Recognizing research participants’ fluid positionalities in (post-)conflict zones

This article re-conceptualizes the highly ambivalent relationships between researchers and research participants in conflict zones, with a focus on recognizing respondents’ multiple and fluid positionalities. Standardized and dominant approaches to qualitative research are largely based on essentialist and infantilized portrayals of research participants and neo-colonial assumptions regarding the research relationship: informants are presumed to be inevitably vulnerable and in need of external protection, while the researcher is positioned as the omnipresent expert in control of the research process. In reality, however, research participants rarely exclusively occupy the ‘oppressed victimhood’ axis of identity and frequently take on active roles in the research and data collection process in a myriad of ways. I elucidate how especially in (post-)conflict zones, research participants frequently re-shape power dynamics by exercising agency over the researcher and the research process. While previous studies have considered how informants’ agency can shape processes of knowledge production, in this article I expand this focus by examining how key-informants can, and frequently do, facilitate the researchers’ safety and security. I specifically draw on personal experiences of empirical research in Northern Uganda. I demonstrate how in a particular moment of post-conflict insecurity – while being trapped in-between the exchange of gunfire between the Ugandan police and an armed group – one of my key-informants ensured my physical protection and safety, thereby exercising power over me and the research relationship. The key-informants in this context thus occupied multiple positionalities – ranging from informant to protector, evidencing that research relationships are never static but rather contextual, shift and fluctuate. Such ambivalent and fluid power dynamics are more reflective of the lived realities of qualitative research and can influence the research process by positioning researchers and research participants on more equal terms.


Introduction
security and physical well-being. Illustrated through this scenario, Okello -and various other research participants, as demonstrated further below -exercised varying levels of agency and control over me and the research process at large, thereby occupying multiple positionalities. This evidences that the research relationship is contextually dependent and never static but 'shifts and fluctuates depending on the different constellations of identity and power at play' (Thapar-Björkert and Henry, 2004: 371).
In this article, I utilize this particular incident -which I further reflect upon below -to reflexively recognize the multiple positionalities occupied by research participants during research in (post-)conflict zones, ranging from respondent to protector. Echoing Thapar-Björkert and Henry, 'the researched are not only and always in a subordinate position in relation to the researcher' (2004: 373) -as it is often portrayed -'but can negotiate and challenge conventional and uni-directional axis of power' (373). To situate my argument, it is important to emphasize that highlighting the circumstances under which respondents can also exercise control does not intend to downplay the significant risks that research participants in conflict zones are exposed to, or to negate the inherent structural power disparities that typically characterize research relationships.
All too often, however, the representation of research relationships during (post-) conflict zones is characterized by an unreconstructed view, assuming and simultaneously reproducing an asymmetrical power paradigm between researchers and informants, in favour of the former (Malejacq and Mukhopadhyay, 2016), thereby generating essentialist and infantilizing portrayals of research participants (Connor et al., 2017). Within these frameworks -which frequently take on neo-colonial tendencies -scholars are positioned and position themselves to be in control of passive research subjects and the entire research process at large, while informants are presumed to be inevitably vulnerable, always in need of external protection and without any agency. In praxis and reality, however, research participants rarely exclusively occupy the 'oppressed victimhood' axis of identity, while the researchers' place is often one of only limited power. Instead, respondents frequently exercise agency and control over the researcher and the research process in different ways (Fujii, 2012).
