The issue of ’vulnerability’ in researching political elites

ABSTRACT

This paper examines the concept of “vulnerability” in the context of the ethics of researching education policy elites. This concept may seem inapplicable: if elites have power, can they be vulnerable? However, I draw on discussions in the research ethics literature to point out that vulnerability is a matter of degree, relevant to all participants. I argue that people may be susceptible to particular kinds of threat, that relate to different sources and types of potential harm, which must be taken into account in any judgement about whether additional precautions are required to protect them. I illustrate some of the complexities surrounding the issue of vulnerability by discussing how these apply in the context of researching members of a Greek political elite involved in negotiations, on behalf of the Greek Government, with the Organisation for Economic Cooperation and Development. I argue that protecting elite participants from harm was an important consideration in this research, even though this had to be balanced against my responsibility to document the exercise of power and its consequences for Greek society. My paper contributes both to the literature on researching elites and to the field of research ethics, by problematising further the concept of “vulneralbility”.

KEYWORDS:

• Education policy elites
• vulnerability
• harm
• research ethics
• ethical judgements

Introduction

Much sociological research in education and elsewhere has focused on understanding the perspectives and experiences of “marginalised” or subordinate groups, along with those of members of the occupations and organisations involved with them. However, since the 1980s and 90s a small tradition of work has emerged concerned with “studying up”: focusing on the role of elites (e.g. Ball, 1990; Halpin and Troyna, 1994; Walford, 1994). Initially, studies of politicians, government officials, pressure groups, and their networks, tended to assume that they belonged to close-knit groups which could be easily identified (Walford, 2012). However, the shift in the 2000s towards the notion of the governance of education reframed this research to include global policy actors who occupy “multiple spaces” and policy networks, being “simultaneously national and transnational”, and therefore more difficult to identify (Grek, 2021, p. 18; see also Yates, 2004). The aim of this focus on elites was to document the operation of elite power both in shaping educational institutions and in resisting “struggles for social transformation” (Ozga, Gewirtz, Halpin, & Troyna, 1994, p. 123). The term “elite” is not always defined within this literature but is generally used to describe individuals or groups who ostensibly have closer proximity to power or who are able to claim distinctive professional expertise (Ball, 1994; see also Morris, 2009; Khan, 2012; Maxwell & Darchy-Koechlin, 2015). This clearly covers a range of people playing a variety of roles, and this needs to be taken into account in any study, as well as in addressing methodological and ethical issues in this field.

However, despite the growing body of work on educational elites, there have been few reflexive accounts analysing the ways in which this research is conducted, and the challenges encountered (but see Addey & Piattoeva, 2021; Duke, 2002; Lancaster, 2017). In particular, the ethical aspects of research associated with interviewing “elites” involved in educational policymaking remain underexplored (Addey & Piattoeva, 2021). This paper focuses specifically on one of these issues, one that has more usually been addressed in studying marginalised or subordinate groups: vulnerability. The suggestion that this is also relevant in studying elites has been made by one or two previous writers: Smith (2006, p. 651) comments that ‘we need to make room to consider the possibility of “vulnerable elites”; while Neal and Mclaughlin (2009) refer to a multi-layered vulnerability in their reflective account of their research on the Commissioners of the Parekh Report. At face value, the concept may seem inapplicable to research on elites, because of the assumption that vulnerability and power are negatively related: since elites have power, it might be concluded that they cannot be vulnerable. However, I will suggest that the issue is more complicated than this. Indeed, discussions in the research ethics literature have already drawn attention to problems with the concept of vulnerability.

Subject to particular criticism has been the idea – built into research ethics codes, and the procedures employed by Institutional Review Boards (IRBs) and Research Ethics Committees (RECs) – that certain categories of research participant are intrinsically vulnerable, so that heightened protections must always be placed upon research projects involving them.¹ Typical categories of the vulnerable include: young children, people with learning difficulties, sex workers, prisoners, psychiatric patients, migrants, people in poverty, and those suffering from serious illnesses. However, there can be reasonable disagreement about who should, and who should not, be included in this list. For instance, some commentators challenge treating children as vulnerable because this implies that they lack competence or are powerless (Farrell, 2005; Morrow & Richards, 1999; Wright, 2015). Furthermore, members of all the categories designated as vulnerable can vary considerably in relevant respects, for example as a result of their intersectional positions. More fundamentally, at least one writer has challenged the very appropriateness of the concept of vulnerability in the context of research ethics (Van den Hoonaard, 2018, 2020; see also Levine et al., 2004). Van den Hoonaard argues that not only is it difficult to determine who is vulnerable, but also that designating potential research participants as vulnerable often results in their being excluded from studies, so that their perspectives are ignored.

