The ethics of ethical debates in peace and conflict research: Notes towards the development of a research covenant

Abstract

This article outlines the case for peace and conflict researchers to formulate a research covenant to better shape their ethical obligations and responsibilities. This is an urgent necessity given that ethical debates have in some proponents become emotive and are not conducted in an ethical manner. In coming to this assessment, the article reviews trends in the research ethics literature and draws out some of the generic issues addressed in a review of the personal reflexivity that an assortment of individual peace and conflict researchers have engaged in when recounting their fieldwork experiences. These generic issues are reformulated in an attempt to codify appropriate ethical practice in peace and conflict research, and they go towards determining the contents of the research covenant. It is suggested that the research covenant is a more ethical way to debate the ethics of peace and conflict research.

Introduction

There are elements of continuity and change in contemporary debates about ethics in social science. These debates have always been emotive but one change has been the added emotional dimension given by some proponents who link ethical practice to issues of academic freedom. In the past, we may have debated with vigour the ethics of covert research, for example, but now we throw around abrasive accusations that allege ethical malpractices which threaten the very nature of higher education research and impugn whole organisations rather than individual researchers (see, for example, the accusations against the British Sociological Association (BSA), among others, in Lowman and Palys (2014)). This raises the interesting question of the ethics of ethical debates in social research.

Three new developments have transformed the ethics of ethical debates. The first is the strict tightening of ethical regulatory frameworks in social science, with the application of scientific and medical ethical codes to humanities and social science research as part of the neo-liberal marketisation of higher education. The resistance towards these processes has impacted on the ethics of ethical debates as some proponents urge a cut loose, hang free and disregard the law approach under the impulse of protecting academic freedom in ethical practice against ever-tougher restrictions in these regulatory frameworks. The second development is the emergence of the ‘new ethics’ (a term used by Hammersley and Traianou, 2014), where conventional ethical practice is supplemented by ethical practices designed to promote social justice and human rights. If what matters is the normative purpose of the research rather than its scientific standards, a ‘new ethic’ impacts on researchers’ ethical research practice, making research ethics subservient to wider moral precepts.

The third change has been the development of the new field of peace and conflict research. This kind of research throws the ‘new ethics’ into particularly high relief because of the normative commitment most of its researchers have towards making a difference to the lives of people affected by conflict, but the higher risks and dangers associated with research in conflict zones or in societies emerging precariously out of conflict also significantly tests the relevance and applicability of conventional research ethics codes. Adaptability in ‘dangerous fields’, as Kovats-Bernat (2002) refers to these research sites, is necessary in order to manage its risks, making flexibility more important than conformity to ethical codes.

Many of the most emotive debates about the ethics of ethical practice concern the experiences of peace and conflict researchers, and is reflected well in the Alice Goffman Affair, which involved academic reaction to Goffman’s (2014) research on police relations with the Black community in Philadelphia, and the Boston College Affair (see Palys and Lowman, 2012), involving the debate around the collapse of Boston College’s oral history archive on the conflict in Northern Ireland. These two cases teach salutary lessons about the ethical challenges and moral dilemmas posed by peace and conflict research.

The purpose of this article is fourfold. First, it will elaborate on some of the above methodological trends as they affect the peace and conflict research. This field is very recent and some researchers have developed reflexive accounts of their personal experiences when undertaking research in conflict zones or during peace processes. The article therefore aims, second, to give a flavour of the content of some of these personal reflections as they relate to ethical practice. Since this field is so new and will be unfamiliar to most readers, the ethical challenges and moral dilemmas the field poses researchers justify such a highlight, especially given that much of it is conducted by PhD and early career researchers who are given little training in this area and are open to naivety as a result. Third, the article attempts to draw from these diverse personal reflections a codification of the ethical, legal and moral challenges in peace and conflict research that systematises its opportunities and threats. Finally, the article aims to reduce some of the emotion involved in the ethics of ethical debates within peace and conflict research by developing the idea of a ‘research covenant’ for peace and conflict research that might be useful to new researchers entering the field. We begin, however, with some reflections on developments in the methodological debate about research ethics.

Debating ethics

Despite the strong elements of continuity in contemporary debates about ethics, as familiar issues remain highly relevant, such as the boundary between participation and observation, the independence of researchers from sponsors and the appropriateness of covert observation, ethical debates in the research methods literature have undergone a transition over recent decades, reflecting broader developments in global social science. Debates have shifted from narrow concerns with ethics, towards an interest in sensitivity and risk, and within risk, there has been a shift from the risks faced by respondents to those faced by researchers. While honesty, integrity, trust and confidentiality are enduring issues despite these changes, these conventional concerns have taken on more significant importance as a result of these trends.

Ethical debates in the research methods literature were once primarily about the ethical implications of some research practices (mostly the propriety of covert research) and of some data collection techniques (mostly participant observation), and debates about honesty, integrity, trust and confidentiality were framed within this limited focus (see, for example, the debate between Bulmer (1982) and Homan (1980, 1991)). The emergence of sensitivity as a methodological issue in the late 1980s (see Lee, 1994; Renzetti and Lee, 1993), and within that the identification of what was called dangerous research (Lee, 1995; Sluka, 1995), gave new meaning to ethical debates and broadened the range of issues and problems provoked by sensitive research. Brewer (1993), for example, noted how the sensitivity of the topic and its location not only raised new ethical challenges, it complicated the problems routinely experienced in all kinds of research (also see Brewer, 2000: 83ff, 2012b: 69–73).

