Recognizing the never quite absent: de facto usage,ethical issues,and applications of covert research in difficult research contexts

Abstract

This article scrutinizes and reconceptualizes covert research in social science. Surveying recent literature about this research method, we reflect on the ethical and safety implications stemming from the widespread, even if well-intended, lack of transparency characterizing many research projects, especially ones conducted in difficult research contexts. While the standard definition of covert research holds that researchers deliberately do not declare to research subjects that academic research is taking place, we argue that the remoteness of Western academia from most researched contexts often a priori renders field research at least partially covert, irrespective of the researcher’s intentions. This is because its aims, utility, and expected outputs are hard to understand for research participants unrelated to academia. We illustrate this argument by analyzing our own fieldwork experiences in Tajikistan and Kyrgyzstan. In conclusion, we emphasize the need to critically reflect on the de facto use of covert techniques by social researchers.

Introduction

The standard definition of covert research in social science textbooks indicates that covertness stems from the researcher’s deliberate decision to declare neither their identity as a researcher to research subjects nor that the research is taking place (e.g. Bryman, 2012: 433–436). This understanding supposes the researcher’s incognito engagement with individuals or a social group, and a simultaneous observation in an intentional and conscious manner. While this research method was largely criticized in the last few decades, especially on ethical grounds (Calvey, 2017: 39 ff.), clearly, the advantage of covert research is that it limits the problems of access to field sites and research participants. By ‘going covert’, researchers do not ask for any research permission in a chosen locality. Covertness also limits researchers’ visibility, because their status is not revealed on the ground. As Calvey (2017, 13) posits, ‘covert research is clearly not to everyone’s analytic taste but the commitment is to explore different and creative ways of constructing ethnographic narratives’. Against negative connotations and moral strings attached to the ‘covert’ label, Calvey (2017: 174) presents and compellingly substantiates a nuanced argument that covert methods have always been part of social science research and present a crucial, even if submerged and stigmatized, tradition of inquiry.

In this article, we further scrutinize covert research in social science. Building on Calvey’s work (2017), we argue that some degree of covertness is an intrinsic feature of all social science research involving human beings, although it might be more pronounced in difficult research contexts. Covertness is an outcome of dynamics in ‘the field’ and results from diverging perceptions of what constitutes research—and particularly transparent, overt research—on the part of the researcher, on the one hand, and research participants, on the other. Contrary to widely held opinions, we thus emphasize that even if researchers try to follow the ethical protocol and design for overt, transparent research, asymmetries in understanding and diverging perceptions can inevitably render research covert.

To substantiate this argument, we draw on recent debates and our own research experience. Recent conversations on academic freedom (Russo and Russo, 2017) and issues related to ethics and access to field sites (Glasius et al., 2018) reveal that field researchers (as well as journalists or civil society activists) increasingly face challenges when investigating topics revolving around conflict, security, and the nature of political regimes. These debates coincide and often overlap with discussions that have been emerging in the last few years with regard to research in settings described as ‘authoritarian’, ‘closed’, ‘restricted’, or ‘violent’ (e.g. Area, 2013; Social Science Quarterly, 2016; Glasius et al., 2018; Bliesemann de Guevara and Bøås, 2020). This literature has largely adopts a problem-solving approach toward issues regarding the safety of researchers and, although more rarely, research participants. It addresses questions about how to recognize red lines, assess and navigate risks in the field, store data in a secure way, and offer best practices. As we argue, in doing so, these contributions often advocate various forms and degrees of covert research as a way to mitigate potential risks in field research.

In our discussion about covert research, we provide examples from the field of Central Asian studies, which refers to the five post-Soviet states of Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, and Uzbekistan. Yet, the insights we offer are not limited to Central Asia or other non-Western spaces. Rather, as we show, they reflect the nature of fieldwork conducted within the framework of Western academia, and its disconnectedness from the realities in field sites both in the West and non-West. In practice, the dichotomy of covert versus overt research simply does not hold.

Our review of the application of the covert research lens in the case of Central Asian studies, including our own research, suggests that this method is de facto used on a large scale. This seems to be driven by demands stemming from the global political economy of knowledge production, in which Western academia is deeply embedded. This does not mean that we condemn covert research, given that some covert techniques are simply inevitable in field research. In line with recent anthropological debates on ‘informed consent’, an approach that presumably allows the researchers to provide full information to research participants and thus avoid covertness (Bell, 2014), we point to the limits of this idea in the field. In this light, we argue that researchers still need to reflect on the importance and implications of covert research all the more. Drawing on literature conceptualizing and applying covert research (Calvey, 2008, 2017; Spicker, 2011), we demonstrate the value of a discussion about covert techniques, both in the Central Asian context and in field-based social research at large.

