Critical Ethnography in National Security Institutions: Methodological and Ethical Reflections

Abstract

Although ethnographic methods have yielded important insights into the inner workings of many political institutions, their use in the study of institutions engaged in national security policy making and practice has been more limited. Common features of security institutions raise methodological and ethical concerns for participant observers: they are hard to access, handle classified information, and their proximity to the use of violence creates challenges for researchers who adopt critical perspectives toward the framing and practice of national security. This article reflects on how critical ethnographers might respond to these challenges, drawing on experiences from a 14-month ethnographically influenced study of UK government departments involved in national security policy making. I argue that even though questions of informed consent and avoiding harm are complicated by the core functions of security institutions and their cultures of busyness and secrecy, navigating these tensions can itself generate useful ethnographic insights into institutional cultures.

Ethnographic methods, in which researchers observe and immerse themselves in the everyday life of a community, can offer unparalleled insights into the cultures and practices of political institutions (Bayard de Volo and Schatz 2004; Schatz 2009). Yet ethnographic studies of national security policy making and practice are limited by the secretive character of security institutions (Pouliot 2008, 285). Researchers have nonetheless conducted ethnographic studies of militaries (Baker et al. 2016), ministries of foreign affairs and defense (Neumann 2012; Stepputat 2012), weapons laboratories (Gusterson 1996), and defense research centers (Cohn 2006). Common features of security institutions raise several methodological concerns for researchers: they are hard to access, handle sensitive (often classified) information, and hold institutional power that shapes the research relationship. Their participation in or proximity to the use of violence generates ethical dilemmas, particularly for researchers engaging in critical ethnography that seeks to challenge social injustices (Thomas 1993), who may critique the framing and practice of national security from an ethical perspective.

This article reflects on these methodological and ethical issues from the perspective of a feminist researcher informed by a broadly antimilitarist politics, drawing on my experiences studying cross-departmental teams of UK government officials working on national security policies. Through 182 hours of participant observation and 60 qualitative interviews with policy makers over 14 months, I examined how institutional cultures are gendered, racialized, and classed and how this shapes policy discussions. I describe the study as “ethnographically influenced,” rather than “an ethnography,” because the latter often, though not always, implies continuous immersion over long periods (Jeffrey and Troman 2004). Such immersion in security institutions is difficult for anyone but employees, who may be contractually forbidden from publishing observational research. Observing in short bursts over a longer period, as I did, may often be the only approach available for external researchers studying security institutions. Although this method makes it difficult to generate ethnographic insights by observing patterns emerging from everyday practice over time, it is consistent with the insight of feminists and “patchwork ethnographers” (Günel, Varma, and Watanabe 2020) that all knowledge is partial and the researcher must account for how this partiality shapes her findings (Haraway 1988).

Such immersion in security institutions is difficult for anyone but employees, who may be contractually forbidden from publishing observational research. Observing in short bursts over a longer period, as I did, may often be the only approach available for external researchers studying security institutions.

I retained an “ethnographic sensibility” (Schatz 2009, 5) through a focus on meaning-making in routine and seemingly mundane practices as I worked all day alongside officials in open-plan offices, observing their informal conversations and formal meetings. Like Neumann (2012) and Stepputat (2012), I gained access to government ministries by working for them, consulting on policy processes in exchange for research access. Institutions that value discretion may have little incentive to accommodate external researchers unless they are gaining something in return, raising methodological and ethical questions that I discuss later. I first address questions around informed consent and security-sensitive information before discussing the balancing of potential harms and benefits resulting from research and the potential for researcher complicity in harms enacted by security institutions.

SECURING INFORMED CONSENT

Gaining participants’ informed consent is an ethical imperative in all research with human subjects, but it gains additional significance in institutions that handle officially classified information. Researchers may require security clearance to access certain offices or meetings, including those in which sensitive information circulates freely, the publication of which could have negative consequences for the researcher, participants, or both. Making explicit agreements about what information cannot be published is therefore crucial. Participants in meetings I observed were informed in advance that they should not discuss any classified information. Although I cannot know exactly what insights I missed because of this limitation, not being invited to higher-level national security meetings meant that I observed few interactions among senior civil servants, which would no doubt have shed a different light on institutional cultures.