The development of the idea of ‘risk society’ in the late 1990s, which became paradigmatic within global social science, articulated the notion that late modernity was exposed to new forms of structural risk and that citizens had been sensitised, indeed over-sensitised, to risks in their everyday life. The risk society was applied within the research methods field to make us aware of the risks respondents faced by some kinds of research, which was then broadened to cover risks also faced by researchers. To reinforce the importance of the latter emphasis, a code of conduct was developed in 2001 by the Social Research Association for the safety of researchers, and Qualiti, a node of the Economic and Social Research Council’s (ESRC) National Centre for Research Methods based at Cardiff University, published a report in 2007 on the emotional and physical harm suffered by qualitative researchers (see Bloor et al., 2007, 2010). Lee-Treweek and Linkogle’s (2000) edited collection on danger, risk and ethics was one of the most significant transition points when the research methods literature made the shift from locating discussions of ethics within the ideas of sensitivity and danger towards locating ethics within the idea of risk.

The imposition of science and medicine-based ethical regulatory frameworks in social science has made ethical clearance an obligation, and within this, priority tends to be given to risk. Risk assessments are de rigueur for ethical clearance even for undergraduate dissertation projects nowadays. The primary focus within ethical debates is thus on identifying potential risks, minimising harm to respondents and researchers alike, respecting the autonomy of respondents, tough ethical clearance monitoring for researchers by ethical review committees and guarantees of preserving the privacy of respondents and protecting confidentiality. Hammersley and Traianou (2014) refer to this as the dominant or standard form of research ethics discourse, with risk being the hub around which the conventional ethical issues of harm, confidentiality, informed consent, anonymity, trust and integrity are organised. This highlights the strong elements of continuity and change within contemporary ethics debates.

They refer, however, to the emergence of an alternative normative ethical discourse, which they call the new ethics, where the purpose of the research ought to be about realising social justice, the furtherance of human rights and an obligation to promote social care. This discourse represents a more profound change. It is represented best by world-leading specialists in qualitative methodology like Lincoln (1995) and Denzin (see Denzin and Giardina (2007b), as well as in research methods textbook by Mertens and Ginsberg (2009), Denzin and Giardina (2007a) and Denzin and Lincoln (2011), among others. (For a summary of this debate in a rival textbook, see Hammersley and Traianou (2012).) The ‘new ethics’, however, can also be located in broader changes within social science about public value, public sociology and the normative dimensions of social science that envision social science with a normative purpose to eliminate social suffering and promote the social good (on which see Brewer, 2013). Yet, the new ethics is not a complete break with the past for it also has roots in long-standing methodological practices like participatory research, action research and feminist methodology, all of which involve empowering respondents and re-ordering the research relationship.

Hammersley and Traianou (2014) dismiss the new ethics because it is in contradiction with the central commitment of social researchers to produce ‘sound knowledge’ (p. 10). This is a harsh judgement for science and ethical responsibility are not mutually exclusive (see Brewer, 2014). However, the tension between the standard ethical discourse and the new ethics is thrown into particularly high relief by considering ethical issues involved in research on conflict, violence and peace. The principle of risk avoidance and ‘do no harm’ must be paramount when working in dangerous fields, but this sort of research also involves trying to do some good, of helping to transform the conflict, promote the peace and improve the lives of people in the midst of conflict or emerging from it.

Peace and conflict research is a field particularly useful for showing how some researchers have tried to balance their commitments to avoiding harm with doing good at the same time. While it helps disprove Hammersley and Traianou’s hasty assessment that the new ethics always produces unsound knowledge, this kind of research also poses very great ethical challenges that are not anticipated in the standard view of research ethics and involve some significant departures from it and thus from the textbook portrayal of ideal ethical practice. In trying to keep to conventional ethics codes of practice, peace and conflict researchers are often faced with moral and legal dilemmas that the codes are inadequate to meet. Some peace and conflict researchers, thus, have deemed standard ethical practices largely irrelevant and inapplicable in dangerous fields and have suggested adopting what they call a ‘localized ethic’ (see Kovats-Bernat, 2002). This is indicative of how peace and conflict researchers have to be adaptable in their approach to ethical issues when in the field, leading to both marked variations between them in their ethical practice and to significant changes from the ethical practices of earlier generations.