This requires us to reflexively re-assess research relationships and re-conceptualize binary constructions of 'the researched' as ubiquitously powerless vis-à-vis the researcher as the only powerful actor in this dyad. Instead, 'the research space can be seen as one of complex power relations, where power is put into play by a number of actors' (Thapar-Bjökert and Henry, 2004: 365) and factors in diverse ways and under different circumstances. Drawing on these premises, Malejacq and Mukhopadhyay (2016) consequently urge scholars that 'the "curtain" shielding fieldwork from greater scrutiny must be pulled back for a more honest conversation about the challenges involved in pursuing this brand of scholarship [in (post-)conflict zones] today' (1012). This normative call constitutes the background of this article, focusing on key-informants' multiple and fluid positionalities during research in post-conflict Northern Uganda, with a particular focus on who provides security for whom under what circumstances -a conundrum that has not yet been sufficiently addressed in the growing body of material on ethics, methods and positionalities in (post-)conflict zones (see Cronen-Furman and Lake, 2018;Baele et al., 2018). 3 In recent years, previous studies have begun to acknowledge that respondents in conflict-affected settings can influence knowledge production processes and exercise agency over which information to disclose (Gerharz, 2009;Mwambari, 2019a;Vlassenroot, 2006), while a growing body of literature also reflects on the dangers of research in warzones (Baird, 2018;Sluka, 2012). Despite these growing ethics debates, however, it thus far remains only insufficiently explored to what extent and how research participants can -and in reality often do -navigate the researchers' safety and security in violence-ridden settings, thereby momentarily controlling the research process. In particular, how precisely these dynamics play out in 'the field' -embedded in empirical reflections -remains only inadequately addressed. Drawing on my experience of conducting research in post-conflict Northern Uganda, I intend to contribute towards addressing this prevailing gap, thereby engaging in a longer-term and multi-faceted process of re-conceptualizing the highly ambivalent relationships between researchers and research participants. I elucidate how especially in (post-)conflict zones, research participants frequently (re-)shape power dynamics by exercising agency and control over the research process, including over researchers' security-thus occupying multiple positionalities that go beyond the stereotypical representation of the vulnerable respondent in need of external protection.
The article commences by providing a brief overview regarding the unreconstructed view of research relationships that dominates discussions around research methods and ethics. I then proceed with critical conceptual-methodological reflections on positionalities and reflexivity during research in conflict zones more broadly, before briefly contextualizing the research context from which my observations and argumentations derive. To illustrate my core argument regarding research participants' multiple positionalities, I then revisit the introductory vignette presented above, and analyse additional instances that occurred during my research where key-informants exercised different forms of control during the process. Based upon these empirical experiences and reflections, I emphasize the importance of re-conceptualizing dominant understandings of research relationships and subjectivities, and in particular of recognizing respondents' fluid positionalities in (post-)conflict zones.

Research relationships: an unreconstructed view
Research in (post-)conflict zones is typically characterized by a myriad of methodological and ethical challenges (Fujii, 2012;Sriram et al., 2009;Wood, 2006). One persistent ethical challenge refers to the different and at times highly fluid and ambiguous positionalities embodied by researchers and research participants involved in the research process (Bond, 2018). A wealth of critical research in recent years -particularly drawing on ethnographic (Bhattacharya, 2014;Millar, 2017) and feminist epistemologies (Ackerly and True, 2008;Jakoby, 2006) -has begun to recognize and deconstruct the varying manifestations of power disparities between researchers and informants (Eriksson Baaz and Utas, 2019;Malejacq and Mukhopadhyay, 2016). This constantly growing body of literature importantly lays open how external (and most often white) researchers enjoy a variety of relative political, socio-economic, cultural and gendered privileges in comparison to their research participants, and how these are shaped by and contextualized within unequal power structures, which become particularly exacerbated in (post-)conflict zones (Thapar-Bjökert and Henry, 2004).
Discussions about research relationships in many ways lie at the heart of broader debates about research ethics and about qualitative research more broadly. Initially drawn into the social sciences and humanities from a medicalized framework, ethics procedures and guidelines are predominantly framed as aiming to protect respondents from (additional) harms and risks due to the research process. These procedures have thereby registered significant advances into re-calibrating unequal power relationships and contributed towards addressing a troubling history of unethical research (Connor et al., 2017). In their dominant application, ethical guidelines and procedures broadly centre around the methodological and ethical principle of 'do no harm' (Fujii, 2012), as well as the protection of research participants' identities -whereby privacy, confidentiality and anonymity frequently constitute key pillars (Baez, 2002). Institutional Review Boards (IRBs) and Research Ethics Committees (RECs) also specifically demand that research involving human subjects is based upon informed consent of those interviewed.
Yet, despite significant advances to minimize research participants' exposures to risk, scholars also increasingly point towards various shortcomings and limitations inherent to research ethics processes (Bhattacharya, 2014;Parkinson and Wood, 2015), inter alia criticizing that these procedures have largely become 'a risk management exercise at the behest of the host institution or funding body' (Tolich and Fitzgerald, 2006: 72). Critique has been raised by researchers that rather than genuinely seeking to protect research participants and being truly concerned with respondents' safety and well-being, ethics guidelines and procedures -in the ways in which they are currently administered by RECs and IRBs -instead primarily seek to protect universities or research councils from legal repercussions arising from any research conducted under their auspices (see Kohn and Shore, 2017;Bhattacharya, 2014). At the university (in Germany) where I am currently based, for instance, the ethics committee is housed within the university's legal affairs department -indicating the committee's legal emphasis, rather than its ethical concerns per se.