There is a more general criticism, too. This is that research ethics cannot properly be a matter simply of following a set of rules, such as those specifying who is, and who is not, “vulnerable”. This is because, in doing research, there are multiple ethical and methodological considerations that must be taken into account, and these will often be in conflict. So, judgements have to be made in particular situations about what is the priority and how conflicting ideals should be interpreted and reconciled (Hammersley & Traianou, 2012). The strategies and tactics researchers use to achieve their research goals depend, to a large extent, on such judgements.

In this paper I will explore the issue of vulnerability in the context of my own recent research. I will begin by discussing further some of the complexities surrounding the concept of “vulnerability”, arguing that there needs to be clarification of in what respects particular participants may be vulnerable. I will conclude by discussing how these complexities apply in the context of researching political elites, drawing on research in which members of a Greek political elite engaged in negotiations with global policy actors were interviewed (Traianou, 2021).

What is vulnerability?

“Vulnerability” is a relational term: it is necessary to specify to what a person is deemed to be vulnerable. In the research ethics literature, the main concern about vulnerability has been the danger that people may agree to participate in research even though they do not sufficiently understand what the research is about, and what demands it will make on them, and/or that they will consent because they do not feel free to refuse to participate. So, here, “vulnerable” implies an incapacity to provide genuine informed consent. The vulnerable are those who, for one reason or another, are judged to be incapable of being informed, of recognising their own best interests, and/or of exercising free consent. The remedy recommended may be that a third-party exercises informed consent on their behalf, or that they should be excluded from the research because it is uncertain whether informed consent could be obtained.

There is, however, another, rather different, meaning that is associated with the word “vulnerable” in the context of social research. This concerns vulnerability to particular kinds of harm that may arise directly or indirectly from the research process. These harms can be quite various: they may be physical, as for example in the case of studying people with serious illnesses where the research could cause undue fatigue or stress. But harm can also be financial, for example if the research reduces the occupational work being done or deters the clientele of a business. There may also be potential reputational damage, arising if research participants are publicly identified, whether through unwarranted access to the data or as a result of being portrayed in research reports. Given this diversity in the form that harm can take, the remedies adopted may need to be different, whether this is curtailing data collection when people show signs of stress or fatigue, ensuring that the research does not interfere with occupational activities, or securing the data and trying to protect the anonymity of people mentioned in research reports.

While these two kinds of vulnerability may overlap, it should be clear that people can be vulnerable in one of these senses but not in the other. Given that members of elites would normally be regarded as quite capable of exercising informed consent, my main focus in this paper will be on “vulnerability” as “susceptibility to the threat of harm”.

Of course, vulnerability is a matter of degree (Gordon, 2020, p. 35). This is true of both senses of the term I have outlined. In relation to informed consent, participants will often vary substantially in how well informed they are about the research, and they may also vary in how free they are to consent or to refuse consent to participate in a research project (Wiles, 2003, Chapter 3). People differ in their capacity to understand the purposes of a research project and what demands it will make on them, and also in their willingness to learn about this. As regards consent, they may be subject to pressure from an employer or a peer-group (Hammersley & Traianou, 2012, pp. 82–98). In much the same way, the likelihood and severity of harms of particular kinds will vary considerably across different categories of people (Hammersley & Traianou, 2012, pp. 62–4); even though, to differing degrees, everyone is vulnerable to some kinds of harm.²

So the most significant question is not which categories of people are vulnerable, but rather is about degrees or “layers” of vulnerability (Luna, 2009). Given that vulnerability indicates a requirement for extra protection, a threshold has to be set above which this must be provided. Necessarily, then, judgements need to be made about how vulnerable the participants in a study are to harms of particular kinds. Furthermore, it must be remembered that individual members of any category deemed vulnerable will vary in their level of vulnerability, so it is essential to take account of all the relevant social characteristics and relations of the particular people involved. Part of what is entailed here is their intersectional position: people are members of multiple categories, and judgements about their vulnerability may differ depending upon which category is prioritised, and how the categories are related. In short, treating “vulnerability” as a relational term implies that people are not vulnerable per se, but may be susceptible to particular kinds of threat, that relate to different sources and types of potential harm, ranging through physical injury, psychological damage, material loss, and tarnished reputation, as well as obstruction to ongoing activities.³

