Introduction

Nobody has come here to interview us about how we went to school and the obstacles that were put in our way by our families and community. In this town, both males and females who went to school became prominent in various fields of endeavors. We all had our share of becoming first in vital positions in this country, but the females fought a different battle and worked twice as much to get there. Why do they write about only the men? I am glad you are here to take our stories. Please write them and make sure that our names and other identifying details are clearly stated and mentioned in your thesis and other publications. Our lives and experiences are an open secret being told young ones orally, but our names do not appear in the books and other writings. There are no monuments to honor us too (Akuffo, 2015).

The extract is from a research participant I met whilst conducting research for my master’s thesis in Ghana. Researchers would be excited when they come across a participant in the extract above, who do not only meet their inclusion criteria but is also willing and enthused about participating in their research project. The situation above, nonetheless, may put the researcher in a complex ethical dilemma. It raises questions about privacy and confidentiality which is maintained by anonymizing or de-individualizing data. It also draws our attention to research ethics boards and their position on anonymity.

Over the last decades, there has been a significant push by ethics institutions to make anonymity or confidentiality a default position for ethical clearance. This is based on the consensus that researchers owe a duty of care and protection to research participants (Kelly, 2009). The practice has consequently become a routine response, easily promised, casually adopted, and often uncritically accepted if one intends on completing ethical forms (Atkinson, 2009; Bosk, 2008, p.92; Jacobson et al., 2007; Kaiser, 2009; Librett & Perrone, 2010; Wiles et al., 2012). Operationalizing anonymity, however, go beyond the routine practice of promising anonymity. This has sparked debates in the literature on anonymity. Scholars have reached conflicting conclusions about anonymity.

Some scholars believe that adopting anonymity as the standard default position in research is key to ensuring research integrity. This school of thought believes that anonymity is important in guaranteeing privacy, confidentiality, and the protection of participants. Anonymity is key for the protection of varying participants such as children (Alderson & Morrow, 2011; Bradbury-Jones, 2014; Furey et al., 2010; Tisdall et al., 2009), older people with vulnerabilities (Kipnis, 2001), and ethnic minority (Karnieli-Miller et al., 2009). Anonymization is important in providing safe spaces for research participants who offer counter narratives on dominant political positions, such as wars and other sensitive topics, such as religion and politics (Stein, 2010). Data on such sensitive topics is considered Footnote 1special category data. Research participants in such situations are encouraged to participate in the research once anonymity is assured (Kelly, 2009, p. 440).

There are, however, other schools of thought which question the assumptions underpinning anonymity. Walford (2005), for instance questions the rationale behind anonymizing. He argues that anonymity is objectionable and usually unachievable (pp. 83–93). Commenting on the difficulty in achieving anonymity, Van den Hoonaard (2003), similarly argues that anonymity is undermined in various stages of ethnographic and qualitative research process. Specifically, attempts at ensuring anonymity leads to the underuse of data (Thomson et al., 2005; Van den Hoonaard, 2003, pp. 141–151). It also leads to the decoupling of place and the removal of contextual, historical, and geographical information that has potential value for the exploration of the subject under study (Nespor, 2000, p. 549, Thomson et al. 2005). The difficulties associated with anonymizing historical and geographical information hamper wider dissemination of qualitative research output (Tilley & Woodthorpe, 2011).

Crow and Wiles (2008), in objecting to anonymity argue that whilst the goal to anonymize is predicated on the protection of research participants, anonymization may rarely protect research participants from the right to rebut, libel or reply. Further, researchers also risk legal action from incomplete anonymization. Does complete anonymity imply both external and internal confidentiality (Tolich, 2004, p. 101)? Researchers casually promise confidentiality without an examination of who they want participants data to remain anonymous to. Is it the University or research institution who are usually the data controller of research done under their auspices? Is it to make data anonymous to other participants, outsiders, or dissemination outlets (Kaiser, 2009)? Overall, it is observed that the nature of research and circumstances or context presents practical challenges to anonymization in certain situations (Coffey et al., 1996; Kelly, 2009; Walford, 2005).

Even though the sustained debate in extant literature on anonymity, has centered the justification of the routine practice of anonymity on one hand, and the impracticability on the other, the literature also addresses anonymity and the assumed power dynamics between the researcher and the participant (Jacobson et al., 2007). In furthering the debate on naming participants, another set of scholarship forefront the benefit that may accrue to participants if they are named or given credit for their data. Svalastog and, Eriksson (2010, pp. 104–110), argues for disclosure of participants to avoid situations where the benefit of owning information through the commercialization of Indigenous knowledge shift to the author. Others have associated being named with pride and having a voice (McCormack et al., 2012; Yanar et al., 2016, p.125). Whether the argument on naming is about benefits or pride, allowing participants to choose how their data is managed is also about democratizing the research process to ensure greater collaboration between participants and researchers on ownership of data. The degree of involvement also has implication on the power dynamics between a researcher and participant.

The attempt to limit the presumed power of the researcher and to give participants real choice about the levels of participation seems to be coalescing around two main things. That is expanding the Footnote 2rights of participants and guaranteeing privacy and confidentiality. This is echoed strongly in research ethics literature because of the potential to curb the power conundrum, which scholars have described to be the genesis of all ethical imperatives in research (Haggerty, 2004; Wax, 1977).