IRB processes also often require risk and safety assessments, emphasizing the researchers' responsibilities to ensure the safety, security, protection and well-being of both researchers and research participants (Baird, 2018). Frequently, the guiding frameworks of these processes are based on simplistic assumptions that research participants are inherently vulnerable and in need of the intervening researchers' protection, and that researches will always and exclusively have to ensure for others' and their own safety. As argued by Kohn and Shore, for REC processes, 'researchers are required to respond to the near-impossible demands to guarantee that participants are protected from assumed emotional [and physical] "harm" arising from any interpersonal encounters' (2017: 231). The assumptions behind these demands, Kohn and Shore argue, are often 'paternalistic, distrustful and ethnocentric' (Kohn and Shore, 2017).
To illustrate, the National Guidelines for Research involving Humans as Research Participants of the Ugandan National Council for Science and Technology (UNCST)which were applicable for this study -repeatedly emphasize the researchers' responsibility to 'safeguard the [. . .] safety of research participants' (2017: 10, 23). Likewise, a 2019 report of the American Political Science Association (APSA) Ad-Hoc Committee on Human Subjects Research emphasizes the researcher's responsibilities ' [. . .] for the safety and security of their research subjects, including physical harm, emotional harm and third party harms' (2019: 2). In addition to those formal ethics guidelines, existing scholarship on these topics similarly typically 'assert[s] that a researcher should think through his or her safety and the safety of those involved [. . .] (Mwambari, 2019a: 4;see Mertus, 2009). These reiterations are fundamentally important to uphold the 'research ethics 101' (Fujii, 2012) principles, and ultimately to protect research participants from harms and risks. At the same time, however, such assessments leave very little space for acknowledging how others -including research participants -can also become involved and tasked with providing for the researchers' (as well as their own) security -a gap that my reflections in this article seek to address.
Scholars have hence begun to criticize that dominant and standardized approaches to research ethics are frequently structured around an unreconstructed and patronizing view of research relationships (Thapar-Björkert and Henry, 2004) that 'take[s] a paternalistic approach to research participants' (Connor et al., 2017: 1) and that ignores a myriad of psycho-social as well as physical challenges and harms that can arise from research in fragile settings. As argued by Connor et al. (2017: 1), these 'processes have fundamentally corrupted the research process and infantilized the [research] subject [. . .]'. Reflecting upon her experience of researching ex-combatants in Sierra Leone, Sayra Van Den Berg for instance points out how research participants often 'felt a sense of being used and then discarded once the research was over' (2019: 2) -illustrating how informants are often treated passively and subjected to research processes, rather than having their agency fully acknowledged and respected. Mwambari, drawing on research with local research assistants (RAs) in Northern Uganda, similarly observers that especially in sub-Saharan Africa, academic fieldwork -and related assumptions about ethics -'still relies on colonial practices of a white man going to find information in remote African places, inhabited by "vulnerable" and underdeveloped communities' (2019b). He critically attests that 'in most cases, the foreign researcher ends up being labelled as an "expert", while the local assistant's [and respondents'] voice[s] [are] silent or silenced' (Mwambari, 2019b) -representing neo-colonial conceptions of research relationships and a patronizing view towards research participants. In recent years, certain (often marginalized) branches of qualitative (social science) scholarship have attempted to move beyond such narrow representations, and critical scholars have directed attention towards inherent and often problematic power relationships between researchers and research participants (Schulz, 2019;Thapar-Bjökert and Henry, 2004). These power asymmetries between those conducting the research and those subjected to it appear to be particularly stratified during research on sensitive issues in politicized contexts, when engaging populations in vulnerable situations, and especially during research in (post-)conflict settings (Baele et al., 2018;Sriram et al., 2009). To eventually overcome hierarchical research relationships and discrepancies of authority and control, researchers working in fragile or violence-affected contexts increasingly reflect upon questions of power, ethics and the politics of knowledge production (Bunting and Quirk, 2019; Cronen-Furman and Lake, 2018), while collaborative and participatory research processes likewise gain growing popularity (Robins and Wilson, 2015;Lundy and McGovern, 2006). Centralizing research participants' collaboration, participation and empowerment (Salmon, 2007), these approaches to research inter alia intend to 'ground knowledge production in the everyday lives of those most affected' (Robins and Wilson, 2015: 236), as well as to challenge power discrepancies and exploitative methodological approaches (Pittaway et al., 2010). Required for these methodological approaches is, amongst others, critical reflexivity on the side of researchers, about their subjectivities and about research participants' positionalities.

