The space between us: feminist values and humanitarian power dynamics in research with refugees

ABSTRACT International humanitarian and development agencies striving to promote gender equality and women’s empowerment sometimes neglect to recognise the power hierarchies present in their own engagement with communities. Drawing on research on Syrian refugees and humanitarian workers in Jordan, this article explores the research and monitoring and evaluation practices of international humanitarian agencies. It suggests that the emphasis on generating evidence has resulted in more transactional and less relational engagement with refugees. This paper asks how feminist values can inform research with refugees, and explores how these values may provide less-extractive ways of engaging with displaced populations.


KEYWORDS
Feminist values; international humanitarian agencies; refugees; power dynamics; research; monitoring and evaluation Introduction Within a humanitarian refugee response, many agencies include activities that aim to promote gender equality and women's empowerment. But although these activities focus on addressing the power hierarchies affecting women's lives, international humanitarian agencies have reflected less than they need to on the power relations they themselves perpetuate through their research and monitoring and evaluation (M&E) practices with refugee populations. Based on research on Syrian refugees in Jordan and humanitarian workers, and informed by my own experiences as a international aid worker, 1 I argue in this article that some of the original motivations of humanitarianismincluding notions of solidarity and proximity to communitieshave become less of a priority than the drive towards generating evidence and data.
My research focused on power hierarchies within the research and M&E practices of international humanitarian agencies in relation to a refugee response. I suggest here that humanitarian practice has deviated from its original motivations and principles. The pressures to deliver on gender analysis, justify interventions, and ultimately demonstrate impact, can result in extractive relationships with refugees. In the drive to gather data, interactions with refugees can become clinical and at times exploitative, focused on getting the greatest amount of information in the least amount of time. Instead of an open approach to asking refugees about their experiences, the questions asked by humanitarian agencies during research and M&E may be laden with assumptions about refugees, and infused by preset, topdown humanitarian agendas.
I suggest here that feminist analysis can offer insights into power imbalances between researchers and refugee communities, and research informed by feminist values can offer potential to redress them. But it is important to note that proximity, reflexivity, and ensuring more open approaches to research and M&E are not new or unique to feminist research. Academic literature has long critiqued and problematised the power hierarchies within humanitarian agencies, arguing for a greater awareness of how power shapes interactions with refugee communities. 2 Similarly, humanitarian (and development) agencies have a long history of using participatory and relational approaches to engaging with communities and privileging their knowledge. 3 However, I argue that the drive for data and evidence has resulted in slippage between these principles and the practices of humanitarian agencies.

Feminist research
Feminist research aims to create 'useful knowledge that will make a difference to women's lives' (Letherby 2003, 4). It seeks to challenge the intersecting power hierarchies that negatively affect women, which makes it distinct from other approaches to research (Brooks and Hesse-Biber 2006). Conducting research that is informed by feminist values is about more than merely seeking to understand the oppression of women. It extends to the process of understanding itself, specifically to how power dynamics affect the way research is conducted. Feminist research draws on other critical approaches to research, situating issues of methodology, and concepts like reflexivity and reciprocity, within a broader focus on addressing unequal power dynamics.
In feminist research, the researcher is explicitly recognised as someone who is subjective. Notions of neutrality, objectivity, and the idea of finding 'truth', which tend to characterise traditional research (as well as research conducted by humanitarian agencies), are seen as inherently flawed (Brooke and Hesse-Biber 2006, 14). Instead, the process of producing knowledge is positioned as complex and influenced by the values and backgrounds of those conducting research. Research informed by feminist values recognises that the motivations and position of the researcher mattersand can even add to the research. The practice of 'reflexivity', which focuses on the question of how knowledge is produced, is a way of reflecting on the research process. Feminist researchers use reflexivity to consider not only position and background, but also the power hierarchies that influence knowledge production throughout the research process (Wickramasinghe 2010).
Research methods which are participatory and flexible also play a role in addressing power imbalances between researchers and those being researched (Liamputtong 2007). This is less of a challenge in much humanitarian research, where agencies commonly incorporate participatory methods such as focus group discussions; however, the way methods are used is also relevant for feminist research. For example, open-ended questions that allow people to express their views, instead of purposive questions where the researcher's assumptions and views are privileged, can be a way of challenging inequalities in knowledge production processes. Refugee researcher, Liisa Malkki, emphasises relational practices of engaging with refugees. This includes listening to what is important to participants, 'leav[ing] some stones unturned' and 'not prying' (Malkki 1995, 51). Through this approach, she suggests, the focus is not on establishing the 'facts' like a 'detective' would (ibid.), but rather the relationship with participants is what matters most.
In being relational, research informed by feminist values also requires attention to the principle of 'reciprocity'. Reciprocity is based on the idea that research should benefit both researcher and research participants, that it involves 'give-and-take' (Huisman 2008, 374). This principle is particularly relevant for work in communities where reciprocal social relations have cultural importance, as in the Middle East region (Deeb and Harb 2013, 21), and for feminist research, which recognises that reciprocity is particularly critical because of power imbalances between the researcher and research participants. Giving something back to participantswhether it is time, relationship, information, or even something materialcan be a way of reducing the power imbalances with research participants. These descriptions of research informed by feminist values may sometimes contrast with the research and M&E practices of humanitarian agencies. This includes research and M&E focused on gender equality and women's empowerment, which should, in theory, reflect greater awareness of power hierarchiesbecause this is what such work is seeking to address.

