Exceptions to blanket anonymity for the publication of interviews with refugees: African refugees in Israel as a case study

Abstract

Literature on the ethics of researching refugees, both as participants and partners, presents strong arguments for why anonymity is the safer option in the event of questionable consent. However, blanket anonymity, without asking refugee interviewees if they wish to be anonymous, may cause more harm than good in certain contexts. One such context which this article will explore is the context of Israel, where a working Refugee Status Determination (RSD) system has yet to be established. This case study highlights that, even in extreme circumstances where identity exposure can create risks, there may be a case for allowing identity exposure and perhaps an obligation on the part of the researcher to publish a name if this is the strong wish of the interviewee. This article builds on literature addressing confidentiality of identities (Duvell et al., 2010; Giordano et al., 2007; Kelly, 2009; Lahman et al., 2011; Lee, 1993; Mackenzie et al., 2007; Pittaway et al., 2010) and literature addressing informed consent (Hugman et al., 2011; Mackenzie et al., 2007; Pittaway et al., 2010; Zion et al., 2010). Primary sources demonstrating the use of waiving anonymity were collected between 2007 and 2012. Data were collected from refugee and activist civil society publications, television and newspaper media reports, participatory observation in the Israeli Knesset, and in-depth interviews with refugees and asylum seekers from South Sudan, Sudan, DR Congo, Liberia and Eritrea living in Tel Aviv, Jerusalem and Eilat.

Introduction

The exposure of a real name attached to a real story can carry great risks for refugees who face the possibility of deportation. Yet, anonymity may mean forgoing certain benefits relating to livelihoods, self-empowerment and advocacy.

Duvell et al. (2010), Lahman et al. (2011), Lee (1993), Mackenzie et al. (2007) and Pittaway et al. (2010) emphasize the need to secure information on refugees’ identity. Literature often warns researchers against relying on verbal and written consent alone when refugees do not have self-autonomy (Cooper et al., 2004; Duvell et al., 2010; Feinberg, 1989; Mackenzie et al., 2007: 302). At the same time, Mackenzie warns against ‘paternalism’ and assuming that refugees do not have the capacity for autonomy, supporting a ‘relational understanding of autonomy’ (Mackenzie et al., 2007: 309, citing Friedman, 2003; Mackenzie and Stoljar, 2000; Meyers, 1989) which entails a ‘negative obligation’ on the researcher ‘to ensure that participation in research does not compromise or further erode participants’ capacities for self-determination …’ (Mackenzie et al., 2007: 3010).

Although there may be a ‘negative obligation’ (Mackenzie et al., 2007: 3010) not to publish names, it may still be ethically permissible, and even preferable, to publish names in certain contexts when this is the wish of the refugees interviewed. Although this may seem obvious when there is little risk to the well-being and life of the refugees, it is less obvious when refugees insist on name exposure that could create harm in the future. Therefore, such a claim demands a careful analysis of the particular contexts where refugees’ choices ought to be respected even if this may cause harm in the future.

This article will use a single case study of research with refugees in Israel to better capture the importance of context to emphasize that a blanket policy of never publishing real names may at times undermine the principle of autonomy and ‘doing no harm’ that researchers hope to uphold. The reason Israel was chosen as a case study is because of the very high media and academic presence and frequency of political discussion (Mutz, 2006: 49). Within this context, names of refugees were often exposed before researchers interviewed refugees. At the same time, Israel was and is an environment in which refugee protection is often ad hoc and susceptible to media exposure and public opinion shaped by frequent political discussion. Identity exposure may have garnered public sympathy in some contexts but risked deportation in others.

The article will first present the wider debate surrounding name disclosure, the ability to ensure consent in refugees’ research participation, and debates regarding autonomy and consent more generally. Next, this article will show why anonymity may cause more harm than good in situations where name exposure assists in livelihood strategies and where a name is likely to be leaked by the press regardless of the actions of a particular researcher. In such situations where neither the refugee nor the researcher can ensure confidentiality in other spheres, refugees may have an interest in having their names published within academic research, which often has the advantage of publicizing the wider context and life histories that refugees offer. Finally, notions of autonomy and principles of consent provide strong ethical reasons against blanket anonymity.

