education are undermined as a result of “living with a witch”, in this article, our goal is to document and discuss the methodological and ethical challenges the researchers experienced while researching this sensitive topic (witchcraft) with a vulnerable group (children of witches, accused witches) in witches’ camps. By exploring the ethical and methodological dilemmas that the authors faced researching the “children of witches” rights and access to education, our goal is to play a role in assisting researchers to prepare adequately regarding the potential methodological and ethical challenges that might come up in the field.

Ethics Approval and Safeguarding Measures

The ethics approval for the wider study, on which this reflexive piece is based, was granted by the University for Development Studies in Ghana. The approval process lasted approximately four months and involved an extensive review to ensure the study protected the rights and welfare of a vulnerable population, especially the children. Key issues raised by the ethics committee included: (a) Informed consent and assent – we clarified that consent would be sought from the children’s parents, while also obtaining assent from the children themselves, using age-appropriate and culturally sensitive explanations of the study and their right to withdraw at any time; (b) Safeguarding and minimising harm – given the sensitivity of discussing witchcraft accusations and stigma, we outlined procedures preventing re-traumatisation (such as discontinuing interviews if distress occurred); (c) Privacy and confidentiality – the committee expressed concern about protecting identities in the camp settings: we addressed this by using pseudonyms in the main study, and removing identifying details from transcripts; (d) Avoiding undue influence – to mitigate potential coercion from providing household gifts (rice and cooking oil), we assured the committee that these would be given only after interviews, regardless of the nature of responses. These safeguards were accepted by the committee as sufficient to protect participants and researchers throughout the study.

Researcher Positionality

Researchers are increasingly required to identify and articulate their positionality in their research because their worldview and the positions they take relative to their research can impact how research is conducted, its outcomes, and results (Holmes, 2020). In other words, researchers ought to reflexively discuss how their position influenced the research process (Berger, 2015). Furthermore, because the researcher is an active agent in acquiring knowledge, it is essential to explicitly acknowledge the researcher’s position due to the nature of the relationship between the researcher’s background and their participants (Bukamal, 2022). In what follows, we briefly discuss our positionality relative to the study.

First, both researchers who undertook this study are Dagbamba and, as such, are natives of the research context. Both of them speak the indigenous language, Dagbani, and understand the local customs and traditions. In this sense, we considered ourselves as “insiders” because we shared these particular attributes with the participants (Braun & Clarke, 2013). Although Dwyer and Buckle (2009) argue that researchers are located either as “insiders” or “outsiders”, in reality, our position in this research was fluid and changing. It was difficult to label ourselves as either complete “insiders” or “outsiders”. While we saw ourselves as “insiders”, the participants saw us as “outsiders” because we are academic researchers and belong to the middle class in Ghana. Yet, during our interactions and discussions with the participants, at various points, they again saw us as “insiders” because we shared common traits such as language, ethnicity, and religion.

Also, the second author has a PhD and an MPhil degree in anthropology, where his doctoral thesis focused on witchcraft accusations and their impact on children’s and women’s welfare. The first author, on the other hand, has both a PhD and an MSc degree in Social Policy with a special focus on children’s welfare and rights. Considering our academic background, both of us were conscious of the power dynamics that are often at play between researchers and participants. This power dynamics generally stems from differences in social class and demographic characteristics (Biswas et al., 2022). We believe that our research participants, especially the adult gatekeepers (accused witches), might perceive us as middle-class. Accordingly, being aware of these power dynamics, we approached both the research topic and participants with a lot of sensitivity and respect.

Third, as already stated, the researchers’ position influences their research. Accordingly, researchers ought to reflexively discuss how their position influenced the research process (Berger, 2015). Since people from all socio-economic strata in Ghana believe in witchcraft (Adinkrah, 2017, 2019), we expected that the accused witches and the children would be cautious of strangers. Initially, when we made contact with the participants in the witches’ camps, they thought we also believed in witchcraft and so assumed we saw them as witches. However, through our interactions with the participants, they realised that we were non-judgemental and that we did not see them as witches. Subsequently, the children and parents opened up to us and became more willing to share their lived experiences in the witches’ camps. Relatedly, Kacen and Chaitin (2006) noted that the researchers’ experiences, worldviews, and backgrounds can influence how they access research sites, collect, analyse and interpret data. Similarly, our background and non-judgmental position on the subject of witchcraft facilitated access to the research participants.