There is sometimes a tendency to assume that researchers have an all-embracing obligation to protect participants from harm. But they can have no responsibility to keep participants safe from threats that are endemic in the environments in which they live and work. Furthermore, there are some occasions when researchers are ethically required to act in ways that may cause harm to research participants. It is common for codes of ethics, and for RECs/IRBs, to require that researchers report abuse or crimes that they witness in the field and this may well result in harm for those responsible. Researchers are also usually absolved of responsibility for harm to key decision-makers when they reveal discrepancies between stated policies and what actually happens in practice. Clearly, the limits to researchers’ responsibility for harm must be recognised.

There is another aspect of this: while it is often assumed that researchers have power over participants, hence the need for ethical regulation, this is by no means always true. Indeed, social researchers rarely have complete control over the situations where they operate; and they are never all-powerful in relation to participants. Many researchers are postgraduate students or junior members of staff, and they may themselves belong to a marginalised minority or oppressed group. Meanwhile, the research participants can be members of a high-status profession, high-ranking managers in a large corporation, or even Government ministers (see, for instance, Grek, 2011, 2021; Ozga et al., 1994). Nor are such participants reluctant to use their power in relation to research, perhaps asking for confidential information in return for providing data (see Alcadipani & Hodgson, 2009, p. 136). In fact, researchers are always dependent, to some degree, on cooperation from people in the field in order to obtain data: gatekeepers can refuse access to sites that a researcher wishes to observe or prevent informants giving interviews; participants may refuse to be observed or interviewed and could even actively obstruct the research. These problems are particularly likely in research with participants from high-status groups or elites (Neal & Mclaughlin, 2009), though they may arise in other situations too (Kim, 2023).

It is perhaps also worth pointing out that vulnerability is not an issue that relates only to research participants, it applies to researchers as well. This is most obvious in “dangerous fieldwork”, and there is a small literature dealing with this (see Lee, 1995; Nordstrom and Robben, 1995; Behar, 1997; Lyng, 1998; Lee-Treweek and Linkogle, 2000; Downey et al., 2007; Bloor et al., 2010; Luxardo et al., 2011; Chevalier, 2015; Laar, 2014; Sampson, 2019). However, threats of harm of various kinds can arise in any research.

In the remainder of this paper, I will illustrate the complexities involved in dealing with the vulnerabilities of participants, and of researchers, from my own research on the making of national education reforms under conditions of structural adjustment and severe austerity in Greece (Traianou, 2021 see also Traianou, 2023). I will focus on the strategies and tactics I employed to negotiate my research relationship with members of a political elite, and how my concern about respects in which they may be vulnerable arose.

The research project: studying a Greek political elite

During 2015–18 the Leftist Syriza-led Government of Greece worked closely with the OECD in the production of two policy reviews for education (OECD, 2017, 2018). The work was carried out as part of the reform package of the third “structural adjustment programme” (SAP) which Greece agreed with its external creditors (IMF, European Central Bank, European Commission – officially known as “the Institutions”) in August 2015. In relation to education, the OECD, asked for the post-2010 New Public Management (NPM) reforms to be implemented (MoU, 2015, p. 22), entered a lengthy engagement with the Greek Government, in several phases across a period of two and a half years. As I have argued elsewhere (Traianou, 2021), the third SAP was more controversial (in terms of its implementation) and more complex (in terms of its negotiation) than the previous two programmes. For one thing, despite Greece’s policy change after 2015 to engage constructively with the Institutions and the OECD in the implementation of the third memorandum of understanding (MoU), this did not resolve all their differences. Unlike “crisis countries” like Portugal (Carvalho & Costa, 2017) and Spain (Verger & Curran, 2014), where the OECD was an invited adviser, Greece had not requested the organisation’s involvement, indeed the Government continued to have strong reservations about the core policy framework it was promoting, relating to school autonomy and accountability. My project focused on the ways in which the composition of each of the two reviews was shaped by the interplay of national and external actors with diverse political orientations, and the effects of this process on national policy. I argued that the two reviews helped significantly to modify the education conditionalities of the third MoU, by delaying NPM reforms and by depoliticising teacher tenure. Indeed, the reports helped to unlock EU funding for the making of teacher permanent appointments (Traianou, 2021).