Fieldwork was carried out in Israel in Tel Aviv, Jerusalem and Eilat between July 2007 and November 2010, and for an additional two weeks in September 2011. Policy documents, bylaws and interviews via Skype were also conducted between September 2011 and February 2012 from outside of Israel. The fieldwork carried out included in-depth interviews with 15 asylum seekers from South Sudan, Eritrea, Darfur, Liberia and DR Congo. These interviewees were interviewed one to four times each throughout the fieldwork period, in addition to continuous informal communication with some of the interviewees within the framework of language classes organized in Jerusalem. In addition to interviews, the author took part in advocacy efforts with refugees and NGOs in the Israeli Knesset, meeting with Members and Ministers of Knesset and sitting in committee meetings pertaining to refugee issues. Therefore, there was a strong element of participatory observation. Being involved in the process of advocacy, and not only research for knowledge, greatly assisted the author in gaining perspective of what research could impact policy change.

The broader debate on confidentiality

Drawing upon Bok’s (1983) premise that confidentially includes the right to share information, Giordano addresses what a researcher should do if, as Ryen (2004) notes, it is unclear if participants wish to be anonymous, or if participant confidentiality is of ‘pragmatic and ethical benefit’. Even if there is a benefit, this ‘ultimately compels us once again’ (Ryen, 2004) to ask why participants should not be able to choose when disclosure of their names is of ‘pragmatic and ethical benefit’. Unintended consequences for the refugee and research would not necessarily justify ‘absolute confidentiality used and enforced in all cases’ (Giordano et al., 2007: 270).

Even anonymity as a default position has been questioned (Walford, 2005). Kelly (2009) cites Grace (2002), who argues that identity can be uncovered through context so the promise of anonymity is a weak one. Kelly responds that it is possible to protect identity enough to promise anonymity ‘in the first instance’ (Kelly, 2009: 433) so long as expectations of the limits of protecting anonymity are clear. Opposing anonymity because many wish to be identified ignores the argument that many may not wish to participate in a study if they were to be identified. Because revealing one person’s identity can reveal the identities of other participants, compromise is often not possible, thus justifying the default position.

Kelly rightly criticizes those who argue that, because anonymity encourages over-generalizing about research findings, anonymity should not be the default. This argument itself, Kelly notes, generalizes about research that makes anonymity the default. However, it is unclear why the choice to make anonymity the default should not also be dependent on context. Even O’Reilly et al. (2012), who asked research subjects if the policy of anonymity was a good idea and learned that some wished to disclose their names, placed anonymity as the default, potentially influencing preferences.

Perhaps non-disclosure ought to be the default because once a name is released, it cannot be retracted, but a name can be exposed if it is anonymized in the first instance. Furthermore, perhaps a completely context-dependent approach is impossible, as the researchers’ judgment on the risks of disclosure in a given context is itself dependent on aggregate data. For example, if children are far more likely to feel pressure to consent to participating in research to begin with, as seen in the research by Dyregrov et al. (2000), and if children are less likely to wish to disclose their identity, as was seen in the single qualitative case of O’Reilly et al. (2012), then perhaps there are reasons for always choosing anonymity as a default, or even blanket anonymity, when interviewing children.

Although this broader discussion addresses general harms caused from name disclosure, there is a related debate on the inability to ensure true consent when refugees’ choices are all ‘potentially horrendous in their consequences’, such as refugees who are in detention centres in Australia (Zion et al., 2010: 52, citing Raz, 1986). ¹ Zion et al. argue against any research in such situations, anonymous or otherwise. Yet, Strous and Jotkowitz (2010) argue that not allowing asylum seekers to participate, out of fear that they do not have the capacity to make autonomous decisions, can undermine their autonomy. Similarly, Rousseau and Kirmayer (2010: 65) argue that, even if ‘freedom to consent … is in question, deciding for the detained refugees whether or not they are “free” is a further attack on their … agency.’ These arguments could similarly be raised in favour of allowing refugees and asylum seekers to choose whether they wish to be anonymous.

However, let us assume that it is possible for researchers to decide objectively that an individual does not have the capacity to consent. If those whose capacity we are measuring could decide for themselves if they had this capacity, the concept may be meaningless, or lead to an infinite regression where we would need to ask if they had the capacity to decide if they have the capacity to decide, etc. Admittedly, a similar infinite regression arises if we assume researchers have the capacity to decide that they have the capacity to decide refugees’ capacity to decide. Putting this debate aside, even if subjects do not have the capacity to consent, Strous and Jotkowitz (2010), in their response to Zion et al. (2010), note that there are many examples where ‘utilitarian logic trumps following adequate ethical analysis’. This is quite extreme. Perhaps ‘adequate ethical analysis’ includes other values besides the capacity to consent. Stoljar (2000), citing Christman (1991) (my emphasis), states that ‘an agent is autonomous with respect to a certain preference or desire only if she did not resist its development when attending to the process of its development or would have resisted had she attended to the process’ (Stoljar, 2000: 100). If a refugee cannot attend to a process because the choices are constrained, what they would have decided may be more important than their lack of ability to attend to a process. In addition, Frankfurt (1988) has argued that even if one’s values are formulated in non-autonomous circumstances, perhaps desires are autonomous if the agent holding them identifies with them (Arneson, 1994: 57).