The differing responses made by peace and conflict researchers to these challenges and dilemmas are worth highlighting for three reasons. First, differences between researchers in how they responded in situ to them have led to a particularly vituperative, emotive and aggressive debate among them about the ethical limits of conflict and peace research, especially with respect to the question of confidentiality and whether it can be guaranteed within or outside the law, but also around the nature of informed consent, and the problem of over-identification with respondents in the field. Second, identifying these dilemmas and challenges has a capacity-building function by drawing them to the attention of new and unfamiliar researchers so that they can be better prepared when they enter dangerous fields. This is particularly useful, for as Bloor et al. (2010) note, PhD students, junior researchers and early career researchers are often those most at risk of harm, being let down by principal investigators and supervisors who are themselves unaware of the risks, dilemmas and challenges of dangerous fields or who are not managing these challenges effectively enough. Third, because these challenges and dilemmas have not yet been codified but are discussed in isolated, hard-to-access and disparate personal reflections on research practice by individual peace and conflict researchers, there is some value in drawing the discussion together in this article. The best way to do this is to begin by reviewing the personal reflections of some of the leading peace and conflict researchers in order to prepare for our later discussion on a new research covenant that systematises and codifies the ethical responsibilities of this kind of research.

Ethics in peace and conflict research

In a recent reflection on his research experiences as a PhD student doing fieldwork in Palestine, Browne (2013) encouraged peace and conflict researchers to be more reflexive, urging on them honesty and openness about its difficulties. He recommended the use of research diaries as ways for ‘novice researchers’ to ‘record the personal’ (Browne, 2013: 423). In fact, there has been a great deal of reflexivity, but it is disparate. In what follows, I review some of the reflexive accounts of fieldwork experiences by an assortment of peace and conflict researchers. Not all such researchers have been reflexive on their methodology and approach, so the review is highly selective, and while there are a few edited collections that some time ago addressed social research in divided and violent societies (e.g. Nordstrom and Robben, 1995; Smyth and Robinson, 2001), my review concentrates on ethical issues. It raises points about ethics which are worthy of broader debate which are at the moment concealed within more disparate and dispersed personal reflections across a diverse set of journals and books. While this will be done on an individual basis, some of the generic, recurring issues in peace and conflict research that focus the broader debate stare out. These are the need to plan for adaptability, the problems around the nature of informed consent, responsibilities towards confidentiality and what it means, the problem of managing risk, the need for data management and security, the problem of ‘guilty knowledge’ and how to manage it, complications around dissemination and publication, and the limits of ethical codes. In this latter respect, the discussion that follows demonstrates that ethical practice by peace and conflict researchers when in the field can contradict professional ethics codes, and conformity to these codes can have the opposite effect to that intended by exposing respondents and researchers to greater threat and risk.

We begin with the personal reflections of the sociologist Gladys Ganiel (2013) based on her work on religion, conflict and peace in Zimbabwe. This is a useful starting place because they are contained within an edited collection designed to address ethics for graduate students, and Ganiel couches her reflections firmly within the normative commitments of the new ethics. She wrote of the moral dimension to her research, feeling it must make a contribution to transforming the violence (Ganiel, 2013: 168). She argued that this kind of research therefore challenged the idea of a notionally value-free tradition of positivist social science, since researchers needed to be fully embedded ethically in the worlds they are researching if they are to make a difference (which she recognised was also a strong feature of action research and feminist methodology).

Despite this tension with the value-free understanding of ethics, Ganiel (2013) recognised that peace and conflict research offers what she calls an ‘ethical opportunity’ (p. 167), the successful negotiation of which needs to be planned for at the outset. Researchers must try to anticipate known problems and plan for how they will manage the unexpected contingencies that arise in dangerous fields so that they can produce reliable (or ‘sound’) knowledge. This involves planning for their own personal safety and for the personal safety of respondents, and consideration given to how they will communicate and disseminate their results without exposing their respondents – and perhaps themselves – to risk. She recounted her experiences in Zimbabwe with respect to confidentiality and informed consent and argued that the risks required a departure from standard ethical practice in that she opted for oral consent not written in order for there to be no official written record of people’s participation in the research (Ganiel, 2013: 173). To further manage the risks of respondents’ exposure, tapes were erased immediately they were transcribed, and transcripts were erased from her laptop in the field once sent out of country electronically. This departure from standard ethical practice was done to protect her respondents from harm, something obliged under conventional ethical codes but which was threatened by those very codes if she had stuck to them.

Ganiel felt sufficiently under the constraint of conventional ethical practices, however, to reflect on what informed consent means if it was only orally given. She saw it as not deceiving people about the research or their participation (Ganiel, 2013: 174), which meant not making guarantees, promises and reassurances that knowingly could not be kept. She particularly warned of reckless self-serving publication and dissemination that places respondents at risk (Ganiel, 2013: 176), and of the need to discuss publication risks with respondents when negotiating informed consent.

The anthropologist Elisabeth Wood’s (2006) work in El Salvador focused on local indigenous community involvement in insurgency and was conducted while the civil war was ongoing. She described that the main ethical challenge was to adhere to the principle of ‘do no harm’ (Wood, 2006: 373), particularly in protecting the anonymity and confidentiality of respondents ‘to the extent possible’ (p. 373). She argued that where it was thought at the planning stage that the research could not be ethically conducted (such as where the principle of doing no harm could not be kept or anonymity and confidentiality could not be guaranteed), the research should be curtailed or not attempted (Wood, 2006: 374). In her case, she reported how she hung back on publication of results in order not to expose respondents to risk (Wood, 2006: 379).