Empirically, this contribution draws on our long-term fieldwork on interactions between local and international actors within the development aid system in Tajikistan (Kluczewska, 2019), and on community security and peacebuilding by national and international actors in Kyrgyzstan (Lottholz, 2018). Although we followed the ethical regulations of our institutions, defended our PhD theses, and published findings in academic outlets that were endorsed by our research participants, we could not help the feeling that we, at least partially, failed in our attempts to be overt in the sense of fully clarifying the aims, outcomes, and implications of our research to the people that feature in it. Our two reflective vignettes illustrate that our own framings inevitably encountered established local framings that were incompatible with them, and that even collaborative and participatory research designs could not help to fully overcome these gaps. The article reflects on the ethics and practical implications of covert research to foreground a more proactive and reflective approach to dealing with the split between the lifeworlds of local communities and policy-makers, on the one hand, and the world of Western academic knowledge production, on the other. We proceed with an introduction into the debate on covert research and its usefulness across disciplines, to then discuss its de facto usage and appeal in the Central Asian context and, finally, in our own research.

Covert research: a demonized, but crucial part of social inquiry

From unethical to unavoidable?

In contemporary social science in Western academia, covert research has largely been perceived as inappropriate, illegitimate, or unethical. One can hardly imagine research similar to Laud Humphreys’ (1970) The tearoom trade: impersonal sex in public places taking place today. In this ethnographic study, which gained prominence mostly in the light of criticism on ethical grounds (Lenza, 2004), Humphreys observed homosexual encounters between men in public places in the United States, known as ‘tea-rooming’ in gay slang. He further followed his research subjects to their homes, which led him to claim in his book that more than half of the men involved in homosexual activities were simultaneously in stable heterosexual relationships. Humphreys not only conducted his research without gaining consent from the men whom he had been observing during sexual encounters, but also on several occasions deliberately deceived them by presenting himself as a health service employee, which allowed him to interview and collect statistical information on his research subjects. Without going into details of Humphreys’ study, it is important to point out that his research represents a clear-cut, extreme case of covert research (Calvey, 2017: 64).

The criticism of approaches similar to Humphreys’ study contributed to a gradual rise in ethics review boards in Western research institutions. Their aim has been to review research designs and methods proposed by social scientists to protect research participants from potential harm. The concept of ‘informed consent’ became an ‘ethical panacea’ (Corrigan, 2003) to the problem of covertness in field research. Informed consent suggests that if researchers inform research participants about research and the latter approve it, research can be considered overt and thus ethical. However, as various authors have argued, covert and overt research should not be treated as opposites (Calvey, 2008; McKenzie, 2009). Covertness is not necessarily a planned strategy to deliberately deceive respondents. The messy and spontaneously evolving reality of research on the ground often complicates researchers’ initial intentions to remain overt. In his 2017 monograph, Calvey (2017) argues that some degree of covertness is necessary to make research viable and instead of trying to avoid it, it is important to constructively reflect on its implications. Given this significant and unavoidable role of covert techniques, field researchers should focus on maintaining situated, context-specific ethics (Calvey, 2008). Spicker (2011: 130) has even argued against linking covertness with deception, by stating that ‘undeclared, incremental learning and discovery is part of the normal process of academic learning and investigation’, although such covert elements should certainly be subject to an ethical reflection. Another important aspect is that sometimes the intent to reveal crime, violence, and exploitation might morally justify covert research, as demonstrated in Scheper-Hughes’s widely acclaimed ‘Undercover ethnography of the organs-trafficking underworld’ (Calvey, 2017: 91–92).

A nuanced approach seems most appropriate in light of the nature of fieldwork in social sciences, where, regardless of signed informed consent forms, the aim and nature of field research often remains hard to grasp for people unfamiliar with it (Cramer et al., 2015: 148-149). In this regard, some researchers have advocated a practical rather than procedural ethics during fieldwork (e.g. Corrigan, 2003: 788; Calvey 2008: 912–913; Heathershaw and Mullojonov, 2020). While agreeing that this is needed, our argument, however, is more fundamental. Field research can still be overt despite both situated and procedural ethics. This is because, as we argue below, research participants’ understandings of the meaning and implications of research vary widely.