Questions arise relating to who determines which data is anonymized and how the data is anonymized. Other questions pertain to identity and voice. Coffey (2012, p. 2), aptly captures this when she asks, when we anonymize, are we giving or taking voice? The research situations in which participants need safe spaces to tell their stories are also situations where participants voices are marginalized (Laoire, 2007). In this context, if the default position of a researcher is to anonymize, then are we stealing voices or giving voices? What if the participants have personal stake in the research? Do we respect their autonomy and the right to make confidentiality decisions? Dealing with such complexities can be challenging and makes the goal of achieving confidentiality impractical (Giordano et al., 2007).

As profoundly revealing as the scholarship in support of anonymization, and the discussions around the need to rethink the application of anonymity are, much is left out. What happens when participants reject anonymity, the institutionally accepted protection protocol needed to ensure confidentiality? The literature on participants autonomy to waive confidentiality is replete. Some argue that participants autonomy is not absolute and should be considered in relation to other moral obligations (Beauchamp & Childress, 2001). For example, Giordino et al. (2007, p. 273) encourages researchers to explore the asymmetry of knowledge between the research participants and researchers regarding their ability to negotiate the implication of waiving confidentiality to decide whether to respect participants choices or override them. Others suggest the burden of such decisions be shifted to the research participant by allowing them to choose exactly how they want their data managed (Anabo et al., 2019; Kaiser, 2009; Ryen, 2004). The central message of this paper is not to debate which school of thought is right on the issue. Rather, it argues that if the urges or motivation behind such participants choices are known, we can better apply their choices in writing research reports or disseminating research findings.

Whilst participants have the right to waive confidentiality, there is little regard for the considerable variation of motivations that goes into participants’ decision to participate in research, and the implication on the safeguarding principles such as confidentiality, anonymity, and privacy. As such, there is minimal guidance in research ethics literature to navigate the loops. Questions that could arise from the extract above remains unsatisfactorily answered. Questions include: Why do individuals participate in research? Why will they refuse anonymity and or confidentiality? Who exactly do they want their personal data to remain visible to? Is it the research institution, other participants, Journals, other dissemination outlets?

What ethical dilemmas could arise from such refusal? What meanings do diverse groups or context attach to confidentiality? What is privacy to individuals in different contexts? Once participants consent has been sought pre-commencement of a research, can participants renegotiate the terms of, and nature of their participation going forward? How do we enable participants to share in the construction of a study aside the interview data they provide without compromising on institutional ethical clearance on anonymity, confidentiality, and protection of research participants? How does cultural context determine the application of anonymity and confidentiality? And how can ethical dilemma arising from the application of anonymity and confidentiality be renegotiated when doing research in the social world? Do we just adhere to participants right to waive confidentiality without an understanding of their confidentiality or no confidentiality choices? I am yet to come across any code of ethics that specifically asks researchers to investigate to enable an understanding of participants motivation for their confidentiality choices. In adhering to such rights to wave confidentiality, this manuscript calls for researchers to explore the motivations behind participants choices within their socio-cultural contexts to determine whether waiving confidentiality serves the ethically sound interest of participants.

We can draw on researchers’ post research reflections to further the debate. In this paper I reflect on my experience as a qualitative researcher whilst researching the first generation educated women in Ghana and the extant literature on research ethics and anonymity to further the debate. I argue the importance of exploring the motivation behind research participants interest in taking part in the research, and the socio-cultural context of the study area to enable a smooth operationalization of anonymity.

The paper is structured as follows: I begin by delineating the context of the research and the research design. Subsequently, I reflect on the context of researching the first generation of educated women in Ghana. This is followed by a reflection of the (un)anticipated ethical issues. I consider the research participants wish to have their names and personal data disclosed in all research outputs and why a rejection of confidentiality may lead to ethical dilemma. Further, I explore what privacy means in different socio-cultural contexts and how that influence participants acceptance or rejection thereof under contextualizing confidentiality. Drawing on my experience and extant literature, the discussion section explores nuances in confidentiality and the practical guidelines on administering confidentiality. I conclude by suggesting practical solutions to overcoming the ethical challenges. By reflecting on my fieldwork experiences, the goal is to address contextual ethical dilemmas or questions that confront researchers when operationalizing anonymity in a practical way.

Context: On Researching First Generation Educated Women (FGEW) in Ghana

The research upon which this paper is written was done in 2015 for my master’s thesis. I researched older women who were the first generation of educated women in their families to acquire education up to the tertiary level. The research explored women’s access to education for those who were born between the years 1917 to 1957. The aim was to uncover the micro-politics around their access to education. Researching this group of women was necessary for three reasons. First, mainstream education literature rarely examines the lives of older women, and their education access experiences. Second, they were the first to acquire tertiary education despite gatekeeping systems which allowed men access to education and impeded women’s access. Finally, exploring inter-household micro-politics of girl’s education was a paradigmatic shift in how access to education is defined. Access is usually defined along large scale macro-structural issues such as education reforms, school resources, infrastructure, fee-free policies, and supply-side policies such as conditional cash transfer (Ahadzie, 2000; Conn, 2014; Garlic, 2013; Glewwe et al., 2014; Kremer et al., 2013; Krishnaratne et al., 2013; Masino & Niño-Zarazúa, 2015). The goal of this project was, therefore, to extend the debate by focusing on micro-level household social relationships and to define access along relationship construction at the family and household levels and the gendered ramification thereof. It also focused on the history of women and education in Ghana.