Parfit (2011) does not view an agent’s inability to consent completely freely as the only factor in deciding how this person should be treated. He distinguishes actual consent, which is not always possible to obtain, from the ‘principle of consent,’ which is that it is wrong to treat someone in a way in which they could not rationally consent. ² If a person is hanging off a cliff, Justo (2010) argues, in a response to Zion et al. (2010), throwing them a rope and telling them to ‘hold on’ may be the morally preferable action, even if this means not telling them full information about the measurements of the rope and the ‘possibility of being scratched while being pulled up from the cliff’ (Justo, 2010: 61). In the context of anonymity, if name disclosure could immensely and immediately help a subject and they ask for their name to be disclosed, their inability to provide true consent may be less relevant than whether they would have rationally consented, even with the harms involved.

Even if the ethics of name disclosure is dependent on actual consent, as opposed to the ‘principle of consent’ or what they would have consented to hypothetically, perhaps the default should be dependent on which default minimizes risks. In other words, the ‘negative right’ could be the right to non-disclosure in some contexts, with the right to disclosure the default. If a name has already been exposed in the media prior to the research, and changing ‘characteristics … beyond recognition’ can ‘blunt the power of the narrative’ (Stein, 2010: 557), then it may be more ethical to make disclosure the default. Although the arguments in this article are mainly to question blanket anonymity, they could also serve as reasons for making the default context-dependent, thus building on Giordano’s analysis of the benefits of context-dependent options for waiving anonymity when anonymity is the default.

African refugees in Israel

The arguments in favour of cancelling blanket anonymity, and allowing refugees the choice to expose their names, is within the context of a single case study of African refugees in Israel. According to data from Israel’s Migration and Population Authority, cited in an October 2012 Knesset Research and Information Center report, there were a total of 22,000 ‘infiltrators’ as of 2009, the vast majority of whom crossed over into Israel, by foot, from Egypt. Of these, 17,630 were under a ‘group protection’ scheme. The total numbers were 62,000 and 53,500, respectively, as of 2012. There were 4300 asylum seekers from Sudan and 1250 from South Sudan in 2009, with 14,850 Sudanese by 2012 and no South Sudanese that same year after all returned when their group protection was revoked, according to the report (Gilad, 2012). However, the author is aware of at least five South Sudanese who are still in Israel as of January 2013. There were 9960 Eritreans and 2030 Ivorians in 2009 and, in 2012, 36,770 Eritreans and no Ivorians, who all returned when their group protection was revoked, according to the report. There were 390 asylum seekers from Congo in 2009 and 440 in 2012 (Gilad, 2012). There are an unknown number from Ethiopia, Ghana, Nigeria and Togo (Gilad, 2011).

Since 2005 an unknown number of asylum seekers have been deported at the border back to Egypt in a policy dubbed ‘Hot return’ (Human Rights Watch, 2008; Kritzman-Amir and Berman, 2010: 640 citing 22, HCJ 7302/07; ARDC, 2011). Those from Eritrea and Darfur have prima facie protection if they manage to stay in the country for more than 72 hours and not be returned immediately at the border, though asylum seekers who have crossed the border at the end of 2012 have been put in detention. Those who have no prima facie protection must apply as individuals in an RSD process that was controlled by UNHCR until 2009, after which the government conducted RSD (US Committee on Refugees and Immigrants, 2009). The process has more recently undergone reforms for further judicial oversight (Population, Immigration and Border Authority, 2011a).

Within Israel, children were not given the right to education in Eilat at the time this research was conducted (interviews at the Nuba Mountain Community Center, 2008; Gilad, 2010; Lior, 2011). Public medical insurance is denied (Kestler-D’Amours, 2011) despite petitions to the Knesset (Knesset Protocol 64, ‘Committee for the Problem of Foreign Workers’, 11 July 2011). Employers often pay refugees less than minimum wage (Furst-Nichols and Jacobson, 2011). The prima facie protection often only means that non-refoulement is respected; basic rights to work, education, and health are not.