In some meetings I was asked not to publish details that were not classified but merely politically delicate, such as the names of states whose bilateral relations with the United Kingdom came up in conversation. To what extent such restrictions limit the analysis depends on the research questions: in my study, I could build a rich picture of how officials interacted with each other and spoke about policy issues without including such details. These exchanges about consent themselves yielded insights into what was or was not considered potentially sensitive or embarrassing. In one meeting with counterterrorism officials, for example, critiques of other states’ operational approaches were (unsurprisingly) considered too sensitive to include in my fieldnotes. Yet masculinist tough talk about “chasing bad guys”—which was more directly pertinent to my research—went uncensored, despite some officials privately expressing embarrassment about such language, thereby highlighting its relative acceptability in that setting. Of course, should security institutions’ assessments of what is politically sensitive overlap with researchers’ interests, such prohibitions could easily inhibit researchers’ ability to make necessary critiques.

Difficulties ensuring that consent is continuously given are inherent to participant observation and are exacerbated by the fast-paced environments common in security institutions: busy officials concentrating on the job at hand easily forget they are being observed. Because ethnographic research seeks to study communities’ everyday practices, the researcher always hopes that participants will, to some extent, let their guard down in this way (Bourgois 1990, 52). This may happen more easily when the researcher actively contributes to the institution’s work, which may require participants to share with the researcher information not intended for public consumption. Consequently, participant observers are urged to continually renegotiate consent throughout the project (DeWalt and DeWalt 2011, 216). Yet I found that the culture of tight deadlines and quickfire conversation made most participants reluctant to continue such negotiations indefinitely. In one meeting where I introduced my research project to a group of participants, I was asked instead to exercise “common sense” over what to include in or exclude from my fieldnotes. Where agreements on consent were similarly unspecific, I decided what data to include based on judgements about the potential harm to participants, while still making explicit agreements concerning any potentially security-sensitive information.

In addition to challenges in ensuring that consent is continuously given, the busy lives of security officials can make it difficult to ensure that consent is “fully” informed. I explained my research to officials verbally and through an information sheet; however, many participants did not want to listen to or read detailed explanations. Of course, any explanation of a research project is inevitably partial, and the researcher cannot control how much participants choose to take in (Fine 1993, 274–77; Horsley, Gillies, and Edwards 2017, 110), such that no clear line exists between “informed” and “uninformed” consent (Lugosi 2006; Thorne 1980). Acknowledging that consent is never fully informed and that this was not entirely within my control, I made information about the project as accessible as possible in the hope of encouraging participants to absorb it; for example, by sharing a blog post written in more engaging language explaining what the research was about. Again, where gray areas arose in relation to informed consent, I turned to the imperative to minimize risks of harm, which I discuss next.

HARMS AND BENEFITS: A QUESTION OF POWER

Researchers have responsibilities to a range of actors affected by their research, including participants, funders, their own institutions, and those whom the research seeks to benefit. Critical researchers attentive to the harms caused by security policies and practices may view these responsibilities as competing at times, because what benefits security institutions may harm populations who are subjected to state violence, and vice versa. Standard ethical guidelines tend to prioritize responsibilities toward participants: the American Anthropological Association (2012), for example, states that ethical obligations to research participants are “usually primary” while acknowledging that “obligations to vulnerable populations are particularly important.” In critical research that explicitly seeks to challenge power structures, weighing these competing claims requires an analysis of power relations.

Critical researchers attentive to the harms caused by security policies and practices may view these responsibilities as competing at times, because what benefits security institutions may harm populations who are subjected to state violence, and vice versa.

In conducting this analysis, I found invaluable Hodgson’s (1999) three-step process for balancing competing ethical responsibilities: (1) identify the groups to which one is accountable and analyze the “historical and contemporary political, economic, and socio-cultural power relations among them” (213); (2) assess the positive and negative consequences to each group that may result from the research; and (3) “determine whether support for the most oppressed group (in a given time and place) is worth the consequences for the other (perhaps differently oppressed) groups” (214).