In conflict zones, the risks to respondents come from disclosure of their involvement to security forces, and to researchers from being vetted by governments, paramilitaries and by local leaders (Wood, 2006: 375). These vetting risks, of course, are faced also by translators (Wood, 2006: 379) and gatekeepers. There is thus an ethical obligation on peace and conflict researchers to consider the impact of the research on a diverse range of people nominally connected to it, which includes, in addition to the above, the effect of secondary or vicarious traumatisation on transcribers and supervisors who read victim narratives (emphasised by McCosker et al. (2001) in their work on ‘domestic’ violence). The risk of re-traumatisation of respondents as they renew the emotional experiences they are accounting in their narratives is an ethical dilemma for all researchers studying emotionally sensitive material not just peace and conflict researchers, but in the latter field, secondary or vicarious traumatisation of the people only nominally involved in the research is a real risk (see Browne and Moffett, 2014: 227).

Wood’s reflections confirm the special ethical challenge posed by informed consent in peace and conflict research. She argued that respondents should be made fully aware of all potential known risks and benefits of their participation, so that they make an informed decision (Wood, 2006: 379). They should also be informed of the risk of the unanticipated and the unexpected. Wood again departed from the ideal textbook portrayal of ethical practice in the standard view of ethics by using oral not written consent (Wood, 2006: 380), in part to get over illiteracy in her respondents but primarily to ensure there was no written record linking respondents to the research that could expose them to risk. Ironically, the ‘do no harm’ principle so central to conventional ethical practice obligated breaching the codes that enforce it because to do otherwise would bring the very harm the codes were designed to avoid.

In the informed consent stage, Wood recounts how she managed the problem of ‘guilty knowledge’ (a term widely in use to refer to disclosure in interview of respondents’ knowledge of or participation in illegal activity; see Fetterman, 1983). It was dealt with by explaining beforehand what respondents could and could not talk about, whether what they said was or was not for publication, and by emphasising that they had the option to decline to answer questions. In this way, the process of informed consent was used to allow respondents to exercise their own control over the content of the interview (Wood, 2006: 381). She did not address the question of the law and what she would have done if data had been requested by state forces, so the promises she made about confidentiality were not put to the test. This is a point we come back to below.

Wood (2006) emphasises the importance of ensuring the security of the data (p. 381) when doing peace and conflict research. Names were not recorded, she did not tape interviews with some respondents, precautions were taken at checkpoints and when crossing borders to conceal data files and transcripts, data were sent back to the United States electronically very quickly (Wood, 2006: 381), and she linked herself to religious figures, using their status and legitimacy in a religious culture to provide some protection and as a cover for her data (p. 382). To further avoid harm, some material was excluded from immediate publication, some of the most sensitive was delayed and not published until a decade later, and some not at all, despite the respondents’ permission to use it (Wood, 2006: 382).

Interestingly, Wood also conducted her research under what we have called here the normative commitment of the new ethics, writing of her sense of ethical responsibility to return something to the community. She met this ethical responsibility through establishing anonymised archives and holding community workshops (Wood, 2006: 383) – there are many other ways of communicating findings back to respondents discussed in Ganiel (2013: 177) – enabling Wood to feel that the research itself became a contribution as a form of intervention mitigating in a small way the harm experienced by respondents as a result of the conflict. Like humanitarian aid, she writes, research can itself be a benefit in dangerous fields. The commitment within the new ethics to both ethical responsibility and science was thus satisfied in Wood’s case by developing reciprocal relationships with respondents which returned something back to them. Nurturing reciprocity in relationships with respondents is itself a way of dealing with the ethical challenges posed by peace and conflict research.

Christopher Kovats-Bernat (2002) is an anthropologist with extensive experience of working in Haiti, and he defines ‘dangerous fields’ as fieldwork sites where data are affected by fear, the threat of force and violence (p. 208). As unstable field sites, they are places where customary approaches to ethics are insufficient and sometimes inapplicable, and do not prepare researchers for the moral dilemmas they face (Kovats-Bernat, 2002: 213–214). In what by now should be a familiar theme, he argues dangerous fields therefore involve a departure from the textbook portrayal of the ideal methodological practice, forcing researchers to innovate. If standard ethical codes cannot meet the challenges faced, the response should not be to operate unethically, regardless of any responsibility and obligation towards respondents, but to use the very local knowledge of people in the field to determine what is ethically permissible and possible. Kovats-Bernat (2002) refers to this as a ‘localized ethic’ (p. 214), which he describes as taking stock of the good advice and recommendations of the local population in deciding what is permissible and doable in the light of the dangers that they are better informed about than the researcher.

This localised ethic may limit practices that ethical codes permit, or require practices that ethical codes do not favour – he also cites the use of oral rather than written consent (Kovats-Bernat, 2002: 215). This localised ethic must also make researchers sensitive to their inability to guarantee safety and their powerlessness to anticipate all the threats faced by respondents (Kovats-Bernat, 2002: 215). He argues that two consequences follow from this. First, peace and conflict researchers need to be cautious in what is promised and what they allege can be guaranteed; and second, they need to think carefully about the balance between what the information is worth against the dangers to respondents and researchers involved in collecting and reporting it (Kovats-Bernat, 2002: 211). In the light of this, he argues that a localised ethic ensures respondents’ own local knowledge dictates ethical practice and defines the ethical limits and parameters of the research.