Rights to both stay in Israel and work can be revoked any day (Furst-Nichols and Jacobson, 2011), and the provision of visas is often arbitrary. For example, in 2007 and 2008, 2000 Eritreans received 6-month renewable B-1 visas, whereas 600 Darfur refugees received the renewable A-5 multi-year temporary residency visas. Darfur refugees who came after 2007, or even after the decision was made in July 2007, did not necessarily receive these visas (Human Rights Watch, 2008). On the other hand, although the RSD process is not following international law, the protection mechanism is not completely arbitrary and may be influenced by public opinion and media attention.

Within this context researchers interviewed over 50 refugees between 2007 and 2010, and the author was personally involved in conducting 15 of these interviews. Interview testimonials were sent to Members of Knesset. The policy was to change all names, in the event that the exposing of a name revealed an identity.

Yet, one refugee, Gabriel, used the text of his interview for a website he started, putting his own full name, suggesting that the researchers did not follow his preference (Kuol, 2008). There may be a need re-evaluate the ethics of encouraging anonymity and assuming anonymity is preferred without asking if this is the preference. The risks and harms of anonymity must be re-evaluated and the autonomy of the interviewee must be addressed.

Risk and harms

Academic researchers do not necessarily have the same moral obligations or roles as a media outlet. At least one role of the academic researcher which may be distinct from media research is to ‘do no harm’. Jacobson and Landau (2003: 192) address the ethical dilemmas that arise from the idea of ‘doing no harm’ and Duvell et al. (2010) argue that ‘the main ethical question to arise is whether the potential social benefits from research outweigh the potential social harms’. The ‘bottom line’ for them was that their ‘research was conducted and disseminated in a way that meant enforcement agencies could not identify the whereabouts of individuals or groups of irregular migrants’ (Duvell et al., 2010). They cite Beyrer and Kass (2002), Lee (1993) and Sieber and Stanley (1988), who address research activities that put participants at risk.

The consequences of exposing names is clearly a factor, even if this potentially undermines autonomy, as Giordano et al. (2007) note. Yet, in some instances, academic attention, while exposing identities to authorities, can also expose identities to a wider public, who may, if they are sympathetic, influence policy-makers in a way that increases protection. If the role of the academic is to ‘do no harm’, and if harm is predicted based on past events, then there may be value in exploring the impact of non-academic media exposure in determining the potential harms and benefits that arise from researchers’ choices with regard to anonymity. In other words, undermining the negative right to disclosure could have negative consequences.

In July 2007, roughly 50 asylum seekers from South Sudan and Darfur made their way in protest from Beer Sheva, Israel to the Rose Garden in front of the Israel Knesset in Jerusalem. After having crossed the border from Egypt on foot, and after finding no municipal assistance in the southern Israeli city of Beer Sheva, they camped out on the lawn in front of the Knesset where, with the help of local civil society groups and charities, they received food, formula for infants, sleeping bags, medical assistance, and media attention. With hundreds of Israelis of all political backgrounds visiting the lawn of the Rose Garden to provide necessities, and with media outlets interviewing for individual testimonials, names and identity-disclosing characteristics were revealed, such that journalist did not need to ‘blunt the power of the narrative’ (Stein, 2010: 557). The emphasis was on individual refugees, rather than on ‘flows, streams, waves, and trickles’ (Turton, 2003: 10).

News articles addressed the plight of refugees, with the focus mostly on those from Darfur, at times ignoring South Sudanese refugees. Media attention came from Associated Press (2007), IRIN (2007), and one iconic image from photographer Yoav Lemmer for Agence France-Press (Lemmer, 2007), later used by other networks, including the BBC (2007a; and see 2007b). The face in this image is clearly exposed and may have helped garner public sympathy. Eventually, slightly more than 500 Darfur refugees, including those in the Rose Garden, were given temporary protection. South Sudnaese, who received less attention, did not. The exposure of the identities of Darfur refugees almost certainly played a role in providing limited protection.

In a meeting with a Minister of Knesset in 2007, a refugee from Darfur active in advocacy for refugees introduced himself with his full name. He was told by a Minister of Knesset, ‘Do not worry. You will not be deported.’ Personal stories, rather than a systematic evaluation of claims alone, may assist in protection. This is an ethical dilemma that Pittaway et al. (2010: 233) address in a study in which research led to resettlement of 100 refugee families ‘in a particular refugee camp in an African nation to countries in the West’ without proper assessment of who in the camp was most at risk and in need of resettlement. However, this example does show that risks are not always an indication of net harm. If Pittaway et al.’s research had not been conducted, it is unclear that anyone would have been resettled, which also would have been a scenario without proper assessment.