Positionalities and reflexivity during research in (post-)conflict settings
Critical reflections on subjectivities and positionalities and on ambivalent power relations between researchers and informants have a relatively rich tradition in critical feminist research (Ackerly and True, 2008;England, 1994;Oakley, 1981) as well as in anthropology (Millar, 2017). Scholars in these (and associated) fields have long been concerned with paying critical attention to how their positions in the research contextintersectional defined by class, ethnicity, religion, gender, sexuality and/or socio-economic status -stand in contrast to research participants and how this shapes and impacts the research process (Bond, 2018;Bouka, 2015;Thomas, 2018).
While this application of reflexivity is relatively common in these disciplines and epistemologies, however, the growing sub-field of peace and conflict studies -as a research area dominated by political science and international relations -all too often avoids these critical exercises. Even though critical scholars within peace and security increasingly reflect upon researchers' identities and positionalities (Cronen-Furman and Lake, 2018;Fujii, 2018;Yacob-Haliso, 2018), the mainstream literature on researching conflict and violence largely avoids these exercises and tends to treat the subject of the researcher as just the neutral instrument for collecting data. In particular, the (neo-) positivist traditions that dominate contemporary peace and conflict research are often centred around 'notions of positionality that can preclude a researcher's steady (re)consideration of her [or his] place within the research environment' (Malejacq andMukhopadhyay, 2016: 1012). As previously observed by Kim England (1994), positivistic epistemologies indeed largely depend upon 'methods that position the researcher as an omnipotent expert in control of both passive research subjects and the research process' (81). Positivist presumptions about interviewing during research likewise commonly assume a simplified interview scenario characterized as a hierarchical and unidirectional process, in which the interviewers control and the interviewed are controlled (Acker et al., 1991). Researchers working in positivist traditions have hence long been educated as holding power over their research subjects and participants 'simply by introducing them into conversation' (Malejacq andMukhopadhyay, 2016: 1013). As a result, it often 'remains common place for political scientists to abstain from delving into the details of how they arrived at their claims' (Malejacq andMukhopadhyay, 2016: 1012). In many ways, this approach both embodies and upholds the above-identified unreconstructed view of research relationships.
The exercise of reflexivity can indeed be a humbling and daunting task, as researchers inevitably have to come to terms with the fluidity, elasticity and limitations of their positions, privileges and power (Acker et al., 1991). Recognizing one's factual position of only limited power would thereby contradict some of the underlying principles and imperatives of positivistic research, in which researchers are expected to exercise unlimited control over the process, in order to account for intervening factors. In many ways, it almost seems as if the status of 'external researcher' is socially constructed as necessitating and/or implying invulnerability, while the label of 'research participant' is often equated with (assumed or prescribed) vulnerability and victimization, particularly so during research in fragile and violent contexts. This ascribed and expected invulnerability of the researcher is in practice often equated with full and uncompromised control over the research process at large, including over research participants. More often than not, these assumptions and the associated roles in this power dyad are heavily gendered, ethnicized and/or racialized (Parashar, 2019), with women from the Southern hemisphere constructed as particularly sub-ordinate and powerless, and the white masculine researcher as extraordinarily powerful.
Critically engaging with these positivist assumptions, scholars across different (sub-) disciplines working on peace and security (Cronen-Furman and Lake, 2018) have begun to break open these categories, primarily by way of foregrounding the contextual limitations of researchers' influence and control (Fujii, 2018) and by questioning the dichotomous paradigm of the powerful researcher vis-à-vis the powerless informant (Fujii, 2012;Thapar-Björkert and Henry, 2004). Such studies primarily recognize that identities during research processes are intersectional and that often -of course highly dependent on context, timing and circumstances -researchers can (momentarily) constitute the less influential party in this power dyad (Malejacq and Mukhophadyay, 2016). Despite implying numerous conceptual and methodological inroads into re-conceptualizing research relationships, however, Thapar-Björkert and Henry critically note that most of these critical reflections 'did not [yet] necessarily result in the acknowledgement that power can also be exercised by participants ' (2004: 363).