Obtaining Parental Consent

As already noted, the children’s participation was part of a broader study that examined the educational lives of children in witches’ camps. Since the children were under the control and influence of their parents, it was critical, per Powell et al.’s (2018) advice, to obtain consent from parents and assent from children. Thus, access to the children’s participation was mediated through their parents, who in all cases were accused of witchcraft. In the camps, these parents were mostly biological mothers, though there were also other women, such as older sisters and aunties, who acted as parents to the children. As noted by Kennan and Dolan (2017), the requirement for informed consent was a key factor influencing both the children’s willingness to participate in the research and the parents’ willingness to permit their involvement. In this study, we held separate meetings with these parents, where we explained the purpose of the research and the reasons for involving their children. All women who participated in the adult study, and who had children between 6 and 17 years, were asked for consent for their children to participate. One of the main ethical challenges we encountered while seeking parental consent for their children’s participation was that some parents requested compensation before granting permission. This ethical dilemma is discussed in greater detail in the following section.

Negotiating Payment and Compensation

The process of seeking consent from the parents presented an ethical dilemma. Many of the women were themselves victims of social marginalisation, and their decisions were often influenced by concerns about food and perceived benefits. Just like other studies, the authors in this research faced the ethical challenge of offering payments to the research participants. This was unsurprising because payments to research participants have been intensely debated, especially in academia (Largent & Lynch, 2017). The debates relate to whether or not payments encourage participation (Kelly et al., 2017), as well as serving the purpose of compensation for research participants’ time, efforts and sometimes expertise (Surmiak, 2020). There are also concerns regarding payments serving as undue inducements and excessive coercion (see Largent & Lynch, 2017; Surmiak, 2020). These debates have resulted in the emergence of models. The models include the market model, whereby payments are seen as an incentive for participants, and the wage-payment model, which sees research participants as offering services by participating in the study (Taplin et al., 2019). Accordingly, participants ought to be remuneration based on the “services” rendered by research participants (Taplin et al., 2019).

The parents told us that we needed to pay the children before they would allow them to participate in the study. Requests for payments were made by all but two of the parents we engaged in the study. However, it is important to note that the requests were often made indirectly and subtly by the parents on behalf of the children. For instance, the parents often asked in the Dagbani language, “ka bo ka yi yen ti bihi maa?”. Meaning, “what are you going to give the children?”. This indirect request could not be taken for granted. Although a literal translation of the parents’ requests implied that the researcher was free to offer the children whatever they deemed appropriate, our positionality within the research context allowed us to interpret this as a request for financial compensation in exchange for the children’s participation. This presented us with a binary ethical choice: either provide the requested payment and proceed with engaging the children, or decline the request and forgo their participation in the research.

The question of compensation was thus a significant methodological and ethical issue. Oftentimes, there was some level of eye contact between the children and the parents when the parents asked about what we were going to give the children. We observed that eye contact was a subtle way of telling the children to wait while they did the “bargain”. Additionally, according to Dagbamba tradition, too, children are required to remain silent when adults are engaged in conversations, even if the conversations are about children (Salifu, 2010). Until children are invited to talk, they are expected to remain quiet. Nevertheless, three children at various times attempted to interrupt the authors’ conversations with their parents, and the children were met with stern warnings and rebuttals. One of the children was even sent away when he made a second attempt at conversing with the authors, when he was not invited to speak. As Salifu (2010, p. 276) notes, among the Dagomba, “boys and children are not allowed to come near when serious matters are being discussed”, especially by people older than them.

Although we did not initially plan to offer compensation to the participants, there were local expectations of “something small” for participants’ time and contributions. Nevertheless, it was clear to us that it was going to be impossible to get the children to participate in the study without addressing the requests made by their parents. Accordingly, the research team agreed to offer each child GHS 50-100 ¹ . Additionally, seeing the difficult situation in which the children and parents lived in the witches’ camps, we decided to provide each participating child with 5 kilograms of bag rice and 0.5 L of cooking oil as our contribution towards improving, even if momentarily, their living conditions.

An important caveat bears highlighting in this regard. Even though we agreed to “give the children something” after their parents insisted, we were aware of the potential influence of non-monetary gifts on participants’ responses. In line with ethical guidance on child research (Graham et al., 2013; Powell et al., 2012), we did not want to create an impression of payment-for-information. While we wanted to recognise the children’s time and effort, we also understood the risk of incentivising participation in a context of poverty. We attempted to address this by explaining to the children and parents that the tokens were a gesture of appreciation, not payment-for-information. Also, the tokens were only given after the interview session. This meant that children received the tokens after their participation, regardless of how much they contributed, to avoid any sense of compulsion. Ultimately, although rice and cooking oil were provided to the household as a culturally appropriate token of appreciation after participation, we believe these items had minimal influence on children’s responses, as they were not the direct recipients and the gifts were given only after interviews were completed. These decisions illustrate the wider tension between universal ethical standards and local realities. What counts as appropriate in one context may be problematic in another.