Even if the emotional appeal from waiving anonymity cannot be shown to have a direct impact on protection, identity exposure is often a necessary byproduct of general media exposure. Waiving anonymity can therefore be necessary to gain greater protection in situations where media attention creates greater sympathy. While working for a short time at Channel 2 News in Israel, the author asked if it was possible to air a piece on refugees. The response was that, unless faces were shown, a piece could not be aired; fewer viewers would watch without a face (participatory observation, 2008). When refugees decide to create their own civil society organizations to fight for fair treatment and basic rights, this often necessitates showing one’s faces and stating one’s name, such as with the B’nai Darfur organization, whose directors often show their faces, including on their website (http://bnaidarfur.org/).

On the opposite end are examples of refugees who receive absolutely no media or research attention as individuals and who remain forever anonymous. Two examples of such forced anonymity stand out. The first example is victims of the ‘Hot Return’ policy at the border with Egypt. A border patrol document reveals that 168 Eritrean asylum seekers were deported on 16 June 2009 (Israel Defense Forces [IDF], 2009). Their names do not appear on any official document. Refugees have testified to other ‘hot returns’ at the border through human rights organizations, including Physicians for Human Rights (Hartman, 2011a). The IDF spokesperson openly stated that refugees are deported to Egypt, choosing to defend the practice (Anshil, 2011) rather than deny it. Particular faces and identities known to the public and NGOs may also be contributing to a policy of non-refoulement. This exposure of identities is not possible during the first few days in Israel, when deportations occur before names can be recorded from UNHCR, NGOs, reporters and researchers.

A second example is in two forms of ‘assisted voluntary return’ (AVR). In one form of AVR, those asylum seekers who could not receive prima facie protection (those who are not from Eritrea, South Sudan or Sudan) were given the option to return with the assistance of a non-governmental organization (NGO Director, 2011, personal interview 22 September; AVR interviewer 1, 2011 personal interview, 12 April; and interviewer 2, 2011, pers. comm., 11 December). A similar programme for South Sudanese and Sudanese also ensured confidentiality and anonymity through another NGO, (South Sudanese AVR Program Director, 2010, personal interview and anonymous Director, 2012, personal interview). For both NGOs, confidentiality and anonymity were ensured in order to protect returnees. The option of opting out of anonymity, in the sense of having both AVR NGOs state their names to the press, was not offered for legitimate security reasons. At the same time, follow-up of returnees is difficult, and in some cases impossible, owing to the anonymity of the identities. This example is not to argue that anonymity should be waived in all circumstances where refugees wish to waive anonymity, but merely that anonymity can mean that, should a refugee or asylum seeker leave, follow-up is difficult.

Yet, in 2011, a well-known Darfurian co-owner of the popular ‘Humus Gan Eden’ restaurant decided to move to South Sudan and was killed after his return. His name, Adam Mohammed, was reported in the Jerusalem Post of 23 December (Hartman, 2011b). The larger public, including the Sudanese and South Sudanese committee, became aware of what had happened to him. Similar knowledge is difficult to obtain when refugees who choose to return, or are forced to return, have no way of publicizing their names before leaving.

Kelly rightly argues that anonymity is the default because one individual revealing their name exposes those who are associated with the individual and also interviewed, such as other teachers and students within a school. Yet, exposing a name, in this instance, could assist in providing guidance for others considering return, and could allow the type of further name discloser that could help provide more information as to the general risks of ‘voluntary return’. Three months after Adam’s death, the author conducted field work in South Sudan over the course of March and April 2012 with those who had returned from Israel between 2010 and 2012. It was possible to interview some returnees because a select few had consented to disclosing their names to the author, by way of the NGOs facilitating return and human rights NGOs who had worked with the refugees before their return. Those whose names had been disclosed introduced the author to other returnees, which invariably led to more name-disclosing, often with identity-defining characteristics disclosed to the author before a formal introduction. Disclosing anonymity of one individual can, as Kelly rightly argues, disclose the names of others. Yet, completely secure anonymity can inhibit researchers from following up on the consequences of policies and, indeed, anonymity itself.

There are times, writes Pittaway et al. (2010: 232), where ‘the desire of the refugees themselves to have their “stories” told to the international community can outweigh consideration of the potential danger to themselves and their communities.’ It is important to remember, however, that the existence of potential dangers can be outweighed by the benefits that telling one’s story can provide. Pittaway et al. promote a relationship of participation, where refugees have a say in how information is used. They argue that ‘ethics should be extended to promoting the interests and well-being of extremely vulnerable research subjects, in ways that make sense for the research subjects, as well as for researchers and their academic institutions and professional bodies’ (Pittaway et al., 2010: 241−242). In some instances waiving anonymity may put a refugee or their community at extreme risk yet also ‘make sense’ for both the interviewee and interviewer.