The problem with this advice, however, is that in peace and conflict research, there will be multiple competing voices, resulting in local knowledge being fractured and contradictory, spread across diverse constituencies, like victims, perpetrators, non-state military combatants, state combatants, the grassroots, civil society, politicians and the like, the strong differences between whom defined the very nature of the conflict and structured the violence. Local ethical frameworks will be manifold and choices may need to be made about whose voices matter in the research. Jonathan Goodhand (2000) makes this point when reflecting on his experiences of researching in Afghanistan, Sri Lanka and Liberia. He identifies three main challenges – practical, methodological and ethical – and focuses on the latter. He argues that conflict heightens and amplifies the ethical challenges faced by researchers, requiring them at the planning stage to develop sufficient understanding of the ethical issues raised by their research (Goodhand, 2000: 15). Consistent with the new ethics, researchers need to develop an ethical framework for decision-making that avoids harm and opens up the possibility of doing some good (Goodhand, 2000: 12). He specifically recommends thinking about whose voices are to be heard and whose knowledge is to count in the research (Goodhand, 2000: 12).

Researchers can avoid doing harm, he writes, by not infringing the security, well-being and safety of respondents, which means that ethically informed decision-making involves peace and conflict researchers thinking about the impact of the research on respondents in war zones (Goodhand, 2000: 13). This requires them to think about how they conduct the research, who they talk to and what to talk about, all in order not to expose individuals and communities to risk (Goodhand, 2000: 13). More than other researchers, peace and conflict researchers must be sensitive to the needs and fears of conflict-affected communities. This involves giving consideration to what topics respondents will feel sensitive about, what will endanger them and what will be the lasting effect of the research long after the researcher has left the field. Their local knowledge will indicate this, but differences between constituencies will almost certainly result in significant variations in what they find sensitive, dangerous and risky. Two moral dilemmas therefore ensue that we can pose as questions. Whom among the voiceless does the peace the conflict researcher give voice to? How does the peace and conflict researcher bear witness to abuse without endangering the abused (or themselves)? Goodhand therefore reinforces Kovats-Bernat’s (2002) point that ethical decision-making in the field of peace and conflict research involves asking whether the research warrants the risks (p. 13). Goodhand (2000) argues further that in not preparing peace and conflict researchers to resolve these issues (or supervisors and principal investigators perhaps persisting with the research regardless that the risks do not make it worth doing), it is unethical to involve researchers who are inexperienced and unfamiliar with working in areas of conflict (p. 13).

One of the features of ethical practice that Goodhand (2000) alerts us to most strongly is the importance of not encouraging unrealistic expectations in respondents – unrealistic expectations about the benefits and impact of the research (p. 14), about the limits of confidentiality and anonymity, and the security of the data. Goodhand does not mention the problem of legal limits on confidentiality. For him, the key principle is to do no harm. The mechanisms he identifies to achieve this should by now be repetitious: by showing sensitivity to ethical issues, predicting likely ethical challenges, reflexivity on the likely effects of the research on people and communities, obtaining informed consent, honesty in outlining the objectives of the research and building reciprocity in relationships with respondents through dialogue with the community and building links with local partners.

In this context, the advice of Brendan Browne and Luke Moffett (2014; also see Browne, 2013), transitional justice researchers who have worked in Uganda and Palestine, is salutary.

Despite all the planning, they urge peace and conflict researchers to prepare for unforeseen contingencies by being flexible and adaptable in the field. They describe how they managed the risks as young PhD researchers by working with non-governmental organisations (NGOs) established in the field (also see Wood, 2006: 383–384), and had to be willing to accept the constraints that come with this, as well as becoming reconciled to an unusually high reliance on gatekeepers (Browne and Moffett, 2014: 228), which brought with it the attendant problem of finding the right gatekeepers (Browne and Moffett, 2014: 228). They describe the great lengths they both went to in protecting the security of their data and their sense of ethical responsibility towards the researched communities. In this last respect, they emphasised the importance of reciprocity as an ethical obligation on researchers, which they refer to as knowledge exchange back to respondents (Browne and Moffett, 2014: 229). They do not consider, however, a special form of harm in peace and conflict research around the problem of guilty knowledge.

The criminologist Dermot Feenan (2002) addressed this issue directly. While the term ‘guilty knowledge’ might be thought inappropriate because of its negative connotations, it is widely in use in the research methods literature and it does not involve a moral judgement of the respondents who disclose it. Some forms of research are at very high risk of revealing guilty knowledge and thus of research data being sought by police, national security forces and the courts. The controversial nature of its data makes peace and conflict research notable among these. There have been several cases of enforced disclosure in the United States (see Feenan, 2002: 762), but at the time at which Feenan wrote, it was journalists who had been affected in the United Kingdom. Lowman and Palys (2014) consider a few subsequent cases in the United Kingdom where social researchers were subject to pressure to disclose the guilty knowledge that had confidentially come their way. Some social researchers believe they have responsibility as law-abiding citizens to report knowledge of illegal behaviour (see, for example, Yablonsky, 1968), others say they would disclose only if there was an intention to harm another person, while others keep to the practice of never revealing guilty knowledge confidentially obtained (see, for example, Lowman and Palys, 2014; Polsky, 1967). Some well-known qualitative researchers have refused to comply with court requests – such as Van Maanen (1983) – although the law suit against him was withdrawn so the issue of disclosure under duress became irrelevant.