Acquiring Children’s Assent

As a principle, the authors wanted the children themselves to also assent to participate in the study, despite the central role parents play in the research context regarding decision-making. We believe that ‘rather than assuming that children assent if their parents agree for them to be a part of a study, it is important that informed consent is gained’ (Smith, 2011, p. 20) from children as well. As Munford and Sanders (2001) note, once children can verbally communicate, researchers should strive to gain their consent. Thus, after securing the adult parents’ consent, we engaged the children to obtain their assent to participate. Assent, as distinct from consent, refers to a child’s affirmative agreement to participate in research (Alderson & Morrow, 2011). The concept of assent recognises the capacities of children and allows them to accept or refuse participation in a study.

We employed age-sensitive and participatory approaches to explain the study to the children. This included using storytelling (mostly) and role play to communicate what the research was about and what their involvement would entail. The idea was to ensure that children understood that participation was not compulsory and that they could opt out at any point without any consequences. Some children expressed hesitation, particularly during initial encounters. They were unsure why strangers were asking them questions. We recognised that merely obtaining verbal assent was not enough; it had to be meaningful and ongoing. Accordingly, to facilitate informed consent, in many cases, we allowed children to observe initial interviews with their peers before deciding to participate. It is worth noting that the assent process often challenged adult authority in the camps. Some parents were surprised when their children refused to participate, having assumed their consent was sufficient. This presented ethical dilemmas around balancing respect for adult authority and upholding children’s agency.

Navigating Researcher–Child Power Dynamics

One of the more subtle but persistent ethical issues we encountered was the power asymmetry between us—the researchers—and the children. Although we share some cultural background with the children, we were still seen as “outsiders” with authority and influence. Studies have shown that researchers must recognise these power dynamics between adult researchers and child participants (Ellis et al., 2023; Woodhead & Faulkner, 2008). Further, such power dynamics are even more pronounced in research contexts where children are socialised to strictly follow adults’ instructions (Woodhead & Faulkner, 2008). Although the impact of power dynamics is real in most research on children, it is especially pronounced when researching vulnerable children (Faldet & Nes, 2021). Having witnessed how parents respond to children’s attempts to voice their opinions, and based on our knowledge and experiences of the research context, we realised how the power dynamics between the researchers and the children might affect the research. As already noted above, in hierarchical societies such as the research context, children’s participation in decision-making, as well as airing their views, is heavily moderated by adults. We were of the view that this reality could impact the research in several ways. For instance, having had their lives dictated by adults, we believed that this would impact voluntary consent, especially as the children were aware that their parents had consented to the children’s participation. Would the children be really able to opt out of the research if they felt the need to do so?

To address the power dynamics between the adult researchers and child participants, we, like other researchers (Ellis et al., 2023; Moore et al., 2018), adopted the strategy of continually informing the children that they could withdraw from the study at any point, without the need to explain why. Furthermore, we reassured the children that if they withdrew from the research, we would keep it confidential and not inform their parents. We also realised that the children had become aware, through other children who had earlier participated in the study, that we were offering food rations to participants as a way of compensating them for their time and participation. Accordingly, to ensure that children would not be compelled to participate against their will, but only agreed to participate because they feared they would not be provided with the benefits, we further reassured them that even if they withdrew from the study, we would still offer them the gifts. Together, these strategies worked, moderating the power imbalance between the researcher and the children, as three children out of the 30 opted out of the study.

Furthermore, since some children learned about the study and the token of appreciation from peers who had already taken part, it is possible that prior knowledge of the study could have shaped expectations and, in some cases, responses. To minimise this influence, interviews were conducted as soon as possible after recruitment, and we varied the order of some questions to reduce the likelihood of rehearsed answers. Moreover, the narrative nature of the questions allowed for a range of perspectives, which helped us identify and probe further. Despite these strategies, we did not find evidence that children were informed about specific interview questions.