Variations of covertness as a way to ensure safety

In parallel with a reassessment of covert research, the aforementioned debates on research in ‘authoritarian’, ‘closed’, or ‘violent’ contexts have been taking place in social sciences. These contributions have addressed research in various types of difficult settings ranging from developing and (post-)conflict countries in Africa and Asia (Glasius et al., 2018; Rivas and Browne, 2019) to transitional contexts in Eastern Europe and the former Soviet Union, including Central Asia (see Area, 2013; Social Science Quarterly, 2016). Even if not directly, they still point to the issue of covert versus overt research. The conclusion arising from these discussions is largely that, in theory, clarity and honesty should be provided about research topics and framings; yet this often becomes impossible in the field. Hence, researchers need to re-frame their research to make it seem more politically neutral and less intrusive on sensitive issues (Loyle, 2016: 930); devise ‘opening narratives’ that put interviewees ‘at ease’ in order to obtain information on potentially sensitive topics (Markowitz 2016, 903); and generally ‘be flexible’ about communication and evolving research designs (Malekzadeh, 2016: 864).

Particularly remarkable contributions in the array of advice are those of Markowitz (2016) and Gentile (2013). Markowitz (2016: 902) argues that the most feasible way to obtain possibly sensitive data and insights from high-level managers in regional and local administrations might be to push one’s way into their offices past their secretaries, either by making up good reasons or simply outsmarting them. However, while it is certainly understandable that there is no need to over-emphasize the uncomfortable aspects of one’s research, creating too much ‘ease’ about one’s research also seems to open up a gray zone between what the researchers are telling their participants they are interested in, and what they are really drawing from an encounter, and what will eventually be published. It appears unlikely that interviewees would still agree for their opinions to appear in the research output, or from the framing of Markowitz’s article, that they could imagine what the final output would be. In his contribution on the different forms of surveillance mechanisms of field researchers, Gentile (2013: 428) also discusses briefly various risk minimization tools pointing to (semi-)covertness, including ‘playing dumb’, that is, not revealing one’s actual understanding of the local politics and society. ‘Playing dumb’ would thus allow researchers to continue research and gather information while avoiding unnecessary attention or causing suspicion.

The ‘bending’ and flexibilization of research aims vis-a-vis research subjects is an example of, as we argue, the actual application of covert research techniques. While these may indeed be pervasive, but also necessary to enable research in the first place (see McKenzie, 2009; Calvey, 2008: 909), proposals for their usage should also include reflection on their ethical implications. In particular, the above arguments point to the necessary realization that researchers relying on such techniques should be aware of the impact of their research on their interlocutors not only during fieldwork but also after it and, in particular, after the publication of findings (Loyle, 2016: 934; Sheely, 2016: 945). Furthermore, concerns about an extractive and one-directional nature of field research have spurred some researchers to embrace more dialogical and cooperative forms of knowledge production. In such collaborative projects, researchers give up their control and authority to fully determine the research design and questions, as well as to interpret and frame research findings (see Hale, 2008). This way, some degree of overtness can be achieved. Below, we describe our own attempts to undertake collaborative research and the way in which aspects of covertness persisted, nevertheless.

Commonly acceptable covert research techniques

As illustrated above, there are some approaches to research that, sometimes subconsciously, make use of covertness and which remain widely accepted in academia today. To synthesize the debate, we identify three such covert research techniques. The first technique is participant observation, which remains a popular method not only among social anthropologists but is also increasingly used in other social science disciplines, such as geography, political science, and sociology. Since the pioneering work of Malinowski, Evans-Pritchard, and Mead, anthropology has, perhaps more than other disciplines, extensively elaborated on the practice and ethics of participant observation. These debates (e.g. Hume and Mulcock, 2004) have pointed to an important distinction between doing participant observation in public, which is generally not considered problematic, and in private spaces, which requires approval for access and imply ethical obligations toward informants. While agreeing on this differentiation, our argument is situated in the gray zone between these two modalities. We argue that participant observation is by default at least partially covert, as researchers cannot realistically disclose their status in all settings in which they conduct observations—for example, when sitting in a park and observing people, or by participating in seminars or conferences to analyze statements, narratives, and behaviors. In his comprehensive survey of studies that have used participant observation, Calvey (2017: 158 ff.) also lists cyber ethnography, which involves the observation of people’s behavior in virtual spaces like forums, chats, or social media groups; and insider or autoethnographies (ibid: 151 ff.), where people write first-person accounts of their work in various institutional settings. While researchers usually cannot afford to inform people whom they observe about this activity nor gain consent for doing so, the key argument in favor of participant observation is that it does little or no harm, as long as anonymization practices are followed. Furthermore, technically it does not present an infringement on personal space, given the often-public nature of, for example, digital content (Calvey, 2017: 59). Nevertheless, these aspects point to the need for reflection on the fact that people are often not aware of their unwitting role as research subjects.