Before proceeding with my argument, however, a cautious caveat is required: acknowledging how research participants can exercise control is not to negate the structural power relations that typically characterize research relationships -often tuned in favour of the intervening researcher -but rather to recognize research participants' agency, to avoid the risk of essentializing and infantilizing their often highly ambivalent experiences and positionalities. At the same time, locating respondents' agency is not intended as de-centring the lived realities of our research participants in favour of our own -which would prove particularly problematic in 'dangerous environments where [research participants'] engagement in the research implies the potential to cause risks for themselves and their communities' (Pearce and Loubere, 2017: 13). Undoubtedly, Western(-based) scholars generally hold considerable structurally embedded material, economic, social and political advantages in comparison to their respondents which cannot be denied. This greatly privileges externally intervening researchers in relation to research participants as well as local researchers and assistants (Bouka, 2015;Mwambari, 2019a), constructing and re-producing unequal power hierarchies inherent to research relationships (Parashar, 2019). Specifically with regards to safety and security provisions, researchers frequently enjoy the relative luxury of having their return tickets from the insecure warzone to their secure home-countries in their back-pockets. As observed by Mwambari, 'especially when the study touches upon sensitive topics, the foreign researcher can travel back to their home country, but often leaving the assistant's [and respondents'] safety in their community compromised' (2019b). Importantly, distributions of power in the context of qualitative research are also conditioned by intersectional identities of researchers and the researched respectively. To illustrate: Even though young European graduate students may generally be perceived (and perceive themselves) as more powerful vis-à-vis an elderly, rurally based female farmer in a war-affected setting -and undeniable enjoy tremendous socio-economic and political privileges in comparison -this division and distribution of power would not necessarily apply in all its facets to the researcher's relationship to an influential local cultural leader or to (former) warlords or rebel commanders. Malejacq and Mukhopadhyay emphasize that 'power asymmetries can be particularly pronounced in research with elites, especially with outsized access to coercion and capital' (2016: 1013). In most societal and cultural (post-)conflict contexts -often characterized by hetero-normative and patriarchal gender identities, relations and hierarchies -these differences get further exacerbated for younger and/or female researchers, as well as scholars from ethnic/racial minority backgrounds (Thapar-Björkert and Henry, 2004;Parashar, 2019).
Yet despite these concerns, we ought to be careful not to uncritically generalize these dominant positionalities and distributions of power universally and ubiquitously across research scenarios, without taking into account localized dynamics and configurations inherent to each individual research scenario. Instead, a more transparent and holistic examination of respondents' positionality and agency during research processes is needed. As demonstrated above, emerging scholarship has indeed begun to reflexively analyse researchers' and research participants' contextual and temporally contingent subjectivities during research in conflict zones (Malejacq and Mukhopadhyay, 2016). Most existing studies employing rigorous reflexivity in this context uncover how research participants can exercise control and agency in their roles as part of data collection processes by purposefully and circumstantially choosing which experiences and viewpoints to share in which situation and with what researchers, and how this carries implications for the quality and representativeness of research findings, respectively (Gerharz, 2009;Mwambari, 2019a;Vlassenroot, 2006). To illustrate, Malejacq and Mukhopadhyay (2016) emphasize that empirical research can be characterized as 'a symbiotic relationship between conversant, each of whom has the capacity to shape the conversation' (1013) and thus to exercise power and influence.
Furthermore, respondents can at times utilize research scenarios for their own personal and political gain, for instance to attain status within their communities. In this context, the researcher can (knowingly or unknowingly) become 'an important element in research participants' strategies to attain certain objectives' (Vlassenroot, 2006: 197). Drawing on research in war-torn Sri Lanka, Gerharz (2009) for instance observes how researchers were at times instrumentalized as promoters and advocates by the Liberation Tigers of Tamil Eelam (LTTE) militant group. Such insights are corroborated through reflections focused on Afghanistan by Malejacq and Mukhopadhyay (2016).