Much literature, including Giordano et al. (2007), Kelly (2009) and Pittaway et al. (2010), argue that anonymity should be the default or the policy to be promoted by the researcher. However, there is often a failure to explore the distinction between ‘anonymity’ and ‘confidentiality’, and these terms are used interchangeably. Lahman et al. (2011) promote ‘Culturally Responsive Relational Reflexive Ethics (CRRRE) oriented research’ where ‘anonymity/confidentiality, and consent are advanced’ (Lahman et al., 2011: 304). Duvell et al. (2010) ask ‘how the researcher will manage to obtain consent in research that has the potential to damage their livelihood and plans?’. One approach they offer ‘is not to promise anything but anonymity and confidentiality …’ (Duvell et al., 2008: 20). Bilger and van Liempt state that ‘besides agreed standards of guaranteeing anonymity and confidentiality, building up trust requires researchers to understand the situation respondents find themselves in’ (Bilger and van Liempt, 2009: 123).

The lack of distinction between anonymity/confidentiality is problematic because, while refugees who wish to stay anonymous to the broader public must receive confidentiality from the researcher, refugees need not be anonymous if they do not wish to be. One can ensure confidentiality for refugees who wish to remain anonymous without always promoting anonymity. This is in no way to overlook the very real risks involved for refugees who choose to no longer remain anonymous. South Sudanese, for example, were deported in 2012 and knew, as of 2011, that this was a real risk (Population and Immigration Authority, 2011b). Yet, if revealing a name in some instances can cause less harm through public awareness of identities that, in some instances, can lead to sympathy, then refugees who are interviewed should make the final decision based on their own assessment of risks.

Another reason to waive anonymity is because anonymity cannot be secured. Kelly rightly argues that the inability of researchers to secure anonymity is not a reason not to offer anonymity to the best of one’s ability in the first instance. Yet, research subjects may have already been exposed to the media or feel that, if there is a risk of names exposed, they prefer to be exposed to an even wider audience. If, as Duvall states, the exposure of irregular status to government authorities can lead to apprehension (Duvell et al., 2010: 232), then one method of avoiding apprehension, if names are already at risk of exposure, may be exposure to an even wider public. In some limited contexts such exposure can generate public sympathy and counteract, to an extent, the negative consequences of name exposure to authorities. This was true for Darfur refugees who risked deportation back to Egypt and from there to Sudan in 2007. Although it is clear that those who did not wish to be exposed should have been ensured confidentiality, it is less obvious that the default option should have been anonymity; as the case was, the default that reporters offered was exposure, and refugees had the option of opposing exposure.

A final reason for the waiving of anonymity is the publicizing of businesses opened by refugees. Faida, a refugee from DR Congo, has her name publicized within the context of her nursery. Her name, Faida Bakaji Tshuma, is central to her livelihood, as her reputation as an excellent caretaker brings her potential parents who wish to enrol their children in her nursery. In 2010 she won the International PlaNet Microfinance Award in Paris; her name was central to this honour. Her full name was also used in academic research on her business model (Kruchik-Krell, working paper), which allowed her to receive further exposure.

This public information about her identity complements her very vocal roles in civil society organizations in Israel. Indeed, her exposure is not only about business: it is about empowering herself by exposing her full identity. It is, perhaps, about a right to share information, as argued by Bok (1983). To assume that the exposure is inducement owing to the refugees’ low socio-economic status possibly ignores the right of refugees to calculate, for themselves, the harms and benefits of exposing their names.

A right, however, is different from a capacity for autonomy, which could include a minimal threshold to assess risks. The above examples do not show a causal link between disclosure of identity and greater safety. Aggregate data would assist in understanding particular contexts, suggesting that, if revealing names would, on average, be more risky, then there would be a strong case for blanket anonymity. It would still be context-specific in that, in the context of refugee research, all subjects should be anonymized. Yet, not only is it unclear what the risks are, and not only may refugees have greater knowledge on the aggregate level, but principles of autonomy, which will be addressed in the next section, are not only dependent on outcomes of decisions or average outcomes of aggregate decisions.

Autonomy

A strong argument for blanket anonymity is that refugees may not be able to provide true consent. If consent is impossible because of coercion, best to choose the option that ensures refugees’ safety. However, if anonymity is not always in the best interest of refugees, and if refugees may wish to make an informed decision and expose their names, then denying the right to waive anonymity may at times deny the right of refugees to autonomous decision-making. If so, there may be a positive obligation to publish a name in some contexts. At the very least, there is no negative obligation not to publish such names.