The second commonly acceptable practice which resembles a covert research technique, concerns the collaboration with gatekeepers when conducting research within organizations, groups, and networks. The implicit assumption is that if a leader or gatekeeper approves such research cooperation, then other members of the given entity will be more likely to follow suit and also participate. This results in selective and unclear patterns of consent, and possibly rejection. While some research participants are aware of ongoing research, others may not have heard about it or fully understood its purpose. For example, Bryman’s (2012: 436) discussion of Mattley’s (2006) work on callers on a sex phone line shows that the line manager knew that research was being undertaken and allowed it, while callers were not informed about it, which seems problematic both from a research ethics and informed consent perspective. Some authors have shown that researchers might still have to engage with each group member to gain trust and secure participation, and that it is crucial to understand this difference between ‘physical’ and ‘social’ access (McKenzie, 2009: 5.2).

The third covert research technique refers to the abovementioned ‘flexible’ approach of framing and communicating the goals and outputs of one’s research depending on the audience. As argued by Bekmurzaev et al. (2018), framing one’s research in a particular manner is important to secure and maintain access to the people and sites in question. The fact that this often implies disclosure of only selective information on a given research project vis-a-vis research participants requires critical reflection about the possible implications of this information asymmetry. As Wilkinson (2015: 399–401) argues, writing a certain framing or terminology ‘out’ of a research project in order to more effectively analyze its unsolicited occurrences and evocations, before ‘writing’ it ‘back in’ in the post-fieldwork stage, should be seen as a legitimate tactic. As we show below, engaging in such information arbitrage may sometimes be the only way to conduct research, both in Central Asia and beyond.

Current knowledge production demands and covert research in Central Asian studies

Western and western-centric knowledge production pushing into covertness

As outlined above, rather than condemning research techniques relying on degrees of covertness as unethical, we argue that there is a need to reflect on their importance and necessity in relation to issues of access and safety faced by both researchers and research participants. Furthermore, such considerations need to account for the global political economy of knowledge production and the dominant role of Western institutions therein. The subordination of research to Western academic markets impacts on the choice of research methodologies and framings and topics. As Bliesemann de Guevara and Kostic (2017: 11) argue, the neoliberalization of knowledge production at Western universities puts pressure on academics, particularly junior ones, to deliver outputs that can compete in the ‘marketplace of ideas’, rather than producing critical knowledge (see Lottholz, 2017). Research grants are awarded to compelling applications, which are time-intensive to prepare and limit the time allotted to empirical fieldwork, so that the contextual immersion necessary to navigate safety and access issues often remains impossible. The simultaneous pressure to deliver fieldwork-based findings and publish in highly rated academic journals, defend one’s work vis-a-vis peer-reviewers, and to maintain an ideally constant flow of new publications with the aim of securing employment affect researchers at various career stages.

To face these challenges, researchers often feel pushed to settle for ‘groundbreaking’, and not rarely ‘sensitive’ research topics, adopting certain angles and data collection techniques to obtain field data. In the specific case of Central Asian studies, Bekmurzaev et al. (2018) observe a tendency to use framings that are ‘politically provocative, if not exerting epistemic violence upon research subjects’ (see also Reeves, 2005, 2014). This might be caused by a scarcity of research funding allocated for research on this region by Western funding bodies and universities. Following the Soviet Union’s collapse in 1991, Central Asia received attention from foreign scholars because of its apparent opening and ‘transition’ from socialism and to democracy and the free market. As a result, until the mid-2000s, the democratization angle dominated Central Asian studies. In recent years, perhaps because of a discontent with the outcomes of the transition, the dominant framings have moved towards conflicts, the rise of authoritarianism, and Islamist radicalization. Although the framing of Central Asia as an unstable region and hotbed of extremism have been challenged on different occasions (e.g. Heathershaw and Megoran, 2011), many research grants continue to be allocated for topics related to violence, political regimes, and corruption, which usually imply criticism of national authorities and black-boxing of states.

Researchers thus find themselves having to satisfy the need for funding, obtaining rich empirical data and successful publishing in this political marketplace (Bliesemann de Guevara and Kostic, 2017: 7). But reproducing dominant Western-centric narratives also exhibits them to the aforementioned problems of access and safety. Furthermore, given that Western universities are generally preoccupied with managing or minimizing risk during fieldwork through standard insurance frameworks (see Bekmurzaev et al., 2018), but are rarely actively helping researchers to reflect on and appropriately prepare for their fieldwork, the latter are often forced to navigate fieldwork challenges on their own and might feel pressured to make use of covert techniques. Below we demonstrate how these constraints have materialized in Central Asian studies.