Apart from ensuring that children voluntarily participate in the study, given the context provided above, we also took steps to prevent children from answering questions they might deem too sensitive or traumatic. The researchers took a cue from Ellis et al. (2023) and shared phrases such as “next topic, please” with the children when we asked them questions, they could not or did not answer. Although some of the children could verbalise their unwillingness to answer specific questions, we believed that this might be challenging for others. Accordingly, we informed the children that they could communicate non-verbally by, for example, shaking their heads to convey their unwillingness to answer the questions asked. Altogether, all of the 17 children declined to answer some questions, while only four of them shook their heads disapprovingly to some questions. In one particular session, Samira (pseudonym) said “next question” to all but three of the questions we asked her.

Also, the power differential was visible in how children responded to questions, often trying to guess what they thought we wanted to hear. We realised that some of them were nervous during the interviews, while others repeated phrases we had used in earlier explanations. To address this, we adopted child-sensitive techniques such as storytelling, which enabled children to express themselves more freely because we departed from questions and answers towards narrations and storytelling. We also spent time building rapport before interviews began and used facilitators from the community who were known to the children. These methods helped address the researcher–participant hierarchy, but they did not eliminate it entirely.

Witchcraft-Related Stigma and Social Exclusion

Globally, stigma is common among different people, and people may be stigmatised for a variety of reasons. People may be stigmatised for behaviours considered socially undesirable: smoking, injecting drugs, and drinking alcohol excessively (Millum et al., 2019). While stigmatised people can be involved in research (Millum et al., 2019), Millum et al. (2019) note that such research is ethically and methodologically challenging. For example, stigmatised participants may be difficult to recruit. They may be at higher risk of certain harms. In the specific case of children, it will be difficult to create and sustain rapport. In the case of this research, the authors identified that the children experienced stigmatisation from witchcraft accusations. In other words, even though the children were not direct victims of witchcraft, they were stigmatised based on their association with their accused parents.

During our research with the children, the authors identified that the children experienced two interrelated challenges. First, the children were stigmatised in the community because they lived in the witches’ camp with their parents. Many of the children initially exhibited signs of withdrawal and disinterest in participating in the study. For instance, many of the children faced stigmatisation from their peers to the extent that some of them had to drop out of school. Others who were courageous enough to remain in school were often found isolated by their peers. They could not mingle and play with their peers. They were often referred to as “children of witches”, and this labelling was enough to distinguish “them” from “others”; the central focus of stigmatisation. As Link and Phelan (2001) note, stigmatised people are often placed in dissimilar groupings to accomplish some degree of separation of “us” from “them”. Because of the stigmatisation, the children were not able to partake in many social activities, especially in school. According to many of the child participants, their relationship with their peers, especially in school, was discomforting. Many of them stated that they felt more comfortable with their parents. In our conversation with one of the children (Taata), for instance, she stated:

I sometimes don’t go out because I feel that I am not welcomed by my peers. They make fun of me by calling me names. They don’t treat me as a human being because they said that I was living with a witch…that my grandmother is a witch. I like staying with my grandmother despite her troubles.

As the extract above shows, even though her grandmother’s strictness was described as ‘troubles’, she still preferred living with her, compared to being around her peers whose parents were not accused of witchcraft. Many of the children who participated in this study shared similar internalising behaviour traits (such as withdrawal) as Taata. We identified that stigma and its associated problem of social exclusion were characteristic features of most of the children (n = 16). During our interaction with a community member, he lamented that:

They are very sad attending school. Sometimes, they are insulted by their colleagues as if they are alleged witches themselves. Even though the community welcomes all of them when they come, their peers still make mockery of them in school.

Another community member noted thus:

They are usually very timid in school. During the break, they look isolated. This is not good for their future. Their performances have not been encouraging except a few. If they are given the needed support, they can do wonders in school but…I wish that something was done about this by school authorities and parents.

Strategies Employed to Reduce Stigmatisation and Exclusion

Breaking Barriers Using CSOs and RGs

Over the years, the role of civil society organisations (CSOs) and religious groups (RGs) in the witches’ camps has been significant. This is because everyday life in the camps is characterised by the accused witches and their children constantly struggling to address matters such as housing, water and food. Many participants during the wider study recounted harrowing stories of how they managed to survive. For example, the participants explained that they had great difficulty accessing clean, pipe-borne water. Many of them depended on water sources from dams, rivers and streams. Accordingly, in many of the witches’ camps, several hand-dug wells were constructed with the support of some CSOs and RGs. In the absence of state support, CSOs and RGs have stepped in to provide their basic needs. Shunned by their family and community, and without state support, the participants argued that they “would not have survived” in the witches’ camps but the CSOs and RGs. Thus, the CSOs and RGs act as “stabilising” forces (Musah, 2020, p. 31) in the witches’ camps. The central role of CSOs and RGs in the lives of participants has, over the years, created a deep relationship between participants and the CSOs and RGs. The authors therefore leveraged this relationship to create, and indeed maintain, trust and rapport throughout the entire process.