This argument more closely ties into the broader debate on autonomy and consent. Although the term ‘autonomy’, discussed earlier, is a highly debated and contested phrase, it is used widely in medical research and closely associated with the principal of informed consent (Frankel and Siang, 1999). Mackenzie et al. (2007) cite Feinberg (1989), who differentiates between the capacity for autonomy and the right to autonomy. The former ‘involves determining the direction of one’s life, making considered choices among one’s options, and acting in accordance with one’s own beliefs and values.’ The latter ‘entitles persons to the social, political and economic protections that enable them to exercise these capacities for self-determination’ (Mackenzie, 2007: 302). ‘Agency’, vs autonomy, is the word used by Hugman et al. (2011) when they state that ‘informed consent is a central concern of research ethics as an expression of human agency’ (Hugman et al., 2011: 656).

Regardless of how one defines autonomy or consent, and regardless of the extent that actual consent is central in circumstances where choices are all ‘potentially horrendous in their consequences’ (Raz, 1986), there appears to be a consensus that having autonomy and agency is not having the ability to make risk-free decisions. Rather, it is either having the ability to assess these risks or, alternatively, only being treated in a way in which one would not object to being treated were one able to assess the risks. Regarding the first possible definition, Meyer (2006) states that ‘limitations on refugees’ ability to achieve intended outcomes should not be interpreted as a lack of agency’ (Meyer, 2006: 29). Assuming that individuals really do have capacity for autonomy, the less we assume that these individuals have the capacity for autonomy, the more we may take away their right to autonomy, as argued by Strous and Kotkowitz (2010). However, unlike Strous and Kotkowitz, this argument recognizes that if refugees really do not have autonomy, assuming they do not will not undermine what they already do not have.

Therefore, as Mackenzie et al. rightly point out, listening to the declared consent of refugees is not sufficient to confirm the right to autonomy if one defines capacity to autonomy as the ability to assess risks:

… consent provides protection on the assumption that participants are autonomous, understand the implications of giving consent and are in relatively equal positions of power with researchers. These assumptions are unjustified in many research contexts in which research participants are vulnerable, but particularly in the context of refugee research in situations of crisis and conflict’ (Mackenzie et al., 2007: 302).

Mackenzie et al. (2007) cites Feinberg (1989), who argues that ‘Spending protracted periods in camps may compromise refugees’ capacities for autonomous agency’ (Feinberg, 1989), a similar argument provided by Zion et al. (2010). Duvell et al. (2010), citing Cooper (2004), argue that the capacity to make informed decisions may be limited by illiteracy and the ‘ability to follow public debates, understand the media discourses and assess the risks involved in the research that concerns them’ (Duvell et al., 2010: 232). This is clearly not the case for the leaders of Bnei Darfur, who were active in civil society discourse, or for Faida, who is a highly educated professional who speaks fluent Hebrew, English and French and runs her own business. Yet, this could be the case for refugees who had never learned Hebrew, were illiterate, and/or were without the basic freedoms in detention centres.

Yet, rather than argue that actual free and informed consent is not the only value, as Justo does in evoking the woman hanging from a cliff, Mackenzie et al. (2007) warn against ‘paternalism’ and assuming that refugees do not have the capacity for autonomy. In other words, they focus on the risks of assuming no autonomy, not the type of relationship a researcher might have when there is no way to ensure autonomy. In this way, they do not address the same dilemmas raised by Zion et al. (2010), Justo (2010) and, more generally, by Parfit (2011). Therefore, Mackenzie et al. suggest a ‘relational understanding of autonomy (Friedman, 2003; Mackenzie and Stoljar, 2000; Meyers, 1989)’ (Mackenzie et al., 2007: 309). Such an approach entails a ‘negative obligation’ on the researcher ‘to ensure that participation in research does not compromise or further erode participants’ capacities for self-determination or their scope for exercising these capacities’ (Mackenzie et al., 2007: 3010).

Mackenzie et al.’s suggestion of ‘iterative consent’, used by Miller (2004) and Jansen and Davis (1998), is meant to build trust and ensure that the wishes of refugees are being followed. Again, this is an attempt to ensure actual consent, and not to decide what actions should be taken if actual consent is not possible, as Zion et al. (2010) claim was the case in Australian detention centres. Similar to ‘iterated consent’ are suggestions by Hugman et al. (2011: 662), who outline ‘participative action research’, which ‘necessarily involves the continuing relationship of researchers with participants.’ This in some ways addresses the problems raised by Ryen (2004) that researchers may not know if subjects wish to be anonymous; this, too, does not address what researchers should do if there is no way of entirely knowing if subjects wish to be anonymous, because choices are all ‘potentially horrendous in their consequences’ (Zion et al., 2010: 52, citing Raz, 1986).