De facto usage and appeal of covert research in Central Asia

As in other ‘developing’ and conflict-affected areas, researching social issues in Central Asia on the ground can be a challenge, particularly for outsiders, as rumors about outside forces instigating conflict and destabilization are still in high circulation. Such alertness is due to recent political upheaval such as the color revolutions of the 2000s and violent clashes in Andijan, Uzbekistan in 2005, Osh and southern Kyrgyzstan in 2010, or Zhanaozen, Kazakhstan in 2011, which are surrounded by conspiratorial narratives on the role of Western and other external actors. These are, however, part of a broader pattern of mistrust dating back to Soviet times. Particularly foreign and Western-affiliated researchers face allegations (jokingly or sincere ones) that they might be spies sent by their governments (Lottholz, 2017). Given this general climate of mistrust or at least discomfort with being interviewed by researchers from abroad or just having them around, any researcher, including those trying to conduct context-sensitive research, is more or less forced to transcend the clear-cut boundaries of overt research.

A valuable example of the three discussed techniques—that is, the use of participant observation, reliance on gatekeepers, and (re-)framing of one’s research—can be found in Trevisani’s (2010) study on Land and power in Khorezm in Uzbekistan. In its introduction, Trevisani discusses how the general context of the country would not have allowed him to openly carry out research on the relation between the state (its territorial subdivisions and agricultural agencies) and farmers, different groups of laborers, and the wider population. In this light, Trevisani (2010: 16–17) made the ‘considered choice’ to affiliate himself with a German-Uzbek cooperation project, the aim of which was to transfer agricultural and technological knowledge and support the modernization of production in Uzbekistan. It turned out that being able to ‘play the project card’ and becoming ‘a colleague’ of various administrators and staff of the local Farmer and Peasant Association was vital for gathering data, as he had soon realized the ‘climate of fear and mistrust’ among agricultural workers in the area, who tried to evade his interview requests, gave only standard answers and generally hid behind a ‘wall of indifference’ (2010: 18). The collegial and informal ties that Trevisani was able to build under the umbrella of technical cooperation enabled him to access people’s everyday life and what he calls the ‘peasant world’, which was dominated by informal flows of information, with ideas and decisions on land reform being shaped at weddings and other private celebrations (2010: 20–21).

In this sense, it can be argued that Trevisani has employed all three covert research techniques that we have identified above: his way of flexibly communicating his research aim (the third technique) was in fact two-pronged, where his key research interest was declared as residual or ‘the rest’ within a larger ‘project’ serving as an entry point (the second technique), which was regarded as legitimate and valuable by the local population (2010: 30). Given the relative uselessness of conducting properly communicated interviews and a household survey, none of which yielded meaningful results or data (2010: 10, 27), Trevisani further employed the first technique, by gaining access to observe agricultural practices and processes with the permission of administrators and farmers, during which, albeit unexpectedly, ‘[m]ost of [his] relevant insights into the agricultural system emerged’ (2010: 23).

Most of the more recent works on the political economy, politics, and societal processes in the Central Asian region, however, do not provide such explicit methodological reflections on the (implicit) use of covert research techniques. Calvey’s (2017: 174) finding that ‘much covert research [. . .] has been emergent and unwitting rather than deliberately planned’ and ‘conducted by drift rather than design’ seems to accurately describe anthropological and ethnographically inclined research on Central Asia. One example is Pelkmans’ (2017) book titled Fragile conviction: changing ideological landscapes in urban Kyrgyzstan, which retells conversations that occurred when he and one informant ‘drove up a hill overlooking [the town], picked up some beers along the way, and sat in the grass, relaxing and talking about the things that were on our minds’ (2017: 77) or when he met an old friend of his (95). Furthermore, Pelkmans describes how he ‘took part in a three-day dawat, or proselytizing tour, of the Tablighi Jamaat, a conservative piety movement’ (102) where the bonding with other group members allowed him to collect ‘revealing’ anecdotes and stories (114). Pelkmans certainly demonstrates the ambition to understand the lifeworlds and struggles of his informants from an emic point of view. Nevertheless, or perhaps all the more, his presence in the field as a researcher and the fact that he interacted with people as part of a research project, moves into the background of his bonding.