I employed a historical case study methodology, treating Greece as a “bounded system” (Stake, 1995, p. 47): as a case with distinct characteristics compared with other EU and OECD member states as they navigated post-2008 crises (Traianou, 2021). I worked from what were acknowledged to be key documents in the educational policy history of the 2010–2019 period. I also kept a systematic record of media coverage across the political spectrum and of sectoral websites during 2015–2019. Predominantly, though, the project relied on narrative interviews with seven policy actors who were part of a particular “elite” policy network, responsible for direct negotiations with the OECD and the Institutions. This network included leading politicians, academics, and teachers. Its members were explicitly affiliated to Syriza, the governing party at the time. Some of these actors had been seconded from their permanent posts (academia, schools) to work on aspects of education policy and for this purpose they held key positions in the two organisations responsible for education policy in Greece – the Ministry of Education (MoE) and the Institute of Education Policy (IEP). All seven of my (male) informants had played significant roles in the negotiations with the OECD and the Institutions and had contributed to the production of the two reports (OECD, 2017, 2018). Most of them had interconnected public biographies and overlapping social connections, as well as links with other national and international networks.

Several of these actors were involved in organising a public forum/dialogue for education and in holding press conferences where they presented the progress of the negotiations. In other words, they were “visible” and seemingly easy to contact (Alvesalo-Kuusi & Whyte, 2018). But, as I discovered, they also placed high value on “privacy and exclusion” (Alvesalo-Kuusi & Whyte, 2018). Indeed, gaining in-depth data about a highly significant process of crisis negotiation between a national government and international organisations proved far from easy.

I also realised that my own positionality – my identity as a Greek national who works abroad, my political orientation, my relationship to some of the informants, and my academic affiliations – created complex relationships. To some extent I could operate as an insider researcher, someone who shared “an identity, language, and experiential base with the study participants” (Dwyer & Buckle, 2009, p. 54). This enabled me to delve into the tense social-political climate at the time of the fieldwork. But being a member of a group “does not denote complete sameness within that group” (ibid. see also Adler & Adler, 1987). I was not a member of the policy network I was studying. Indeed, having worked and lived abroad for more than twenty years I was an outsider researcher (Mannay, 2010) to the Greek political context more generally, and was no doubt seen that way by the participants. However, I had met two of my informants before the 2015 elections in research contexts, and I believed that I was regarded as someone who shared similar political ideas – though I was sceptical and quite critical of the post-2015 Government’s actions and policy direction. But, in any case, as Howard Becker (1967) pointed out long ago, however hard researchers may try to be impartial, they will often be seen as being on one side or the other. I hoped that my insider-outsider status would facilitate the research.

The interviews took place in a number of different settings: in cafes and bistros, but also sometimes in the policy actors’ offices. They were conducted between April 2019 and November 2021 in a schedule that was to some extent disrupted by the pandemic, as well as fluctuations of national politics. Four of the interviews were conducted when the informants were still in their official capacities. The rest of the interviews were conducted after the 2019 general election, when the informants had left office. Each interview lasted between 45 minutes to an hour. The interviews focused on the actors’ interpretations, tactical decisions and evaluations in the course of their encounters with global organisations and other national actors and policy communities.

In the next section I discuss how relationships were negotiated throughout the research process: from gaining access, through collecting and analysing interview data, to publishing the findings.

Navigating between strategy and ethics in the research process

As already noted, I was aware from the start that it was not going to be easy to gain access to elites; and that, once access had been obtained, it would not be simple to get my questions answered (Walford, 2012). Social relations in the field are sometimes presented as a matter of finding strategies to build rapport, in other words to encourage people to provide the data required for the research project (Ostrander, 1993). Any reluctance or resistance is treated as a barrier to overcome (Harvey, 2011). In my dealings with members of a Greek political elite, I was aware that some of them sought to limit and manage my access to data, and I recognised that I needed to find ways to build trust with them so that they would provide entry to more of the data. However, at the same time, I was aware that the blocking tactics and reluctance on the part of my informants were not simply motivated by a desire to promote their own interests, they were also trying to protect themselves from reputational damage. Moreover, this was a reasonable concern on their part, and one that I ought to respect. After all, they had been involved in a high-stakes game. The contested nature of the OECD’s agenda was reflected in the media as well as in Parliament where, for both the Right and the Left, the OECD was viewed as the vehicle for the third MoU package. Trade Unions and most of the political Left were critical of government strategy. In their narratives, my informants expressed their disappointment at the lack of support from the political Left, about the hostility of the Institutions and of some members of the OECD teams, but they also emphasised the political significance of the two OECD reports in achieving modifications to the conditionalities of the third MoU. It was clear that, in this contested context, their actions during the negotiations had implications for their reputational standing with significant others that may well affect their future careers. In short, I concluded that there were ways in which they were vulnerable and should be treated as such.