It may appear, however, that iterative consent at least assists in ensuring consent if such consent is possible. Yet, this again returns us to the image of a person hanging from a cliff; iterative consent would take time, and this could lead to the individual falling from the cliff. It is not clear that this person can even truly give completely voluntary consent. It may be necessary to ask, as does Stoljar, if a person would have resisted had they this capacity to resist. In other words, would individuals have resisted to name exposure had they the capacity to do so. If the answer is ‘no’, then perhaps one can argue that a potential research subject does have autonomy if defined not as the ability to consent, but the ability to make choices that they would not have opposed had they been able to consent. It may also be ethical to ask, as does Parfit, what a person could rationally consent to, even if they cannot provide actual consent. If a researcher has strong reasons to believe, and a refugee strongly insists, that they would have consented had they the capacity to do so, then it is not clear that iterative consent is necessary or even preferred.

For example, one of the situations that most demands quick consent, and where the waiving of anonymity can result in concrete benefits, are precisely those times in which there is great urgency before a scheduled deportation. A family from the Ivory Coast chose to publish their names, faces and histories in Haaretz newspaper in an article discussing a planned deportation (Weiler-Polack and Lecker, 2012). If Weiler-Polack and Lecker had followed the ethics guidelines of iterative consent before interviewing or waiving anonymity, then the benefits that would make the interview ‘reciprocal’ for the refugee would have been destroyed. This particular family chose to expose their names at the last minute, because it is at the last minute when such exposure is deemed necessary. The very urgent and uncertain future for many refugees does not mean consent is induced; on the contrary, waiving anonymity can be a calculated and worthwhile decision that can lead to long-term protection. However, even if this means consent is induced, it could be ethical for the researcher – and a reporter – to still publish names because it is what the family would have wanted had they the autonomous capacity to make the decision.

Let us assume, however, that autonomy is dependent on capacity, and not on what an individual would have preferred had they the capacity to assess the risks. Let us assume that this capacity is undermined owing to unequal power relations between researcher and subject. This could be because subjects really do not identify with their stated preferences (Arneson, 1994: 57, citing Frankfurt, 1988), but feel pressure to declare a consent they do not feel or because, in contradistinction with Frankfurt, autonomy is historically contingent and so genuine desires are formulated as a result of coercive circumstances of unequal power relations between researcher and subject. Even if this is the case, adding an explicit ‘relationship’ is not necessarily adding any more equality of assurance of consent to the relationship. For example, the members of Bnai Darfur, whose lives would be at risk were they to be deported back to Sudan, nonetheless show their faces on their website, in addition to media sources. This information is used by media outlets and researchers (Sudilovsky, 2009) despite the lack of a ‘continuing relationship of researchers and participants’ (Hugman et al., 2011). The most effective way of empowerment may be for refugees to publish their own work, and have this cited by researchers; such a scenario would have no relationship other than a professional respect for the integrity of the other’s ideas.

If this is the case, a refugee may not want to build up a relationship with a researcher over time but does wish to have their stories told and their names exposed through the researcher. In such instances, ‘iterative consent for iterative consent’ would be necessary. This could lead to a paradox. If iterative consent for iterative consent needs to be obtained, then iterative consent is forced upon the refugee, either way. Or, alternatively, an infinite regression arises, in which the iterative consent for iterative consent demands iterative consent, etc. In this case, in addition to the problem of consent and an infinite regression, there are particular potential harms created by iterative consent, because the passage of time could undermine the benefits of pushing off a deportation. If autonomy is both a capacity and right, as Mackenzie et al. (2007) argue, demanding iterative consent may, at times, undermine both.

Conclusion

The importance of ensuring anonymity when this is the wish of a refugee interviewed, or when a refugee does not indicate that they wish their name shown, is central. Yet, if the capacity for autonomy means that refugees can weigh risks and consider indirect and intangible benefits, then to deny them this option − by assuming that they cannot weigh the risks and consider benefits – both undermines their rights to autonomy and possibly their opportunity for protection. The case of Israel presents an example where, in some instances, revealing an identity can provide livelihoods, protection and autonomy. In some instances found in the case of Israel, iterative consent would undermine these benefits.

In the context where the risks are too great, a responsible researcher should assume blanket anonymity even if this undermines autonomy. However, in contexts where there is already name exposure through other mediums, and where name exposure is central to refugees’ strategies for securing their own rights, it is unclear that refugees should have to choose between having their stories told anonymously or not having them told at all.