In my study, participants held positions of relative power as (mostly) British citizens in secure, well-paid employment, with the power to influence government policies. The research was ultimately intended to benefit people with less power: those experiencing the negative impacts of UK policies, be they Yemeni civilians endangered by British armaments sold to the Saudi military or British Muslims surveilled under counterterrorism laws, for example. Although I held ethical responsibilities to minimize the risks of harm to both groups, given the unequal power relations between them, at times I judged that my ethical responsibilities toward the latter group outweighed some minor potential harms to my participants. Of course, those affected by security policies are not a homogeneous group sharing common interests that are easily translated into a clear set of ethical responsibilities toward them. Claims to speak for marginalized people can function to silence or misrepresent them (Mohanty 1984; Spivak 1988), and British antimilitarists have sometimes caricatured or ignored the views of those most affected by British militarism (Al Shami 2018; Syria Solidarity Movement UK 2015). Yet holding positions of power—including the power to represent others through research—confers a responsibility to advocate for justice for the less powerful while making the decisions behind this process transparent so that researchers can be held accountable for them (Alcoff 1991).

The most significant potential harm to participants came from the possibility that publishing their views could damage their professional relationships or hinder their prospects for promotion, risks I mitigated by protecting their anonymity by using pseudonyms and vague descriptors. An arguably less injurious but more likely possibility was that the research would produce critiques of participants’ work with which they did not agree, which some may interpret as a form of harm (see Mosse 2011). In deciding to what extent to discuss the normative commitments that informed my research with participants, I followed Hodgson’s principles. Like many researchers (e.g., Bourgois 1990; Gray 2016; Wall 2011), I rarely discussed my political leanings with participants, though a quick Google search would have revealed them. This influenced the language I chose to use in conversation with officials: for instance, I only rarely used the term “militarism,” understood as the normalization and legitimization of the use of organized political violence. I did so partly because it was not part of the policy lexicon and some officials found it unclear; however, it is also a term typically used only by antimilitarists. Although it is common for researchers not to disclose their political beliefs, this practice raises questions around participants’ right to know the details of the kind of study they are participating in (Fine 1993, 274–77). Like Wall (2011) in his ethnography of militarism in his hometown, I determined that any minor “harm” caused to powerful institutions by being subject to political critique was outweighed by greater harms to those bearing the brunt of militaristic policies, which my study sought to contribute in some small way to addressing. Furthermore, although the imperative to protect individual participants from public critique was paramount (e.g., by anonymizing them), few would argue that government departments or policies should be shielded from criticism in the same way.

CO-OPTATION AND COMPLICITY

Scholars of critical military studies have raised concerns about what Enloe (2010, 1107) calls the “risks of scholarly militarization”: the danger that, when studying militaries, the critical researcher becomes complicit in the very harms she seeks to critique. For example, in an attempt to maintain access or be taken seriously by participants, the researcher may engage in or decline to challenge militaristic practices and ways of thinking, a concern that applies across security institutions more broadly. Indeed, research in national security institutions typically has characteristics that create pressures to become complicit in harmful practices: the institutional setting gives participants a degree of power over the researcher, whose access would be easy to revoke and difficult to regain, and serious forms of violence are often central to the organizations’ work, rendering critique more sensitive (Becker and Aiello 2013, 67).

Research in national security institutions typically has characteristics that create pressures to become complicit in harmful practices: the institutional setting gives participants a degree of power over the researcher, whose access would be easy to revoke and difficult to regain, and serious forms of violence are often central to the organizations’ work, rendering critique more sensitive.

In meetings, informal chats, and interviews, some participants expressed militaristic or colonial attitudes that (in my view) contribute to real-world harms. Yet when research access is precarious, it is easiest to keep quiet, lest challenging harmful views cause participants to withdraw their participation or become guarded (Horsley, Gillies, and Edwards 2017; Neal 1995, 528). Because a belief in the legitimacy of militarism is core to what many security policy makers do, to oppose that belief would be a serious condemnation. In some instances, I gently challenged such views, but often I chose not to. I even adapted my own language to avoid causing discomfort: for example, after one participant suggested my use of the term “arms trade” could be read as hostile, I started referring to “defense exports” instead. The stressful experience of this pressure to conform sometimes prevented me from asking sensitive questions; yet, reflecting on my own affective responses also afforded insights into how institutional norms are maintained (Cohn 2006, 106; Shesterinina 2019).