I submitted my proposal to the Institute of Statistical, Social and Economic Research (ISSER) of the University of Ghana and received ethical clearance on March 27, 2015. Further, I presented my proposal at the University of Ghana’s Department of Sociology seminar. Two key issues were raised. First, how I plan to gain access to my targeted participants since there were no relevant data sets on the FGEW, and second why I did not name the study location in the thesis title. Following their advice, I narrowed the study to one traditional area. I was, however, reluctant to include the name of the study location in the thesis title because I wanted to ensure double barred protection, a disguise, or a type of conventional confidentiality where the study sites, and the research participants were all anonymized (Kaiser, 2009; van der Geest 2003). I included the name of the study location, Kyebi eventually given the consensus among faculty members in my department to do so. Their arguments were two-fold. Firstly, naming the location was important given that I was dealing with a historical event and an elite group of people known for becoming first in various fields of endeavors. This suggestion concurs with Nespor’s recommendation that anonymizing place could decouple events from historical, geographical, and other contextual details (2000, p. 549). Svalastog and, Eriksson, (2010) makes similar arguments that removing contextual identifiers or abstracting data could decontextualize data from its socio-historic context.

The process with the ethical review board and the departmental inputs I received from the seminar helped shaped my research at the planning, designing and conceptualization stage. I anticipated and planned for some of the ethical issues whilst others emerged during data collection and even after the research has been written up.

I researched fifteen older women at Kyebi whose ages ranged between 64 and 104 years. Kyebi is the traditional capital of the royal Akyem kingdom in the Eastern region of Ghana. It is a typical matrilineal society where kinship ties and family ties connected the people. I gathered data through exploratory autobiographical narrative which enabled the research participants to recollect and re-tell their experiences of education access (Connelly & Clandinin, 2006, p. 477; Saleh et al., 2014, p. 272), informal conversational interviews, and rarely semi-structured interviews. All fifteen participants gave their informed consent before the interviews, and participation was voluntary. I interviewed the women about the micro-politics that played out in their families around their access to education. To enable me to explore the in whose interest issues and access to education as embedded in relationship construction, I interviewed every family member alive in the FGEW’s family who took part in the politics of education access including the few key decision-makers and gatekeepers of education access who were alive. I conducted the research within families who were royals and non-royals because of the differential education access experience of the two types of families.

The participants shared their experiences of education access and re-told their stories about the resistance activities that some family members undertook against gatekeepers of access in their families when women became the first casualties of non-enrolment education decisions. In the cause of data collection, participants were forthcoming with information about other family members (decision-makers and gatekeepers) who permitted, denied, and or enabled their access to education. The interviews were emotive in nature and the extent of disclosure was troubling because of disclosures of conflicts among family members concerning resource dilution within families. The need to de-individualize their data to ensure confidentiality was imperative to prevent potential rife within families that may occur if anybody could identify the participants in the research output. But how do a researcher reconcile participants’ interest to have their names and personal data published in all research output, versus the standard ethical practice to protect participants by concealing their identity.

My assumption was that once the purpose of the research, the procedures to ensure anonymity, confidentiality and the proposed outcome of a study is transparently shared with targeted participants, they would be encouraged to take part in the study. This assumption however oversimplifies the considerable variation of motivations that goes into research participants’ decision to participate in research. Such participants motivation could lead to unanticipated ethical issues which could impact negotiating and renegotiating of confidentiality, access, and consent.

(Un)Anticipated Ethical Issues and the Re-Negotiation of Consent

The major issue I anticipated before fieldwork was the age of my participants (64 to 104) years and the possibility of frailty, memory loses and other types of vulnerabilities. Commenting on age of research participants, Birt & Poland (2018, p. 2), posit that chronological age alone is not a reliable marker of vulnerability. I considered other factors such as autonomy (Agich, 2004; Beauchamp & Childress, 2009), living in their own homes and not confined to institutional settings (CIOMS, 2016), and free from other types of vulnerabilities such as medical and deferential (Kipnis, 2001). I also anticipated that the interviews for the research could not be conducted in the public space. This is because it involves interviewees sharing information about others in a potentially sensitive family conflict situation. The potential emotive nature meant that the interview location was important (Mani, 2006). Thus, I suggested a neutral venue or asked the participant to choose their own venues. All participants chose their homes. Within ethics literature, interview venues are considered micro-geographies where power relationships are enacted and constituted (Elwood & Martin, 2000). Participants, therefore, had more control over the venue process.

Further, I negotiated longer interview duration and accommodated the need to take pauses and long breaks in instances where memory loses were detected. I had a participant who had challenges recalling details of micro-level household education access events. I relied on other participants she mentioned in her education access narrative for data triangulation.

There are issues that cannot be anticipated before fieldwork due to the nature of human social subjects. In my situation, it was difficult to determine unequal relationships as the research participants had contradictory status. Their ages could be construed as vulnerability. Their status as generally educated women, accomplished, and well known in their community, however, could put them in a privileged class. Whilst I did not think their ages presented unequal power relationships, they could be considered elite research participants. They are drawn from a small sample to partake in the research because being a first generation educated woman mark a turning point in the history of education in their families and community. Much of the personal information that can make them easily identifiable is common knowledge in the community. The fact that the interview is done with any of these 15 FGEW is central to understanding the history of women’s education access in this community. Thus, they satisfy the criteria to be considered elite research participants. Consequently, de-individualizing their data or ensuring complete anonymization of these participants was going to be almost unachievable (Walford, 2011; Woods, 2012, p. 4). I was going to anonymize their data anyways even if that meant under using the data or limiting the analyses. My approach at ensuring confidentiality and privacy was to Footnote 3anonymize the research participants. Complete anonymization was, however, impractical since I named the study area. Revealing the study area in the research output was important in understanding the socio-spatial relations that underlie access to education for women (Jessop et al., 2008). Against the backdrop of avoiding rendering the social relations hidden (Nespor, 2000), in the study that sought to construct negotiation of access to education as relationship construction, I could not anonymize both place and participants. I opted to anonymize participants only.