In combination, these examples indicate multiple positions that can be embodied both by researchers and research participants, as well as ambivalent power relationships. As noted by Mwambari in the context of research in Northern Uganda, the 'positionality of any person involved in research [. . .] is not fixed as it evolves as researchers and collaborators relate to each other and to the context being studied' (2019a: 1). This, in turn, forces us to realize that it is not always and exclusively the intervener who controls and the informant who remains passive and subjected (Björkdahl and Selimovic, 2015). Malejacq and Mukhopadhyay therefore conclude that 'the notion that the researcher has control over [the research] context is simply out of alignment with facts on the ground. A researcher can actually be much more dependent on her [or his] informant than the informant is on her [or him]' (2016: 1013). These are important steps towards recognizing research participants' agency in the research process, and to move beyond essentialist and paternalistic assumptions about research relationships, which continue to dominate in debates about qualitative research, particularly so in violent and fragile contexts, as demonstrated above.
While most existing reflexive scholarship on research participants' agency thus concentrates on how respondents can influence knowledge production processes (Gerharz, 2009;Mwambari, 2019a), however, fewer studies have focused on how this may influence the researchers' positioning (Thapar-Bjökert and Henry, 2004). In particular, to what extent and how research participants in (post-)conflict settings can exercise control over the researchers' safety and security remains only insufficiently explored -a gap that I seek to engage within this article.

Fieldwork under fire: security during research in (post-)conflict zones
The empirical reality of research in conflict zones indeed demonstrates that researchers frequently find themselves trapped in situations characterized by insecurities, violence or political turmoil (Mertus, 2009;Nilan, 2002;Rivas and Browne, 2018). Even though there has conventionally been 'little mention [. . .] of the practical matters of survival in perilous field sites and [of researchers'] experiences of violence in the field [. . .]' (Kovats-Bernat, 2002: 208), a growing body of literature increasingly reflects upon these experiences, showing that undertaking field research in violent places can inevitably cause threats to the physical safety of researchers (Nordstrom and Robben, 1996;Sluka, 2012). In many ways, it should not be surprising, but rather obvious, that the widespread levels of violence and atrocities that (per definition) characterize these settings cannot (and do not) stop at the researchers' doorsteps, but instead entangle them as well.
For instance, Ivana Macek (2009), who conducted long-term ethnographic research in Sarajevo during the war writes: 'living in the besieged city alongside Sarajevans, I too had to employ all of my faculties -my intuition and cognition, my senses and emotions -in order to manage from day to day [. . .]' (2009: 11). Reflecting on the inherent insecurities of ethnography in wartime, Macek 'recount[s] [her] own responses to being shot at -sudden depression as my sense of purpose evaporated' (Macek, 2009). In Nordstrom and Robben's insightful selection of essays on Fieldwork Under Fire (1996), a range of scholars similarly reflect on the challenges of encountering and responding to danger in fieldwork in (post-)conflict zones as diverse as Northern Ireland, Croatia, Mozambique or Palestine. As ethnographers conducting studies on violence and culture, the authors thereby reflect on their 'initial confrontation with violent events' (Nordstrom and Robben, 1996: 14).
While this body of reflexive scholarship is primarily focused on how researchers in warzones themselves engage with experiences of being affected by war-related violence and insecurities (Baird, 2018;Nilan, 2002), what remains only marginally explored in these studies is an honest analysis of how research partners and collaborators (willingly or unwillingly) become occupied with and responsible for the security provisions of external researchers, a lacuna that this article engages with. Indeed, as illustrated through the introductory vignette above, in fragile and violent research contexts, research participants -and especially assistants and key informants -frequently (need to) navigate the researchers' security and are tasked with ensuring for their physical and psychological safety. Throughout the growing literature on the ethics and practicalities of conducting research in fragile settings, however, these instances are only insufficiently reflected upon, and it remains only inadequately acknowledged that the positionalities in these situations shift, and how this can change power relations between the researcher and the researched.
By drawing upon personal research experiences from Northern Uganda, in this article, I seek to open up a space for conversations about this and other related experiences. By illustrating how key informants in a particular moment of post-conflict insecurity ensured for my physical safety and well-being -alongside exercising other forms of control throughout the research process -I aim to contribute empirically to the growing literature on positionalities, relationships and reflexivity during research in conflict zones (Baele et al., 2018;Bond, 2018). I thereby aim to contest and correct conventional uni-directional assumptions regarding the researcher's exclusive ability to protect research participants and their respectively assumed inabilities to do so, which ultimately essentializes respondents as ever vulnerable and in need of external protection.