Further examples for this ‘footnote’ and ‘marginal existence’ of reflections on at least some degree of covertness while doing research (Calvey, 2017: 156) are Ismailbekova’s (2017) Blood ties and the native son and Reeves’ (2014) Border work, both of which give compelling insights into patronage politics and lifeworlds in the country’s north and around the contested Kyrgyz-Uzbek/Kyrgyz-Tajik borders in the Ferghana valley. Both works, however, largely (even though not completely) dispense with procedural details on how the authors obtained consent of research participants or communicated their research endeavors—a pattern that seems to not least reflect editorial policy in the two-book series where the works are published. An exception to this trend comes from Beyer’s (2016) study The force of custom, where she declares her intention ‘to deliver this book to my key informants and to the libraries and schools in my two fieldwork villages’ (2016: xxi). The fact that Beyer’s informants asked her ‘to use their real names’ in the manuscript points to an explicit communication about research results and involvement of research participants into a more dialogical form of knowledge production.

Importantly, besides a general choice or aesthetics on the part of book series editors or the authors themselves, the relegation of methodological concerns to a minimal space within monographs is often driven by a concern not to expose subjects involved in given research projects. As we have already argued elsewhere (Bekmurzaev et al., 2018; Lottholz and Kluczewska, 2017), a more effective way to avoid or at least mitigate the risks that participants (and researchers) face when engaging in research on potentially sensitive topics might be to arrange a longer-term cooperation with clearly defined terms of engagement, and preliminary results to be provided by the researcher for feedback and potential corrections by the partnering organizations or individuals. By engaging with local communities, organizations, and academics and attempting to ‘give back’ to research subjects, such cooperation allows equalizing power-laden relations between the researcher and the researched (see Hale, 2008). While this enables their voices and opinions to be included at multiple stages of the knowledge production process, it does not automatically preclude the emergence of covert and non-transparent knowledge collection practices in the course of cooperation, as we show below.

Application in cooperative research: experiences and reflections from Tajikistan and Kyrgyzstan

When the intention to ‘go overt’ encounters established local frames (Kluczewska)

Between 2016 and 2017, I spent 14 months in Tajikistan conducting field research for my PhD thesis on development aid in this country. My doctoral research, based at a British university, was situated at the intersection of interpretivist International Relations (IR) and social anthropology. It looked at global paradigms underlying Western development aid, which are actively promoted in Tajikistan by donors and international organizations through development projects, as well as everyday practices of localization and contestation of these paradigms on the ground (Kluczewska, 2019). Having heard of some instances when foreign researchers put their local interlocutors and research assistants in Tajikistan at risk through the choice of particularly sensitive research questions and secrecy about their research aims, I opted to do my research in an overt way to avoid potential safety issues. An ethical approval from my university was not enough. While my topic was not as sensitive as researching illicit trade or religious radicalization might be, my focus on local contestations of liberal norms could also meet with skepticism among local academics, civil society actors and, in particular, government bodies. In this case, by ‘overt’ I meant to be explicit about my exact research interests with all research participants, as well as partnering with various local institutions during fieldwork. These collaborations included securing an affiliation with a local research institution and arranging participatory research with a local non-governmental organization (NGO).

In the course of my fieldwork, however, I realized that my intentions could not be understood through the established local frames. In the case of academic collaborations, the challenge stemmed from different understandings of social science research, including the use of theories and research methods. I will shortly outline the current condition of Tajik academia to explain this obstacle. While the local academic sector used to receive significant attention and funding from state bodies in Soviet times, albeit in exchange for support and active promotion of Marxist-Leninist ideology (Amsler, 2007), this changed drastically in 1991. As the socialist economic system collapsed, researchers in Central Asia experienced even such trivial difficulties as lack of paper. The situation in Tajikistan was even more acute due to a civil war (1992–1997), lack of economic resources, and mass emigration of local intellectuals. Up until now, there have been no national research grants in the country, and there are only limited opportunities to do fieldwork in the Western sense of the term. There are also very few exchanges with academic staff from other countries. While local academics speak Tajik, knowledge of Russian, lingua franca in the post-Soviet space, is more and more rare, and few researchers speak English. Furthermore, a continuous brain drain of skilled cadres from underfunded universities is taking place. Anthropology is not recognized as a separate discipline in the country, and the idea of ‘fieldwork’ is automatically associated with ethnographic and archeological research (see Reeves, 2014). Ethnography, in turn, continues to be largely viewed, in accordance with the Soviet tradition, as a description of cultural artefacts of ‘indigenous’ communities. As for IR, a predominantly realist and geopolitically oriented tradition of this discipline prevails locally. Thus, IR is understood either as the history of diplomatic relations between nations, or an analysis of superpowers’ rivalries over spheres of interest.