I deliberated with research participant about confidentiality, identification, and the handling of their data. I proposed that their personal data would be anonymized. However, whilst they were enthusiastic about participating, they rejected anonymity. Participants mentioned other family members and how they actively worked to deny their access to education and the consequent conflict over resource dilution within the family. I was going to employ generic codes such as FGEW one, FGEW two, but they insisted that I add all their identifying details in the research output. I explained the risk they will be put into if their personal data is disclosed but they shared contrary opinion. Surprisingly, the desire to have their names and other details published in the research outputs was consensual amongst all 15 FGEW. I felt a sense of powerlessness. This was because I was trying to meet the ethical standards based on my institutional guidelines while attempting to find ways to accommodate participants request. I could not accept their no-confidentiality, or no anonymity clause. Thus, I renegotiated consent. We agreed on employing pseudonyms.

Renegotiating Informed Consent

During the data collection, I began to really assess the possible implication if the personal data, fully identifiable is linked to a participant or their families. That notwithstanding some of the participants reiterated their desire to include their names in the research output. The access stories they shared involved other family members and conflicts over resource dilution. As a researcher, I had thought about what disclosing their identities in the research output could mean, such as the potential divisions that could be created in their families. Consequently, I could not honor their request to use their personal data in the manner they proposed.

I was concerned that some participants may terminate participation if I de-individualize their data totally. I renegotiated consent. Renegotiating consent involved relying on networks. I have lived in the community all my life; thus, I had networks that aided re-access and re-negotiation. Everybody who heard about my research wanted to help. I also relied on my shared identity such as ethnicity, gender, and my similar education position. My shared identity defined my encounters in resolving the issue and served as a unifying factor. This affirms (Soni-Sinha, 2008; Chawla, 2006; Few et al., 2003; Mani, 2006) assertion that characteristics such as age, gender, and education shape encounters with the researched. I argue that these markers coupled with being local can be leveraged in negotiating access because it helps ensure trust (Feldman et al., 2003; Russell, 2005). Renegotiating access also centered what names I could use to replace their names. They seemed happy about the suggestion to replace their names with other names typical to the Akyem matrilineal society so that readers can put it into context. Possibly, including the name of the town in the thesis title partially satisfied their request to share their stories and be publicly acknowledged in academic work.

Mani (2006), commenting on the emotive nature of interviews suggest that participants could be oblivious to it. Relying on Davison (2006), proffered solution to seek consent in a continues manner, I reminded participants about consent during the interview process and after data collection. Though the participant were happy about their participation they tried to renegotiate no confidentiality at every given opportunity. I did not think they were making a rational choice by their demand to have their identifying details published in all research outputs. I had a duty to protect participant in accordance with beneficence as emphasized by Belmont report (Anabo et al., 2019). Therefore, weighing the benefits and balancing them against the potential risks of the internal family conflicts I envisaged, I chose to protect them using Footnote 4pseudonymity as a confidentiality card as initially agreed upon (Beauchamp & Childress, 2001). I take responsibility for the ultimate decision on confidentiality. This is because the researcher decides in the synthesis of the findings what needs to change, how much information or data they would change whether person specific details or contextual details and the implications on the research (Parry & Mautherner, 2004; Wiles et al., 2008).

Reality After Research was Written-Up and Submitted

The thesis was written-up. I gave a summarized report of the findings to participants as they had expressed interest in reading the findings, the statements ‘I would like to read the book after the research. …how do we get to read the book about us? were questions they repeatedly asked every time they saw me in the community. Whilst sharing findings with participants is important for respondent validation (Silverman, 2000) and generating additional data (Lacey & Luff, 2001), in this instance, it was about encouraging participation, making sure the findings resonated with participants expectation and the co-production of knowledge. Whilst participants were happy about the findings, the issues of confidentiality surfaced again. All participants expressed some level of disappointment with the pseudonymization of their identities.

A granddaughter of one of the FGEW I interviewed approached me and narrated:

My grandmother has been so happy because you came to interview her. She is so happy about your work, and she said she could find herself in your book. She was looking forward to her name being associated with some of the information presented but she says that you gave her a different name in the book.

They expected their names mentioned in the work even though I made it clear that it was not going to be possible because of ethical concerns. One participant told me that even though you have changed all my identifying details and given me another Akan name, could you not have just used my name, when I die, who will tell the world that I was part of this memorable thing and that is my story in your book.

In keeping up with what participants wanted, I made sure the presentation of data in all research outputs threaded participants voices as much as possible, quoting at length from their narratives. This was my way of bringing their stories to light as much as possible (Forbat & Henderson, 2005, p. 1125).

I soon realized after the research has been written up and submitted that the research data could be linked to some of the first generation educated women. I did not bridge the protocol on anonymity. In fact, even though I quoted the participants at length in line with their demands, I still sacrificed some of the direct quotes that were just too person specific. I did not release linkable data in the process but some people who read the thesis could identify people they knew in the thesis work. However, there was no way I could have written about them and achieved complete anonymity or protected them from being identified regardless of how the research was presented because of the nature of the research topic and the context and history of the study area.