More than this, members of elites usually have insider information about events and processes that have wide implications for people within their societies. They may feel that some or all of this should be kept secret (or that powerful others would expect this), even though they understand that there are grounds for a public right to know, or that research on these events and processes could be valuable. Certainly, members of political elites often see little personal gain in cooperating with researchers. On the other side, researchers studying elites, very often, aim to get informants to divulge information that they might otherwise have kept back, using strategies similar to those long practiced by investigative journalists (Dexter, 2006). But these strategies raise ethical questions about security issues: should all aspects of intergovernmental negotiations be made public? (see Shils, 1956). As I noted at the start, members of elites have often been treated as “fair game” for researchers to deploy such strategies. In this context, it has sometimes been argued that the usual ethical requirements, such as obtaining informed consent, must be suspended, on the grounds that members of “elites” do not need protection, and indeed that their activities can legitimately be exposed to publicity (Galliher, 1980; Gaztambide-; Fernández, 2015). However, this needs to be reconsidered. In my own research I found that I had to tread a delicate line between pursuing the research as effectively as possible and recognising ethical responsibilities to my informants.

In all my initial contacts with potential interviewees I used my professional credentials, standing, affiliations and personal connections for gaining access. To some extent I relied on contacts who were peripheral to the policy process but whom I could nevertheless mention in my initial contacts with potential interviewees. As Cassell (1988, p. 85) points out, “everyone who might possibly know someone, must be contacted and asked if they will give introductions, vouch for one, and otherwise help one’s enterprise” (see also Winkler, 1987). I tried to present my research topic as “harmless” (Ozga, et al. 1994) by explaining that I was interested in understanding the Europeanisation of education during the post-2015 period. This often produced oral consent to be interviewed, but I was aware that a degree of deception was built into the process (Wiles, 2003). Moreover, despite often warm initial expressions of willingness to meet me, potential interviewees were frequently slow to agree a time and place. They generally wanted to know more about me and my research before they agreed to be interviewed. Often, I would get no further replies to our initial email exchange. Telephone arrangements proved to be more effective. I soon realised that I had to keep on reminding potential informants (by email or phone) that I was “around” and available to conduct the interview at short notice – one of the interviews took place on a Sunday morning, another was conducted on a public holiday. There were indeed times when I felt I was treating potential interviewees as “quarry” to be pursued by any means possible (Dexter, 2006, p. 35), not a comfortable position to be in, but it was hard to see how this could be avoided if access was to be achieved.

I was acutely aware in organising appointments and conducting interviews that “gaining access is not the same as establishing the trust required for getting useful data” (Ostrander, 1993, p. 135; Hammersley and Atkinson, 2019; 2019, Chapter 4). Furthermore, some relationships required careful management both before and after the interview, and within the interview encounter itself. For instance, much of the time, I did not ask challenging questions regarding their perceived “compliance” with the demands of supranational organisations, as implied by some of the press and by the political Left. I presented myself as the “ignorant” outsider willing to learn from knowledgeable insiders. Dodge and Geis (2006, p. 80), in the context of their research with white collar prisoners, talk about a “‘babe in the woods’ approach: pretending ignorance and inviting the informant to ‘teach’ you, as only he or she would be able to do””. My positionality as a Greek national who works abroad, and probably my gender (see Gewirtz & Ozga, 1994), made it relatively easy for me to adopt this approach.

This sort of approach is necessarily highly contingent in what it produces. On one occasion a key informant had agreed for the interview to take place during a break of a conference he was organising. A common contact had put us in touch who I believed had given me a positive recommendation (Winkler, 1987). But the informant had not specified the date of the interview “it’s a very busy period but your research is important. Come to my conference if you’re around this week”, he said. I decided to turn up on the first day of the conference to introduce myself. I also made sure that the interviewee was aware that I was talking about the broad focus of my research to one of the keynote speakers that I knew, and he then introduced me to the informant, who was about to leave the room. The keynote speaker commented to me, in the company of the informant: “I hear you have planned to talk about important things”.