Against this background, my impression was that I was never able to explain to my local colleagues what my research was about, why I was conducting interviews and doing participant observation, and why I did not refer to the theories of Samuel Huntington and Halford John Mackinder, as many expected me to do. This is not to diminish the value of work of local academics or argue which social science is ‘better.’ Rather, these encounters reveal that I could not conduct a fully overt research simply because the notions that I relied on in my research, such as ‘IR,’ ‘ethnographic study’, or ‘fieldwork’ had different connotations among my local partners. I was unable to find or develop alternative, corresponding local frames to explain what my research was about.

My collaboration with a local, internationally funded NGO faced similar challenges, taking into account the meaning that ‘researching development aid’ has gained within international development. During my stay at the NGO, I was helping with project activities, writing new projects, and conducting monitoring to understand how local NGOs position themselves vis-a-vis their donors. Here, my intention to be overt encountered further challenges. My colleagues in the NGO and other interlocutors working in the development sector assumed that because my research was about development aid, it had to be about measuring aid effectiveness in Tajikistan. While I was trying to grasp everyday practices of development work and often reminded my colleagues about it, they continuously assumed that I was interested in evaluation, that is, in comparing project results with the initial project documents, against the log-frame with inputs and outputs. Here, there was also a misunderstanding regarding the aims of my research. While my intention was to highlight multiple, heterogeneous local agency(ies) in everyday development work in an ethnographic manner, my NGO colleagues believed that I was looking for spicy details concerning, for example, frequent instances of bribing and grant selling between NGOs and donors (see Lottholz and Kluczewska, 2017), and pitfalls of project implementation. My explanations that my research was not policy-oriented did not seem convincing either: would such research not be completely useless if it did not provide any practical recommendations?

From the point of view of my academic partners, for my research to be ‘academic’, I should be just sitting in the library and writing a highly abstract and theoretical text. According to my NGO partners, in turn, because my research was about development aid, I should be doing a highly statistical baseline study or a situational analysis relying on large-N surveys in villages. These encounters reveal the differences in understanding what ‘proper’ research entails which foreground different degrees of covertness.

Navigating the c/overtness continuum in participatory research and knowledge production (Lottholz)

Like Karolina, I tried to realize the idea of cooperative, dialogical research and knowledge production in my doctoral research on community security and peacebuilding practices in Kyrgyzstan (Lottholz and Kluczewska, 2017; Lottholz, 2018). Regardless of my attempts to ensure transparency and dialogue as part of an overall ‘cooperative’ approach (Bekmurzaev et al., 2018), the participatory observations I conducted as part of project implementation activities of an NGO network that works on police reform and community security and is based in Kyrgyzstan’s capital Bishkek (see Lottholz, 2018: ch. 7) ended up involving covertness.

In establishing cooperation with the NGO network in question, I offered to the coordinators to conduct an ‘accompanying research’ (Russian: soprovozhdaiushee issledovanie), which would include regular conversations at the head office, participatory observation at project implementation events in select municipalities, and further, individual interviews with members or member organizations of the network in different communities across the country. While this case study proved to be the most exhaustive one within my PhD project, this experience also raised important questions about my status as a researcher and the terms of my involvement in the organization.

The latter aspect became obvious when I accompanied different activists from the organization to project implementation visits in both northern Kyrgyzstan (close to the capital, Bishkek) and in the south-western Batken province. The events I was allowed to attend were held as part of a project aimed at strengthening social partnership between local administrations, law enforcement organs, and the population for public security provision and crime prevention, financed by a United Nations agency. While conducting research in these two places, our general agreement was that I could participate as a volunteer to support and facilitate the running of the respective training events and group meetings. This legitimated my presence on the ground and the use of conversations I witnessed in my research. The organization’s work in a community in the north proved rather tedious, as the local working group was struggling to assemble all its members and sometimes required repeated scheduling of meetings in order to find agreement on the prioritization of security issues and a decision on measures to be taken on them, as stipulated in the project plan. When navigating these issues, the organization’s coordinator for this community decided not to disclose more details about my involvement with the organization. I tacitly agreed with this approach, as my previous experience had shown that being too forthcoming with information about my foreign origin and affiliation could also distract the attention of people involved in security and peacebuilding projects.

My participation in another event in a southwestern community showed that my ‘expertise’ as a scholar could also have positive effects. In this case, when opening a planning event, the activists from the organization’s head office introduced me to the community both in my supporting function for the event and in my researcher position: ‘Philipp is producing a large academic work here. He is writing about how we are building a normal country [kak stroim normalnuiu stranu].’ This minimal introduction sufficed for me to become accepted as part of the organization’s collective endeavor, which was indicated by people’s active interest in my background during breaks, while no one expressed concerns over my presence at the event (see Lottholz, 2018).