What about the participants, the topic and the context made the research participants easily identifiable? First, FGEW, and the name of the town or location in the title rings a bell. The study was about uncovering the micro-politics around education access negotiation for a group of women I called FGEW who were extremely popular and the story about their achievements well known in their community. Thus, they were a selected few who satisfied the inclusion criteria such as marking a turning point in the education histories of their families. This concurs with what Woods (2012), terms as elite research and the difficulty in achieving complete anonymity in such research work. Further, they became prominent people and some of them first in various field of endeavors in Ghana after independence. Deductive disclosure was thus possible (Sieber, 1992; Tolich, 2004). They could be identified through some of the personal identifies that was not altered e.g., age, profession, number of children. Additionally, families at the community levels are easily identified with their family professions. Thus, it was easy to find families who have been lawyers or families that are into national politics or families that have had three generations of people go into similar professions. Moreover, the stories about their lived and perceived experiences are unique to their generation. Residents of the town, especially the elderly ones, would have no challenge connecting the stories in the thesis to the participants.

Second, this was possible partly due to the context. Identifying participants in research through contextual identifies is described as inductive disclosure (Tolich, 2004). First, the study area or polity has always had a structured chieftaincy political organization headquartered at Kyebi in the Eastern Region of Ghana. The town became prominent in the history of Ghana in many ways. Furthermore, macro, or structural education access issues such as infrastructure were resolved in this area even before Ghana attained independence in 1957 (Kwakye, 2007). Coupled with the political leadership’s resolve to get those in the Kyebi polity educated, several people from this area became first in national endeavors. How the FGEW received education is thus, a story that has been told and linked up with resolved macro-access education concerns such as school infrastructure and meso-level access issues such as scholarships.

Whilst the meso-level concerns on access once resolved enabled boys to go to school unimpeded, it was the beginning of education access negotiation for girls at the household levels. They needed permission from gatekeepers, decision-makers and overseers of family resources and education access in their families to go to school. I have written up about the gatekeepers, and the general micro-politics of girls’ education in Kyebi in that generation elsewhere in an upcoming paper. These contexts and the unique and unusual education access process makes them easily identifiable. Thus, anonymizing participants would not hinder those who were familiar with the history, culture, structural education access issues and the local politics of the study area from connecting the dots.

What was missing in the stories that were told about their education access was the micro-level access realities within households and families. This knowledge gap was what my research attempted to bridge. In fact, issues such as the politicking around access and the resolve by some family members to deny access to female members of the families (both royal families and other non-royal families) was an open secret that was informally shared and discussed among town residents. Collecting data on these meant collecting data about sensitive issues of conflicts around resource dilution in families and the potential politicking around who benefits from education access regimes within families. Uncovering these and making the lives of these women visible through written work regardless of how I tried to conceal their names and other details to protect their privacy and ensure confidentiality was going to be incredibly challenging. These issues, however, only became apparent after the research has been written up and submitted.

Contextualizing Confidentiality Based on the Cultural Environment and Participant Motivation for Participating in the Study

Ensuring absolute confidentiality through anonymity is not always possible because of deductive and inductive disclosures. Therefore, the idea of participants refusing or rejecting measures that would ensure confidentiality even if ethically cleared, in the procedural ethics process is concerning. Choices about anonymity, confidentiality, and other aspect of the research design is taken before fieldwork and submitted to mandatory institutional research ethics committees for approval (Jacobson et al., 2007). Because of the blanket clause to anonymize in many institutional ethics, researchers are compelled to anonymize without a critical reflection of whether the context provides a compelling reason to do so. Kaiser (2009) calls it procedural ethics. Other scholarship examining ethics review processes have described it as an absurd administrative constraint. In Bosk (2004), scholars commenting on this constraint states that some ethics review practices are “one more inane bureaucratic requirement in one more bureaucratic set of procedures, ill- suited to accomplish the goals that [they are] intended to serve (p. 417).

On the field however, one is met with the differential participants perspectives, culture of the people and the individual preferences about managing their data in ways that ethical regimes provide minimal guidance on.

Whilst providing anonymity may encourage participants to participate in the research (Kelly, 2009, p. 440), especially in cultures where individuality and privacy is highly priced (Crow et al., 2006; van der Geest 2003), it may be different in other cultures. Presenting anonymity as a measure to get participants in communities with cultural variations may prove to be counterproductive. Research participants who want recognition for their data would reject anonymity (Shulman, 1990). Just like the Kyebi situation where participant refused anonymity because they had personal stake in the research.

What do you do when participants reject such safeguarding measures? How do we achieve anonymity when participants do not wish to be anonymized? Our institutional rules and personal notions of rights and wrongs could be tested in such situations because ethics regimes assumes that participants want confidentiality (Crow & Wiles, 2008, p. 5; Grinyer 2004). However, practical fieldwork situations demands that we consider rights and research ethics through the individual participants relativistic lens, so that we do not engage in undue paternalism.

Reflecting on my role, was I ethnocentric in applying institutional ethical principles on participants who intimated to me on too many occasions that they did not want it? Should I have given in to the participants demand about processing their data outside of what I got ethical clearance for? Was I too careful about protecting them from assumed potential risk, I had not carefully examined? Was I tacitly cohesive in de-individualizing their data? Did I employ confidentiality against participants will? Further, I thought about what it meant for power relationships and the co-production of knowledge that researchers are encouraged to addressed during data gathering (Oakley, 1981; Ribbens & Edwards, 1995; Stanley & Wise, 1990, 1993).

I have struggled with this dilemma since 2015 and there are always more questions than answers as to whether I was tacitly coercive in my approach to dealing with the issue. We could approach this by reconciling institutional expectations with the ethical standards of research participants and their communities.