Becoming an active participant in policy-making processes raised further ethical challenges: although there were aspects of these processes I could support, others caused me to continually question whether my participation was justified. Having chosen to study security policy discussions because they can legitimate militaristic and colonial practices, I made myself complicit in that process. This dynamic was also shaped by my positionality: my whiteness and British citizenship likely helped facilitate my access while also protecting me from the worst effects of the policies discussed (e.g., those relating to counterterrorism). In the interests of preventing harm against the most marginalized, I committed to bringing in critical perspectives wherever I could; for example, by inserting wording into draft policy documents reasserting human rights commitments or highlighting the racialized or gendered harms of particular security practices. Although some of this wording stayed, it was often removed, prompting conversations with officials that provided new data about unwritten rules governing policy making and how they are reproduced. Deliberately testing the limits of acceptability in this way therefore became a methodological tool for studying institutional culture, although as an external researcher the rules applied to me were not necessarily the same as those applied to the officials themselves; I accounted for those differences when interpreting the data. From an ethical perspective, however, these conversations and minor changes to the wording of policy documents likely made little difference to security practices, and it is debatable whether they outweighed the potential for my participation to legitimize anti-emancipatory policy regimes (Jackson 2016). Furthermore, the freedom to persist with this critique-as-method approach was perhaps more available to me as a researcher offering consultancy in-kind than to someone paid or employed by the institution; although being on the payroll would grant deeper access, it also comes with its own ethical concerns (Mampilly 2020).

Compromises made to maintain access represent a trade-off between complicity in the moment and the pursuit of longer-term research objectives. As Martín de Almagro (in Holmes et al. 2019, 224) puts it, “Instead of challenging the power dynamics at the time, you prioritize critical feminist research outcomes that will help you … uncover and challenge the broader power dynamics.” These moments of acquiescence helped me build rapport with some participants, enabling me to ask more challenging questions later, which generated useful data and encouraged them to consider more critical perspectives. Whether such trade-offs are worthwhile is unknowable in advance: researchers cannot predict with certainty the impacts of their research nor the severity of harms to which they may contribute. These factors can be challenging to ascertain even retrospectively: the lack of transparency surrounding security institutions makes it difficult to know how the policy documents I helped produce have since been used. Furthermore, just as researchers can use participants’ words to advance political agendas with which the latter may not agree, so participants can do the same in return: researching any political institution, even from a critical perspective, risks the findings being used to further political aims that the researcher does not share (Jackson 2016). These ethical dilemmas may deter some critical researchers from engaging with security institutions, and perhaps rightly so. However, navigating these tensions can also reveal the institutional dynamics of what can or cannot be said and done, and how knowledge is produced and used, thereby advancing critical research agendas.

CONCLUSIONS

Some ethical dilemmas raised by ethnographic studies of security institutions resemble those encountered by ethnographic researchers in other political institutions, which are shaped by difficulties in gaining access, cultures of busyness, and negotiating with participants who wield institutional power. However, these dilemmas may be exaggerated in national security communities, which are often characterized by cultures of secrecy and a sense of exceptional urgency. Moreover, the circulation of security-sensitive information presents unique challenges, and the imbrication of security institutions with systems of violence and oppression creates distinct ethical tensions for critical ethnographers seeking to challenge those systems. In my study of institutional cultures in UK government departments that make national security policy, the tensions between the core beliefs and functions of these institutions and the critical perspectives that animated my research reflected the research problems I sought to address. They spoke to questions about how institutions develop and maintain belief systems and values, whether and how these beliefs and values can be challenged from within, and what are the costs of doing so for differently gendered, racialized, and classed individuals. Navigating ethical dilemmas therefore became a source of learning about institutional cultures and the experience of negotiating, submitting to, and resisting those cultures. The way through these dilemmas is rarely clear-cut. Consent is never fully informed and not always continuously and unambiguously given; what protects one group may harm another; and complicity with the militaristic ideas and practices predominant in security institutions can be both antithetical to and a precondition of critical research. Although the ethical trade-offs will differ for each research project and lead to varying conclusions, there is much to be gained by sitting with these ethical tensions and learning from the uncomfortable situations to which they give rise.