Methodological reflections
Departing from these general conceptual-methodological reflections, in this section I briefly introduce my overall research engagement in Northern Uganda and the specific research project from which I draw my experiences. I present this background to position my reflections and argumentation within their respective context, as well as to locate my own positionality. I have conducted research in post-conflict Northern Uganda since 2011, but the particular examples I draw from in making my argument throughout this article specifically derive from 7 months of doctoral research in Northern Uganda's subregion of Acholiland, between January and July 2016, focused on wartime sexual violence against men. By exploring this highly sensitive and stigmatized issue in a hetero-normative and hetero-patriarchal context, the research project engaged particularly vulnerable and often silenced research participants (Schulz, 2019).
In addition to different data collection techniques -such as key-informant interviews and focus group discussions -I specifically conducted four participatory workshop discussions with male survivors, guided by a participatory action research (PAR) framework. These workshops aimed to facilitate an engaging, emancipatory and empowering research environment in which male survivors were enabled to exercise agency and control, in order to conduct research with people, rather than on them (Robins and Wilson, 2015). An affiliation with a local partner institution -the Refugee Law Project (RLP)also allowed me to situate the research as part of an on-going institutional praxis of sustainably working with male survivors, and to develop close collaborative working relationships with some of my interlocutors.
While conducting the research, and specifically in making my arguments throughout this article, I reflexively acknowledge my own positionality as a young, male, white, European academic, which positioned me as an obvious outsider, if not even stranger, to the context of Acholiland, and therefore also to my research participants. As a result, and despite my participatory research design aimed to contribute towards a longer-term process of democratizing research relationships and practices, my positionality in relation to my research participants is characterized by power asymmetries and enhanced access to various social, cultural and material resources and socio-economic capital. I therefore recognize how my own positionality influenced not only my relationship to research participants but also the research process at large, and the way I am perceived by others in the field. Inspired by Baines (2017), I am also mindful of the critique that outside scholars, such as myself, 'will always be self-limited in their ability to listen to and write outside the yoke of colonialism' (28).

Empirical reflections: recognizing key-informants' positionalities as agents and actors
In the introductory vignette that opened this article, I have sought to illustrate one of my key-informant's multiple positionalities, by reflecting on a particular instance of conflict-related insecurity during empirical field research in Northern Uganda in June 2016. While I was trapped in the compound of a local bar in-between gunfire exchanges involving the Ugandan police and an armed group for several hours, one of my key-informants -whom we refer to as Okello -navigated the post-conflict terrain and ultimately ensured my physical protection and safety. Taking certain personal risks for himself, he provided for my security, thereby temporarily exercising control over me and our research relationship at large.
Prior to this particular incident, we had met regularly over the course of the last few months and for the purpose of our conversations, we also visited each other's homes. Despite this fairly close nature of our contact, however, to this point in June 2016, our relationship was primarily one of external researcher and key-informant. This planned meeting in the bar was indeed the first-ever occasion where we would meet outside the context of a specific research-related scenario. Taking the nature of this relationship into account further underscores the particular dimension of the situation as described above.
Furthermore, during the whole incident, I was not only in close contact with Okello, but various other key-informants also made sure to call, to follow-up on me, give me advice on what to do or simply to ensure that I was safe. Most of them indeed offered practical advice to guarantee my safety. But as elaborated upon in the introduction, especially Okello was instrumental in constantly updating me and ensuring my safety and protection. Considering the security situation that evening, Okello took personal risks, in order to ensure for my protection and well-being, thereby exercising immediate agency and control over me, impacting on our relationship (in the long-term) and occupying different positionalities.
Reflecting on some of the underpinning assumptions, in many ways it seems paradoxical to assume that researchers as external interveners are better positioned and suited to always and exclusively guarantee the safety of respondents, without acknowledging respondents' relative advantages, based upon their local knowledge, familiarity with context and prior experiences. Frequently, researchers purposefully choose to study these populations precisely because of their lived realities during the conflict and post-conflict periods, often to analyse their resilience and coping strategies. We thus cannot assume that we as researchers -who evidently do hold numerous privileges and exercise control in certain situations -are always and ubiquitously in a position to control, navigate power and ensure safety without recognizing other possibilities and positionalities. Such presumptions cannot adequately reflect the complex realities of research and are indeed based upon neo-colonialist assumptions regarding research relationships.