Like other researchers, in employing such tactics I tried to use my academic post to reinforce the image of someone who is established not only internationally but also nationally (see Jober, 2021; Lillie & Ayling, 2020), fearing that I would not be taken seriously by a male informant and a well-established Greek academic. Despite cordial introductions I was taken aback when, as we were approaching the location of the interview – a near-by bistro chosen by the interviewee – he asked me to remind him of my name, what I do and the name of the person who put me in touch with him. Despite this, the informant seemed to speak openly and with confidence about his role in the negotiations: “please record the conversation. I’m a sociologist and therefore familiar with the research process”. But as the bistro was full of people known to the informant, including conference delegates who acknowledged him respectfully as a public figure and academic, I was careful in asking questions that might “put him on the spot”. During the interview, I often felt that the informant was taking control of the research relationship to protect himself (see Dodge & Geis, 2006), and that I was complicit in this.

Some of the interviewees, especially those who occupied a lower position in the hierarchy of the policy network, were very cautious about releasing information: “the memory escapes me” said one of them when I asked him to say more about their meetings with external organisations. Despite our pre 2015 research relationship, this informant got very nervous when I asked if I could record the conversation: “people used to take notes now they record conversations and upload them on social media”, he commented, before he, very reluctantly, agreed to be recorded. On this occasion, I made a conscious decision not to “play-up” my identity as a senior academic. Instead, I presented myself as also wanting to discuss our common research interests about teachers’ working conditions. In this way, I managed to [re]establish trust, which allowed me to pursue my interview agenda. Such strategic use of researcher positionality has been employed by many social scientists researching elites, as a way of gaining access to them, and then building trust with them (Gewirtz & Ozga, 1994; Grek, 2021; Herod, 1999). Unlike the bistro interview, it proved much easier eventually to exercise some pressure for further information from this informant because of our previously existing relationship. But I was aware that doing this raised ethical questions.

After the 2019 general elections, it became easier to arrange interviews with key informants, since they were no longer in post. As I became known and trusted among them, I found that some were interested in knowing who else I had spoken to (Hunter, 1995; Lancaster, 2017) and often made suggestions about whom I should speak to. I was open about who I had already talked with, and this often helped to build further rapport; though I never revealed what they had said, respecting their confidentiality. It was through this route that I eventually managed to get access to a person who had been particularly significant in the negotiations. However, during the interview it was clear that he had prepared what he was going to tell me; and, when my allocated time was coming to an end, he started looking at his watch to warn me that my time was up. At the same time, he showed me an important state document which was not in the public domain, as evidence of the ways in which the national government managed the external pressure to implement reforms. He was initially reluctant to give me a copy: “we’ll see”, he said, “I would not want this document to be quoted in publications”. He did subsequently send it to me; perhaps he needed time to make a judgement about my credibility and political affiliations. I did not quote from the document, though I learned much from it (see also Milana, 2021).

So, while some of my informants were clearly conscious of the need to protect themselves, I also felt a responsibility to protect them. But there was a difficult balance to be struck between this and employing strategic action to gain access to the data I needed. Given the highly politicised context, I treated them as vulnerable since they could be identifiable as sources of information or as participants in the processes of international negotiation I was documenting. This was especially true given the relatively small number of people who were directly involved in these events. I was careful in selecting data for inclusion in published articles, using publicly available documents as cited sources wherever possible. This was partly because of my fear that anonymisation might not be successful (Hammersley & Traianou, 2012 , Chapter 5). Many researchers have discussed their dilemmas about how to balance their commitment to knowledge production with potentially exposing respondents’ identities, choosing to withhold information to reduce the risk of harm to informants (see Baez, 2002). I found that resolving such dilemmas was not easy. One interviewee revealed a particular incident which involved several members of both negotiating teams. The incident, according to the informant, had resulted in a significant change in the content of the final OECD report (OECD, 2018). The informant said to me: “Write about it, very few people know about what happened just before the publication of the final report”. In the end, I decided that the disclosure of this information could potentially jeopardise the delicately balanced politicised policy processes that were still underway.