To create and sustain rapport with the children and their parents, the authors relied on weekly meetings of the CSOs and RGs with the children. At those meetings, the CSOs and RGs created a congenial atmosphere for the children to be engaged. For instance, together with the authors, the CSOs and RGs took time to sufficiently explain the nature, risks and possible benefits of participation. This was often done in languages the children understood. Specifically, the Dagbani language was largely used because it was identified as the common and widely spoken language by the children.

The authors took turns to informally discuss topical issues in the camps and beyond, including school life, life in their original communities and tales of Kaptindaringa (local folklore). The authors also equally ate with the children while freely interacting and sharing interesting stories with them. The authors particularly recounted stories of people who were poor and marginalised and yet were able to address challenges in life. The purpose was to motivate the children to look beyond their present conditions, particularly their stigmatised condition. Motivated, the children took turns to tell what their career intentions were and who their role models were. Interestingly, four children saw the authors as their role models. They argued that they also wanted to go to school and to teach at the university. They wanted to make life comfortable for vulnerable segments of society, especially their parents who were accused of witchcraft. They, therefore, saw our activities in the witches’ camps as a moral activity intended to bring hope to people. In a nutshell, children who were identified to be timid began to open up.

Rapport-Building Using Sports-Based Activities

Sports often play a significant role in shaping young people’s behaviour and addressing social challenges (Crabbe, 2009). Beyond recreation, sports activities are frequently used as tools for a variety of social purposes, such as strengthening community ties (Crabbe, 2009), fostering social capital (Ziakas & Costa, 2010), promoting social change and a sense of belonging (Green, 2009), and tackling issues like crime, delinquency, and drug abuse (Hartmann, 2003). What is missing in the literature is how sports can be used to create and maintain rapport, especially when researching with children. Realising that sports (football, to be specific) can be used as an informal means of social pedagogy (Crabbe, 2009), the authors used sports to create rapport among children in this study.

We organised three different sporting (football) events for the children. To ensure that the children and members of the communities participated in the football matches, the events were organised in the evenings (4–6 pm). That is, after the close of the school session and the ethnographic fieldwork. A few sporting equipment were procured for this purpose. Interestingly (elderly women are usually not into football in northern Ghana), other members of the communities, including the elderly, took a keen interest in the events. The teams that won were given various prizes, including toffees. What should be noted is that the involvement of the authors in the sporting activities created and sustained a good rapport as the authors engaged the children throughout the entire research process.

Finally, given the high levels of stigma associated with witchcraft accusations, it is possible that some participants shaped their responses in ways they perceived to be more socially acceptable to us. While rapport-building, the use of open-ended questions, and the emphasis on voluntary participation helped reduce this risk, social desirability bias cannot be entirely ruled out in this study.

Conclusion

In this paper, we have reflected on the ethical and methodological challenges of researching with children living in Ghana’s witches’ camps, focusing on five key areas: seeking parental consent, acquiring children’s assent, compensating participants, addressing power dynamics, and dealing with stigmatisation. These reflections emerged from our efforts to conduct research that is both ethically sound and methodologically robust within a social setting of exclusion and marginalisation. Among others, we have discussed how our positionality impacted our research, how we addressed the power dynamics between adult researchers and child participants, and finally, how we adopted several trust-building strategies while researching stigmatised children. Our experiences demonstrate that ethical research cannot rely solely on formal frameworks dictated by ethics committees. Instead, it demands sensitivity to the lived realities of vulnerable populations, particularly when working with children in difficult contexts. We found that standard ethical protocols had to be adapted in response to local expectations and cultural norms. For example, the decision to compensate participants, or to engage children through sports and storytelling, was shaped more by local dynamics than by pre-set ethical protocols. Our study highlights why it is crucial for researchers, especially qualitative researchers working with sensitive topics and vulnerable participants, to be well prepared for the challenges discussed in this article and others, and to make plans in advance for how to address them. We hope that these reflections will enrich broader discussions about conducting research with marginalised children, particularly in Global South contexts where standard ethics guidelines may not fully reflect the complexities of local realities. Finally, for future research, especially in contexts marked by social exclusion and marginalisation, it is important to centre the voices and needs of participants in designing ethical safeguards and methodological tools. Researchers must balance adherence to universal ethical principles with an awareness of the context-specific risks that might arise.