Most codes of ethical conduct do not address the problem of court-ordered disclosures (Feenan, 2002: 766). The British Psychological Association, however, recommends that researchers warn respondents at the informed consent stage of the risk that the information may be subject to subpoena, so that respondents can take this into account when deciding to participate (see Feenan, 2002: 767). The ESRC in 2000 warned that while there is a duty to confidentiality, there are exceptions when disclosure is required in relation to criminal investigations or under court order (Feenan, 2002: 767). As the law stands at the moment, research data are not entitled to special protection from court disclosure (Feenan, 2002: 776). We might argue that it should, but the hard fact remains it is not.

Feenan (2002: 772–773) explains the legal position that even where a researcher states that information will be kept confidential, any information that is confided as a result is still subject to the law on confidentiality. Confidentiality legislation works two ways. Researchers have an obligation to keep confidential any information disclosed by a respondent unless given the express consent by them to publicise it (hence the importance of informed consent and the requirement to discuss potential publication, and within that, whether anonymity is required, and if so, how). Researchers, however, are also legally obliged to disclose information regardless of promises of confidentiality when expressly ordered to do so by the courts (hence the problem of guilty knowledge). This means that researchers need to be cautious in their promises and guarantees of confidentiality when courts can force disclosure. This raises the moral dilemma – which is also a profound legal problem – of whether or not to disregard the law.

Of course, the best way to deal with this dilemma is to avoid having to confront it in the first place. It is part of a researcher’s ethical practice to plan for how they will manage the likelihood of coming across guilty knowledge; it is ethically irresponsible to deny it will be an issue and thus to fail to discuss it. Feenan suggests three ways of managing the dilemma. First, being cautious and open about the legal limits of confidentiality when in discussions with respondents at the informed consent stage, for researchers have an ethical responsibility to disclose the possibility that data could be subpoenaed (Feenan, 2002: 775). Second, researchers should plan ahead for how to deal with guilty knowledge by trying to avoid its disclosure in the first place. This requires careful negotiations with the respondent about the limits of the topics to be addressed. The experiences of Wood (2006: 381) in managing guilty knowledge have already been referred to as an example. Third, the calculation should be made whether the information is worth the risks of its exposure. Research where details of offences are sought simply should not be undertaken (Feenan, 2002: 775–776), or its topic should be redesigned so that this is not its primary aim. This confirms the advice of Goodhand (2000: 13) above that some forms of research are not worth the risk of collecting or reporting it.

But the problem remains of what to do if there is potential for the data to be subpoenaed by the court. Peace and conflict researchers have handled this risk in different ways, and the varying choices they made have generated a heated debate. Two cases are worth exploring in detail for the lessons they bear. These lessons are different in each case. The so-called Alice Goffman Affair usefully illustrates how the time-old problems around the boundary between participation and observation within ethnography are heightened for ethnographic researchers studying violence, conflict and peace, while the so-called Boston College Affair highlights how traditional concerns around confidentiality and anonymity can become dangerously problematic when they involve peace and conflict research.

The Alice Goffman Affair revolves around her PhD research on police-community conflict in a Black neighbourhood in Philadelphia which was leading to an over-representation of young Black males ending up in prison (later published as Goffman (2014); for a similar case, see Venkatesh (2008), which reflects on his ethnographic research on drug gangs in Chicago). According to Parry (2015), the Goffman case shows the ethical boundaries of ethnography. The problem surrounds her disclosure in an appendix that she acted voluntarily as driver for an individual on the lookout for the killer of one of her key respondents and declaring that on this night, when willingly driving the car, she had wanted the killer to die. This is unlawful and represents conspiracy to commit a crime. Lubet (2015) argues that Goffman participated in a serious felony by conspiracy to commit murder.

The ethical challenge for her was whether she had been too honest or not honest enough. In one sense, she was too honest in admitting to feeling this emotion and to acting it out in this manner. Lubet (2015) warns ethnographers about expressing ‘bone-deep emotions’, since acting on them is an entirely different matter, from which he infers that impulse control is an indispensable tool for the ethical ethnographer. By not demonstrating this control, Lubet claims Goffman endangered other people’s lives (both the assailant and the intended victim), never mind her own, which was an ethical breach of the worst kind, since the first principle is always ‘do no harm’. He thus cautions young and old researchers alike, ‘if you’re asked to murder someone, just say no’.

In another sense, however, she had not been honest enough, for when debate about her revelation became widespread and nasty, she subsequently claimed she had embellished the story and knew there was no risk of a felony being committed that night. The lesson here is that an ethically responsible researcher needs to resist the temptation to sensationalism, aggrandisement and self-serving forms of publication. A fast buck and easy royalties are not worth the lack of ethical integrity cheap publicity displays. Goffman reported that she subsequently destroyed the data by shredding her notes and disposing of the computer hard drive under fear of being subpoenaed (Lubet, 2015: 6), something she could not have done if there had already been a subpoena without her committing the crime of contempt of court. This is precisely the problem faced by the researchers in the Boston College Affair.