In addition to this particular instance of Okello ensuring my safety in a moment of insecurity, during the course of my research, there were (of course) multiple other examples of research participants and collaborators exercising agency and control over me, our relationships and the research process at large. For instance, when I was suffering from malaria (twice during this period), key-informants checked up on me and brought juices and food, out of concern for my physical and mental well-being. With regards to research processes, various respondents also repeatedly shielded off certain (and for me important) information, before only revealing them later, often towards the end of the research process. Of course, informants may also well be expected to have shielded off certain information entirely from me, in order to protect their stories and narratives from researchers in general, and from me in particular. These dynamics resemble what others have previously documented across different contexts with regards to the politics of knowledge production, of who shares information with whom, where and how (Gerharz, 2009;Vlassenroot, 2006). One of my research assistants similarly admitted that he only chose to put me in contact with some of his close contacts -who became invaluable for the purpose of the study -after a considerable amount of time, and after having gotten to know (and trust) me and my approach to research and ethics, respectively. Accompanying the introductory vignette, these examples serve to illustrate the multiple contexts where research participants exercise agency and control, instead of exclusively occupying the passive axis of identity that is often ascribed to them.
What implications can be drawn from these reflections, then? First, and perhaps most importantly, that the researcher's position is often one of only limited power, and that research participants themselves also serve as agents and actors, including with regards to security provision. While scholars have previously begun to show different ways in which research participants exercise control over data collection processes, this particular point about how respondents can ensure researchers' safety has not yet been sufficiently established in the growing research ethics literature in conflict settings. Erin Baines' observation that 'one's vulnerability in one relationship does not define the person as ever vulnerable ' (2017: 14) thus certainly applies to research participants' subjectivities in (post-)conflict zones. Recognizing these fluid and locally contingent positionalitieswhich are never static but change over time and are dependent on contexts and relationships -can help us to re-conceptualize binary constructions of the ubiquitous powerless 'researched' vis-à-vis the researcher as the only powerful actor (Thapar-Björkert and Henry, 2004) during research in warzones and post-conflict settings.
Furthermore, taking seriously the empirical reality that during research in (post)-conflict settings both researchers and research participants exercise agency in different ways (Björkdahl and Selimovic, 2015) -including with regards to security provision -can contribute towards a longer-term and multi-dimensional process of democratizing and decolonizing research practices and relationships (Salmon, 2007;Smith, 1999), facilitating research carried out with local people, rather than on them (Robins and Wilson, 2015). This, in turn, can get us closer towards research as more than extraction (Bunting and Quirk, 2019), but as relationships based upon trust (Schulz, 2019), following a humanist ethos (Fujii, 2018) and geared towards cooperation, collaboration and mutual reciprocity (Pittaway et al., 2010). Such processes can be expected to facilitate a shift of the locus of power from the researcher towards research participants that results in knowledge production geared more towards the terms of 'the researched' (Robins and Wilson, 2015), and that helps to elevate local perspectives beyond local concerns (Bunting and Quirk, 2019).
Finally, all of these considerations can also carry far-reaching implications for the type of research we conduct and the knowledge we produce. If we acknowledge and accept the researchers' (temporary) position of only limited power and control -and that of research participants as active agents -we can come to terms with the fact that the findings and claims we put forward can only ever be limited and offer partial snap-shots into complicated lived realities of war, conflict and violence, rather than constituting universally applicable, all-encompassing and generalizable insights.

Conclusion
In this article, I have sought to de-and re-construct relationships between researchers and research participants within the context of (post-)conflict settings. By drawing upon personal research experiences from Northern Uganda, I have illustrated how key-informants' identities in post-conflict zones cannot uncritically and universally be relegated to the presumed ever-vulnerable respondent in need of the researchers' protection. Simultaneously, researchers cannot always and uniquely be in control and navigate their respondents. Instead, research participants frequently exercise control over the researcher and thus over research processes at large -for instance when it comes to the provision of security while in the field. As argued by Björkdahl and Selimovic (2015: 169), 'both the researcher and the researched are subjects with agency' -albeit with differing power statuses. My application of reflexivity in this article has shown how in a particular moment of post-conflict insecurity, a key-informant exercised control over me, with wide-reaching implications about the research relationship at large. Recognizing research participants' fluid and multiple positionalities thereby contributes to a process of re-calibrating power relationships inherent to research processes, and carries wider implications for research processes and the politics of knowledge production, beyond neo-colonial and essentialist conceptions of research relationships, power and agency.