Research Purpose Achieved, was Pseudonymizing Participants and Maintaining the Name of the Town Helpful?

One of the main purposes of researching the FGEW in Ghana was to make sure that the micro-politics about their education access was not missing in literature. It was part of bridging the gap in the literature of education which has often concentrated disproportionately on researching macro-level education intervention as a uni-causal input of access. I suppose the purpose was achieved. Most importantly, an older generation about whom little was known in written work was written on. Some scholars have argued that there is limited representation of women and other minorities who are othered (Archer, 2007; Jensen & Lauritsen, 2005; Knight et al., 2004). Indeed, within education research, women’s access experience and the micro-politics and contentions around their access remain marginalized beyond the economic rational postulates. Thus, by concentrating on this population and legitimizing their subjective knowledge, I had done my duty. The lives of the FGEW could be made visible.

Was the research participants objective of participating in the research achieved? Two main motivations underpinned their willingness to participate. Apart from helping me with my research, their motivation was to get their access story told. The research accounted for the differential micro-level education access situations of the participants. Moreover, I made sure the presentation of data threaded participants voices as much as quoting at length from their narratives. This was my way of bringing their stories to light (Forbat & Henderson, 2005, p. 1125). I admit however, that there were other details that I did not include because it was too person specific and thus could be enough to identify the participants easily.

Even though the research purpose was achieved, I decided to evaluate whether others aside the research participants could easily identify the participants. Thus, when I went back to Ghana for my DPhil. Thesis data collection. I gave a copy of my thesis and the subsequent manuscript I had prepared for publications to my mother to read. My mother is in her mid-sixties, and she has lived in Kyebi for about half of her life. After reading them, however, she began asking questions: were you talking about this person in this paragraph? This information looks like it will have been given to you by this person, …but I am so certain she could be the only one to have provided you with such an information.

Was the easy identification of research participants widespread? I gave the thesis to two other people, one prominent figure whose family had lived in Kyebi for about a century and had a great grandfather who was persecuted during the infamous persecution of Kyebi Christians (Kani,1975). He found the thesis profoundly revealing. He also began to mention names of people I had interviewed: you have done well because we know them, but this aspect of their lives is unknown to many except the idea that they were the first generation educated women. Their stories are usually told with distortions and made to look like they went to school unimpeded. I was worried. I gave the manuscript to another person in his early thirties but though he could tell the stories were about the educated women of old in the town, he could not figure out who the actual participants were.

Whilst the feedback I got meant that at least the stories have been told accurately, the idea that the participants could be easily identified by any native who read the thesis was concerning. What then was the point in employing confidentiality? This unintended implication of the research was a lesson I learnt too late.

Was it Fatal? My Mental Dilemma Continued

I thought the purpose of pseudonymizing the research participants had been defeated. What are the implications? Maybe using general codes would have worked. Why did I reveal the name of the town in my research output? Could changing the name of the town have worked? As I reflected on these things, I asked myself why I thought using a single disguise (pseudonymity) could work instead of using double disguises (anonymizing both the study location and participants). The answer is not straight forward.

The ethics committee who approved of the research suggested I contextualize the research as there was no relevant data set on the population I was considering, by including the name of the study location. My department also insisted that I included the name of the town in the title to recognize the geographical situatedness of the study. It is almost impossible to provide double barred protection when place is unmasked in a thesis title because of the possibility of inductive or contextual disclosure. Whilst my approach satisfied ethical requirements, that notwithstanding, as a researcher, the consequences of sharing individual stories about micro-politics of access to education of a population who could be regarded as elite research participants should not be lost on me (Woods, 2012). Further, reporting the findings with rich description of participants such as age, work, number of children in their unique context implies that inductive and deductive identifiers could compromise confidentiality. This contextual detail, was however, imperative to avoid decoupling the study from its context specificity (Nespor, 2000. P. 549). It was also important to embed the study in history and to explore how the FGEW and their families lived and acted in their situated contexts and their negotiation of education access (Thomson et al., 2005).

I blamed myself for not protecting their privacy and identity enough because sometimes informants may not be aware of the possible consequences of the emotive nature of enquiry. Moreover, even though participants were educated, I had better knowledge about the research process, I knew about the implication of research on micro-household politics if participants are identified in a way that my participants did not.

These issues weighed heavily on me, and I wanted closure on the issue. I took time to investigate my research participants, who were still alive and the study area once again. I also began to ask questions about the lives and social organization of the community in details.

Worried Over Nothing? Situating the Rejection of Anonymity Within the Personal Context of Participants

I realized that I was worried over nothing. First, I was not dealing exactly with participants in a potential unequal relationship situation as I thought their contradictory statuses as educated but with other disadvantaged markers; their ages, their gender, and their presumed lack of knowledge about risk in providing data, portray them to be. I was dealing with participants who did not want to see their participation as minimal or a single interview. The goal was to give them voice. Giving participants voice is used in relation to how I threaded their voices in communicating the findings. Even though I synthesized the results to create a combined story of narratives to provide a further layer of anonymity in most instances, however, I also quoted extensively so that a participant can identify themselves in the written work despite the use of pseudonyms. I also engaged participants on the use of pseudonyms. We agreed that the pseudonyms employed would be names that are typical to the study context. This is consistent with Svalastog and Erickson (2010, p. 104) suggestion to engage participants concerning the choice of pseudonyms. Engaging participants on choices of pseudonyms is also about giving participants’ voice (McCormack et al., 2012).