The Boston College Affair involves the development of an oral history archive lodged at the College on Northern Ireland’s conflict. Many leading figures were interviewed, including some of the main ex-combatants in the paramilitary organisations. The archive collapsed when the Police Force of Northern Ireland (PSNI) made successful requests in the US courts for access to the data. The researchers have called for the destruction of the data, but it is now under court jurisdiction. They have also urged that social research data be exempted – in their case retrospectively – from enforced court disclosure. The PSNI used some of the information to make arrests (although there have been no prosecutions yet). In Palys and Lowman’s (2012) detailed chronology of the affair, they state that the Boston researchers gave promises of confidentiality without limit or reservation and without drawing to respondents’ attention that there might be a threat of subpoena (Palys and Lowman, 2012: 272). One of things lost sight of in the affair is that researchers gave promises that information would not be disclosed in respondents’ life times but went ahead and published a highly contentious book that made controversial allegations about current political figures that served to draw attention to the data (Maloney, 2010). The researchers have blamed Boston College for not upholding their promises and guarantees. They have not to my knowledge reflected on whether it was ethically irresponsible to make those guarantees knowing that they could not be kept within the law and then to themselves abrogate the promises by publishing within the life time of some of the people referred to in the book.

While the researchers have attended only to the ethical responsibilities of Boston College, in a follow-up article, Lowman and Palys (2014) strongly defended the ethical propriety of the researchers and argued that it was British sociology that had betrayed the principle of research confidentiality, and that the BSA and myself as its President, among others, were colluding in an attack on academic freedom. What brought on this accusation was that I had written a letter (Brewer, 2012a) to the Times Higher in a personal capacity from my work address, criticising the view that the Boston College researchers were innocent victims caught up in the affair by pointing out some of the ways that they had brought it upon themselves. To my knowledge, the BSA has still not issued any formal statement on the Boston College Affair, but the personal ethical stance I took based on my long experience as a peace and conflict researcher was expanded into a case against the BSA.

The ethical stance I took then – and still take – is that guarantees of confidentiality and anonymity are vital under the moral and legal principle of doing no harm but that, like all aspects of research practice, they are subject to the law. Lowman and Palys (2014) referred to this stance as the ‘law of the land’ approach, which they contrasted pejoratively with their own stance that emphasised the importance of protecting respondents’ anonymity and confidentiality regardless of the law and at all costs, freeing ethics and social research generally from the constraints of the law. The ‘law of the land’ approach was said to subjugate ethics to law. Naturally, I see ethical responsibility working quite differently. Ethical practice requires respect for the law. I cannot see how it is compatible with ethics, even the ‘new ethics’, for researchers to knowingly act criminally or outside the law. The Alice Goffman Affair shows us that.

The BSA and I were accused nonetheless of attacking the first principle of ethical practice because allegedly we offered no protection from harm for respondents. In situ, on-the-hoof decisions lead naturally to variations in ethical practice; this is to be expected. However, reasonable and legitimate questions can also be asked of Lowman and Palys’ alternative ethical position. How ethical is it to offer promises of confidentiality and anonymity that knowingly cannot be kept? Harm is alleged to be risked to respondents when researchers act within the law, but what of the harm to respondents when researchers do not discuss the risk of subpoena and then are not able to keep their promises under court rulings? Lowman and Palys (2012) refer to the law of the land approach as ‘truly disturbing’ (p. 114), but is it not equally disturbing to promote the illusion that confidentiality and anonymity can be protected at all costs? In counter it could be argued that harm is caused when researchers embellish and make promises that knowingly cannot be kept, and that researchers display ethical irresponsibility when the risks of subpoena are not discussed at the informed consent stage. Lowman and Palys assume that confidentiality guarantees should have no limit, but what about harm to researchers when they are subpoenaed or even jailed under contempt of court? The no harm principle that we equally endorse is Janus-faced, for harm is surely done by the suggestion that there are no limits to the guarantee and that researchers are above the law. The law of the land approach is accused of being unethical, but is it ethical to suggest researchers are above the law?

The implications of the Boston College Affair are threefold. Peace and conflict researchers need to be cautious and judicious in what they promise; they need to be open and honest in how they outline the risks faced by respondents; and they need to be careful in how they publish results and the way they write in order to protect against harm. In such an assessment, I have the support of researchers who are experienced in undertaking peace and conflict research in dangerous fields and whose work has been reviewed above.

What I have tried to argue in this review of researchers’ fieldwork experiences is that peace and conflict research poses severe dilemmas and challenges that researchers have had to respond to on the hoof, forcing researchers to make honest calls that others may well disagree with or which differ, sometimes fundamentally, from the textbook portrayal of ideal ethical practice and from tough regulatory frameworks that are often ill designed for the peculiar challenges of dangerous fields. The choices they made in good faith deserve more empathy, for peace and conflict researchers have had to respond without much formal training or any codification of the ethical practices essential for this kind of research.