Nevertheless, the interaction in both communities drew on the covert research technique of participant observation to the extent that the community members, administrators, and law enforcement staff in the respective working groups had no way of knowing what exactly my research was about (or, as in the first community, that I was conducting a research project in the first place). The decision on this arrangement was ultimately not up to me but it was taken by my interlocutors in the police reform organization, with whom I later shared my impressions and initial drafts of analyses from the visits. While the encounters described above illustrate the potential benefits for NGOs to openly communicate their cooperation with researchers (vis-a-vis risks of distraction by suspicions among project working groups), some degree of covertness inevitably shaped the production of my research results and communication among the members of the police reform organization.

Ideally, the production of knowledge in academic-activist cooperation should occur in a ‘dialogical’ manner, where researchers share their outputs and receive feedback from their partner organization. In practice, however, civil society organizations are extremely busy with meeting deadlines, applying for new funding and day-to-day project work, which leaves little room for lengthy discussions of dissertation chapters or academic journal articles. This cooperation was no exception, as the feedback I received on my work was limited to giving an ‘OK’ not only for my interpretation and presentation of the organization’s work and its positive effects, but also for its limitations. It turned out that to receive more detailed feedback and reactions to my assessments and criticisms of the community security work, I would need to contact members of the network personally. Further, the perhaps most important limitation on dialogue was that for a better understanding, most of the organization’s members needed my draft analyses to be translated into Russian, for which I had neither the time nor the resources back then. As a result, similar to the selective approach to communicating my research activities in the organization’s community-level work, the content and main arguments of my work were not too well known nor discussed among its activists but mostly turned around in my interaction with the main gatekeeper, the board member in the organization’s head office in Bishkek (which presents a case of the second discussed covert research technique, i.e. reliance on gatekeepers). In the meantime, I managed to have the relevant dissertation chapter translated into Russian and share it with network members. This did enable a more in-depth discussion and gathering of feedback, but the exchange was again limited to three members only. This demonstrates that even in contexts of research collaboration with activists and professionals who have the literacy and capability to read and comment on findings (which is not always the case, as in certain anthropological research, for instance), the everyday duties and challenges faced by partners may still prevent them from engaging in an in-depth dialogue. This way, despite investing more efforts into dialogical knowledge production than is usual, certain details and wider implications of my research remained covert—a fact that underlines how covertness can only rarely be avoided completely and needs to be embraced and critically reflected upon in the research process.

Conclusion

Covert research is an underestimated, yet crucial part of social inquiry. We have demonstrated that covert research techniques are widely used today, although often without acknowledgment, especially as a way to ensure the safety of researchers and research participants in restricted settings and in relation to sensitive topics. Furthermore, we have identified three commonly acceptable practices in field research, which de facto rely on some degree of covertness: participant observation, reliance on gatekeepers to ensure access, and flexible framing of research objectives depending on the audience.

Building on previous arguments (Calvey, 2008, 2017; Spicker, 2011), we have attempted to further reconceptualize covert research by moving away from the covert versus overt binary. We did so by situating the covert research method in the context of current academia and by highlighting how current demands of knowledge production push researchers towards covertness as a way to generate groundbreaking findings on sensitive topics while trying to stay safe. Moreover, we argued that covertness is less a result of the researcher’s decisions during fieldwork (to be covert or overt) as it depends on diverging perceptions of what constitutes (transparent) research on the part of researchers and research participants, respectively. Covertness is an outcome of spontaneous and often unexpected conditions and is much less subject to deliberate intention and control.

As an alternative to the currently prevailing trend, we have argued that field research that relies on long-term collaborations with local partners can be a way to overcome covertness, or at least reduce it. In this way, knowledge is co-produced rather than extracted or ‘translated’ by researchers. Yet, we have also shown the challenges of covertness in such collaborations, by referring to our own research experiences in Tajikistan and Kyrgyzstan. The remoteness of Western academia from the researched contexts often a priori renders academic research covert, in that it is simply hard to understand locally. In this sense, our reflections on the de facto use of covert research techniques and a call for more dialogical research are not intended to claim that the approach we favor is in any way superior in ethical or moral terms. On the contrary, we extend the argument about the necessity, perhaps even inevitability, of covert research techniques to our own research. We hope to inspire further discussion on the de facto use and ethical implications of covert research, as well as on the possibilities of reducing covertness and achieving more transparency and dialogue in social inquiry.