They wanted recognition for their data, and it would be unduly paternalistic (Giordano et al., 2007; Ryen, 2004) to apply complete or total anonymity given that the data although, emotive could not be described as Footnote 5special category. Baez (2002) refers to such complete anonymization as convention of confidentiality. Convention of confidentiality is useful when dealing with participants with stigmatizing traits.

Further, participants wanted the core of the research to be about narrating their individual personal stories with their names, whilst I wanted the research project to be about understanding the micro-politics of their access to education. Telling their stories with their names in the final output, however, would not address the research question. Conversely, anonymizing all the contextual and personal identifiers will also not address the research question. Because I was dealing with a unique group of people, their personal identifiers such as their education backgrounds, and the contextual identifiers e.g., the meso-education access context, and the history of education access in the location was necessary to contextualize the findings.

The participants personal context for rejecting confidentiality was their motivation to see their names in written works. The extract from my fieldwork note in the introduction should have given me a clue. I instead got surprises because I did not investigate the motivation behind their zeal to participate and how keen they were to determine how the research should be written-up. If I had investigated the motivation behind the call to waive the right to confidentiality, I would appreciate that they were more knowledgeable to negotiate confidentiality. I thought I had a better understanding of the implication of waiving their right to confidentiality. I had assumed asymmetry of this knowledge because despite the potential rife that naming them could cause within their families, they wanted to be acknowledged for their data.

Apart from the earlier reason the research participants gave me during the research five years ago re ‘helping you,’ I realized that it was both about their personal need to get their stories told and their political need to be seen and their achievements recognized and written about.

Worried Over Nothing? Situating the Rejection of Anonymity Within the Social Context and Cultural-Specific Issues

Whilst I was confused about why educated people were bent on rejecting confidentiality, I have come to understand that the research participants in my project were making rational choices in rejecting anonymity or conventional confidentiality. Apart from their personal need to get their names in written works, the nature of the social organization of the people and the community also played a part.

The study area is a typical matrilineal society. It means that tracing one’s genealogy is through the mother line only (Nukunya, 2003). It also presupposes that relationship construction around access such as transfer of traditional offices and properties is through the mother line and may involve maternal uncles, maternal grand uncles, maternal aunties, and everyone considered part of the matriliny. The bond among kin relations is still intact unlike in the urban centers where social change is rapidly leading to different and new ways of family formation (Kpoor, 2014, 2016). These traits have been dealt with fairly in the literature around family formation and descent systems (Nukunya, 2003). When education access is embedded in such relationship constructions, access is no longer a private matter because different matrilineal and extended relations would know and determine access and the nature of such access.

Furthermore, the people are communal in how they go about their activities. People just knew about each other’s businesses and what they do is not private at all. Therefore, the concept and understandings of what privacy and confidentiality mean and how it is approached was different from contexts that are very individualistic. In the Kyebi context, people’s education access issues were, therefore, rarely private. They cherish truth and consider nothing confidential with speaking their truth about their lives. By trying to impose my institutional confidentiality ethics on them, I was overruling what they considered right for themselves and their society.

My reflections do not suggest that qualitative researchers should ditch anonymity. It furthers the debate that anonymity is a decision that should be negotiated with both institutional ethical regimes and research participants whom we seek to protect. Further whilst anonymity is good research practice, it has implications on power relations between the researcher and the researched in view of who decides or have the right to decide what aspect of data is anonymized. Understanding the motivation behind participants participation and the context of the study area is important in operationalizing anonymity.

On reflection, I am still not sure whether I did the right thing with how careful I had been about trying to ensure anonymity or confidentiality and there are still more questions than answers.

Navigating the Loops and Practical Guidelines on Administering Confidentiality

Confidentiality has conventionally been about procedural ethics where researchers are required to state they will ensure confidentiality. A good appreciation of the nuances, however, occurs eventually when operationalizing the proposed study (Miller & Bell, 2012). Especially in research employing emergent qualitative enquiries, it is difficult to tell what the participant is consenting to at the initial stages of the research. Thus, even though confidentiality has it place in how the access to and continues participation of research participants work, because of the ongoing nature of access to participants, the nature and specificity of confidentiality may also change. Whilst knowledge about the fluidity of confidentiality is helpful, a good appreciation of the motivation behind participants confidentiality choices and their willingness to participate in the study, will be helpful in navigating the dilemma that so often arises re-confidentiality.

Increasing calls on exploring nuances in confidentiality coalesce around two approaches: the dominant and alternative approaches. The dominant approach suggests external confidentiality where participants data are processed in a manner that enable them to remain anonymous to outsiders (Tolich, 2004, p. 101). The alternative approach explores internal confidentiality in ways where data could be made anonymous to other participants and their immediate relations (Tolich, 2004). Kaiser (2009) also explores an alternative approach, and he enjoins researchers to focus on two things, the audience of the data and dissemination plan (p. 7) in constructing what he termed a re-envisioned consent form. This form he recommends should include the targeted audience for the final research output. Additionally, negotiating what this form should entail must be ongoing and should present research participants with confidentiality options to choose from.

Indeed, the above debates conclude that that rigid confidentiality is inherently incompatible with emergent qualitative approaches for two reasons: first, contextual identifiers cannot always be removed, further removing deductive identifiers may not always answer the research question. Consequently, they advocate for research ethical applications to be about providing incremental guidelines along the research process in stages. Based on the knowledge from extant literature, the table below summarizes the stages along which incremental ethical guidelines on confidentiality is needed.