I also worried that my analysis could be misinterpreted by the policy actors I had interviewed, or by those in other policy networks, the media, and other political parties. I decided to offer informants the opportunity to send me feedback on my analysis, a technique for establishing credibility that some researchers have called “member checking” (Birt, Scott, Cavers, Campbell, & Walter, 2016; Lincoln & Guba, 1985, p. 314). I did this initially by inviting some of them to a research seminar where I was giving a paper on the study. Not everyone attended or responded to my invitation. Nevertheless, contact was made (via zoom and email) a few months later by two participants. One of them emailed me a draft chapter he had written about the trajectory of Greek education policy. Later, I was contacted again by the same person who encouraged me to talk to him about what I was writing in order to clarify some aspects of policy. He said he was also happy to provide feedback on my draft paper. The day of the meeting, he sent me my draft article with a few corrections in red on the first couple of pages, including corrections to my written English. He did not disagree with my account, he told me, and he made suggestions about other people I could talk to that he could put me in touch with. At that point I decided there was a risk to my autonomy as a researcher (see Traianou & Hammersley, 2021) and did not encourage further communication. This was respected by the informant. Despite his promise, he did not send me any further comments. I did not make any significant changes to my analysis. A few months later I received an invitation to his book launch.

I tentatively concluded that he, and others, felt that they had been protected from unnecessary damage to their reputations. I had endeavoured to do this without compromising my analysis of the complex political events in which these informants had been involved. But difficult, and necessarily questionable, decisions had had to be made in this process.

Conclusion

In this paper, I have discussed the complexity of the concept of vulnerability, and suggested that designating some social categories as vulnerable and others as not vulnerable is unsatisfactory. I examined some of the problems surrounding this concept, noting the need to recognise differences in degrees and types of vulnerability. I also emphasised that there were limits to the responsibility of researchers to protect participants. But this is not to deny that the vulnerability of participants must be taken into account. Research on elites is a particularly interesting case from this point of view, since elite members are usually regarded as powerful, from which it is often concluded that they are not in need of protection. More than this, it is often assumed that researchers can use whatever strategies are necessary to encourage members of elites to divulge the inside information to which they have access.

I argued that a middle way is needed here if research is both to meet ethical standards and to be effective in gaining the data required to understand the role of elites and the consequences of their actions for policy and practice. I illustrated this through a discussion of my own research on members of a political elite in Greece. I showed the strategies it was necessary to employ in order to navigate contacts with potential informants, seeking to overcome the power imbalance between me and them, while at the same time recognising the need to respect their concerns and fears about divulging information. As many other researchers have noted, the research relationship is messy and untidy. It involves difficult encounters and continual self-reflection. I had not anticipated the informants’ degree of emotional involvement in the production of the two OECD reports. I empathised with their accounts while at the same time trying to preserve my autonomy as a researcher.

My paper contributes both to the literature on researching elites and to the field of research ethics, by exploring the concept of “vulnerability” and its practical implications. I argued that protecting elite participants from harm was an important requirement, even though this ethical commitment had to be balanced against my responsibility to document the exercise of power and its consequences for Greek society.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This article is part of a funded research project by the British Academy (BA/Leverhulme, SRG1819\191004) in partnership with the Department for Business, Energy and Industrial Strategy.

Notes on contributors

Anna Traianou

Anna Traianou is Professor of Educational Studies at Goldsmiths, University of London. Her research interests include the situated character of social research ethics and the role of social actors, including political parties and trade unions in education reform.

Notes

1. See, for instance: Research with potentially vulnerable people – UKRI. Van den Hoonaard (2018) reviews what a wide range of ethics regulatory bodies, in several countries, say about vulnerable groups and how they should be treated by researchers A very large number of social categories have been listed as vulnerable at one time or another, see Sieber (1992):93. For discussion of vulnerability in relation to prisoners, see Mitchelson (2017); psychiatric patients, see Bracken-Roche, Bell, and Racine (2016); and migrants see Maillet, Mountz, and Williams (2017). For a brief history of the origins of the concern with vulnerability and subsequent interpretations of the term, see Levine et al. (2004).

2. This arises from the fact that, as Alasdair MacIntyre has argued, we are “rational dependent animals” (MacIntyre, 2019). The connection between vulnerability and dependence has been explored by feminist philosophers, who have stressed the significance of emotions and embodiment: see, for instance Purcell (2013); Mackenzie et al., 2013; Mao (2019); and Polychroniou (2022). They have sometimes argued that “vulnerability” can be positive, implying responsiveness to others (Gilson, 2014).

3. For discussion of types of harm: Hammersley and Traianou (2012):ch3.