It is with this idea of codification in mind that we turn in the next section to the notion of a research covenant to better clarify and shape the ethical obligations and responsibilities of peace and conflict researchers. Such a covenant should accurately reflect the diverse range of ethical challenges and dilemmas of peace and conflict research and point towards improved ethical practices that can guide research in dangerous fields.

A research covenant for ethical practice in peace and conflict research

These thoughts on a research covenant are not fully developed or finalised and are best understood as a prolegomenon or notes towards a formulation. Let me first clear away potential misunderstandings. Some interpretations of the word covenant see it as a legal contract. A legally binding written agreement between researchers and respondents is not the way we should go in improving ethical practice in peace and conflict research. I have in mind more the religious connotations of the term covenant that extend back to Abraham in the Old Testament that portray it as a statement of understanding, expressing obligations and responsibilities that form a moral rather than a legal commitment. Furthermore, my role in articulating this research covenant is, continuing the religious metaphor, that of a diviner not an originator, in that much of the contents of the research covenant are drawn from the experiences that have structured the reflexivity of knowledgeable and practiced peace and conflict researchers discussed above. In other words, the research covenant is embedded in the experiences of several expert peace and conflict researchers. These are people who have shown serious ambition to develop ‘sound knowledge’ and good ethical practice and who have wrestled thoughtfully with the ethical challenges their research has provoked while also feeling themselves under an ethical responsibility to make a difference to the individuals and communities they study.

In this spirit of divination, therefore, I suggest that a research covenant for ethical practice in peace and conflict research should comprise the following moral obligations and commitments:

• The use only of adequately trained and experienced researchers familiar with the difficulties of working in conflict zones and in societies emerging out of conflict.

• The need to plan ahead to anticipate ethical issues and to plan for adaptability and flexibility in the field to deal with the unexpected.

• The need for full, open and comprehensive disclosure in negotiations of informed consent, including any potential risks of subpoena and whether informed consent has to be oral rather than written for some or all respondents.

• Honesty and openness in the status of the guarantees of confidentiality and anonymity.

• A detailed risk assessment to anticipate known risks, and a strategy for the management of risk, whether to respondents, researchers, gatekeepers, translators and all those directly or nominally affected by the research, including a calculation of whether the information is worth the anticipated risks in collecting and reporting it.

• A detailed plan for the management of data and its safety and security.

• Managing the problem of ‘guilty knowledge’ and how best to frame the research to reduce or eliminate the chances of uncovering it.

• Addressing the anticipated complications around dissemination and publication, including consideration of strategies like delaying publication, not publishing some data and developing a writing style that is not sensationalist or which embellishes.

• The need for reciprocity in the relationship with the researched community so that something is returned as a form of intervention, acknowledgement and recognition, that goes some way to making a difference to their lives.

The development of a research covenant for ethical practice with these moral commitments is a necessity for peace and conflict researchers to restore some of the damage done in notorious cases of malpractice. This research covenant can help restore trust between peace and conflict researchers and respondents in dangerous fields in two ways. First, the research covenant makes explicit that the research will not bring harm to respondents by irresponsible, careless, illegal and professionally inept and unethical behaviour on the part of researchers. Second, it makes explicit that the research involves no dishonesty, deception or dissimulation by the researcher, so that what the respondent gets is what the researcher says at the outset in negotiating informed consent.

Conclusion

Peace and conflict research will always present higher risks and greater ethical challenges than research done in non-dangerous fields. Its challenges are several and severe, as this article has highlighted. There are practical and methodological problems involved in this kind of research, and it presents deep emotional strains, ranging from the loneliness and stress involved in working in dangerous and isolated locations to the risk of secondary traumatisation from hearing people’s narratives. This article has focused solely on its ethical challenges.

Despite the severity of these ethical challenges, the adoption of a research covenant can help eliminate, manage and ameliorate the known risks and try to anticipate the unknown ones. Careful planning before entry into dangerous fields is essential. During fieldwork, it is essential that ethical practice is maintained, whether through conventional ethical codes, localised ethical knowledge or a combination of both, but ethical practice within the limits of the law and within the constraints imposed by the risks of the setting. Post-fieldwork, ethical practice requires consideration given towards publication, dissemination and writing style, as well to discharging the ethical responsibility to return something back to respondents.

Above all, ethical debates need to be conducted ethically. This means without emotion or sensationalism and by extending empathy and understanding to the on-the-hoof decisions peace and conflict researchers have had to make when in situ. These on-the-hoof decisions can involve departure from the textbook portrayal of ideal ethical practice and from conventional ethical codes. These deviations and departures need to be understood and accepted by ethics clearance committees, supervisors and principal investigators, who often have unrealistic expectations of what ethical practice means in dangerous fields.

Empathy and understanding are also necessary among researchers themselves who are debating ethical practices, for an ethical debate on ethics practices requires sensitivity being shown towards the different choices circumstances have led some researchers to make. One way of reducing emotional intensity in ethical debates is to codify ethical practice around some generic moral commitments that help better in clarifying and shape the ethical obligations and responsibilities of peace and conflict researchers. The research covenant outlined here is proffered as a prolegomenon towards this.