Table 1 Stages along which incremental ethical guidelines on confidentiality is needed
Full size table

Whilst other scholars focus on the participants and dissemination outlet as audiences in exploring nuances in confidentiality, this paper contribute to the debate by arguing that changes in confidentiality should be enabled by the variation of participants motivation for taking part in the study. Additionally, I argue for the exploration of the cultural and political context of the study site to enable an understanding of participants choice of confidentiality. For example, if the study site is an individualistic culture, the tendency to expect rigid confidentiality could be greater as compared to societies that are communal in their social organizations. Confidentiality could also be constructed to include no confidentiality clause.

Under what conditions should such rejection of anonymity or a no-confidentiality be granted? Whilst Kaiser (2009), suggest the burden of such decisions be shifted to the research participant, leaving this entirely to the participant to decide risk defeat the Belmont beneficence principle (Anabo et al., 2019). Researchers are trained to understand better the research process and can have a perspective on confidentiality having encountered the data in a way that a participant may not. For example, if a vulnerable population who could be in harm’s way if personal data is identified with research data, reject confidentiality, should their agency to determine the nature of confidentiality be respected? Should the burden of what kind of confidentiality be shifted to such a participant? should confidentiality be applied tacitly against their will?

There is no right answer to these questions. Whilst reflexivity is key if confidentiality can be operationalized as a re-iterative process, confidentiality should also be constructed around the types, the how, and the purposes. Importantly, it should be based on the participants motivation(s) for the rejection of confidentiality.

Additionally, researchers could explore the data gathered, and work with participants to determine which part of the data could cause confidentiality issue. This is not straight forward because whilst a researcher’s perspective formed from encounters with institutional ethics regimes and the ethical relativistic lens from participants, could identify data that could cause confidentiality issue, participants determination of which data should remain confidential and which data should be revealed could also be based on their motivation for taking part in the research. Thus, as earlier posited, a key aspect of exploring nuance in confidentiality should include participant’s motivation for participating in the research.

To make the process of exploring participant motivation easier, I affirm the importance of Kaiser (2009)’s new and re-envisioned consent form. Additionally, I also suggest an additional form ‘the participant motivation consent form’ be created. This would enable the researcher to put the participant’s consent preference into perspective when disseminating the research output. Motivation can be examined throughout the research process and a final form can be created for participant to sign. The figure below is an example of how a participant motivation form can be created.

Fig. 1
figure 1

Participant’s motivation consent form

Source: Author’s creation

Envisioning confidentiality this way to include the participant motivation form could have solved the confidentiality dilemma I faced in the Kyebi Ghana situation when I researched the first generation educated women

Full size image

Navigating the Loops: Practical Guidelines on Dealing with the Ethical Quandary Re-Confidentiality

The paper was based on reflections of my fieldwork experience in Ghana and the ethical quandary I faced in operationalizing confidentiality. The dilemma is, what ethical position can one take in scenarios when the researcher and the researched have conflicting perspectives about confidentiality? Should participants voices be drowned in confidentiality conversations around the processing of data they have voluntarily provided?

As illuminated in the paper, participant motivation and the cultural context of participant may influence their confidentiality choices. Participants choices may differ from the ethical standards of the researcher’s institution. Whilst maintaining institutional allegiance in this matter is crucial, researchers must negotiate with the participants against the backdrop of their ethical clearance, the participants choices, their motivation, and the socio-cultural context of the research. When this is done, the consensus that is reached would ensure trust and enables continues participation of participants who may want to terminate their involvement because they want recognition for their data (Few et al., 2003; Karnielli-Miller, 2009; Koch & Harrington 1998; Maxey, 1999). Additionally, allowing participants to negotiate their own terms of participation such as autonomy to determine anonymity or confidentiality is important for correcting the power imbalances between the researcher and the researched (Fisher & Ragsdale, 2005; Hanley 2005; Henry, 2003; Kelman 1972; Koch & Harrington, 1998).

Ethical issues may, however, arise since they may not be familiar with institutional ethical practices structuring the project, therefore, such encounters should be managed to protect reputations, rights, and the integrity of the research. Managing such encounters mean assessing how community values and varying cultural understandings of privacy reconcile with institutional ethics. It also implies assessing whether the society we are researching is a communal or an individualistic society and how that influence the choice of the researched to opt in or out of confidentiality. Managing encounters suggest incorporating participants confidentiality preference and enabling participants with personal stake to share in the construction of the study aside the interview data they provide.

Judging by my own experience, negotiation of anonymity, and confidentiality should be ongoing, and renegotiated throughout the research process to enable a mutual and a balance process of constructing knowledge. There should be an opportunity for researchers to submit ethical concerns that require addressing after an encounter with participants data. This is important to clarify whether based on evidence, changes could be done to the project after the initial clearance pre-operationalization of the research. For example, insurmountable ethical implications could be reviewed by the ethics board after the researcher has encountered participants data, their cultural and non-cultural factors such as kinship ties and community structures which have impact on participants view of privacy, anonymity, and confidentiality. I recommend that the researcher should maintain ownership of the research process whilst enabling participants more freedom to determine and co-produce knowledge.

The question remains, what happens to research when participants refuse the confidentiality needed for their protection? The answer to this is not straight forward. I propose that more reflection papers on how researchers have navigated the loops should be encouraged to enable us to answer this question. We need to bring these reflections on fieldwork experiences to the forefront of research and policy discourse. It will also help us to understand whether researchers have left any ethical footprint on their research location and participants.