Conducting research differently: methods and approach
My doctoral research explored humanitarian narratives on gender norms among Syrian refugees in Jordan. I conducted research during a nine-month period from 2016 to 2017 in three cities (Amman, Irbid, and Zarqa) with Syrian women and men. I sought to understand their day-to-day experiences in Jordan as well as their lives in Syria through a range of methods including: participatory photography; semi-structured interviews; lifestory interviews; and participant observation. I also conducted semi-structured interviews with ten international and local humanitarian workers from different international nongovernment organisations (NGOs), United Nations agencies, and local organisations. These individuals had worked, or were currently working, in Jordan, on gender equality and women's empowerment issues. This article draws heavily on the reflections of these humanitarian workers, particularly their reflections on research and M&E.
As I began my research, I knew already from my experience as a humanitarian aid worker that there is often a lack of awareness of powerand/or a lack of commitment to challenging thisduring the process of engaging with communities. Instead, there is an overriding concern with producing data and evidence. As 'value for money', efficiency, and impact have become more important to humanitarian agencies, it may be that awareness of power imbalances within interactions with communities has slipped from focus.
Recognising these problems associated with how refugees are sometimes engaged by humanitarian actors, I sought to conduct my own research differently. I set out to be less restrictive, to prioritise listening, and to spend time with refugees instead of rushing into questions (Malkki 1995). This was at times a challenging process, particularly in terms of balancing relationships with pressures to complete research outcomes. For life stories, I met some individuals several times over a few months. I chose photography as one of my research methods, seeing it as a means for Syrian refugees to understand their own lives as well as a means of challenging power hierarchies between researchers and participants (Packard 2008). Photography workshops were also a way of 'giving back' to refugees, enabling reflection on topics of interest to them (whether positive or negative), as well as helping them to develop photography skills. During the workshops, which were conducted through local organisations, I built relationships with Syrian women and men.
Through the research process, I tried to reduce the distance between refugees and myself, as well as between refugees and my research assistant (Liamputtong 2007). Conducting feminist research meant letting refugees talk about what was important to them, and even sharing information about myself. I spent time eating with refugees, cooking, and shopping. Deviating from a more clinical approach to engaging with communities, I saw myself as researcher with my own biases, who had a vested interest in understanding the challenges of refugees and who wanted their stories to be represented in all their complexity. I grappled with the challenges of representation throughout the research process. This included how I represented myself: an Australian-educated researcher of Sri Lankan origin, who had previously worked for international agencies, and who was at times mistaken for a domestic worker during fieldwork. I also struggled with how best to represent the lives of my participants in all their complexity.

My research context
The conflict in Syria began in 2011. Since then, approximately 5.7 million Syrians have been registered as refugees in the surrounding countries of Jordan, Turkey, Lebanon, Iraq, and Egypt (United Nations High Commissioner for Refugees (UNHCR) 2019). A further 6.1 million people have been internally displaced (UNHCR 2018). Over half a million Syrians have been killed in the conflict (Reuters 2018). Jordan has a long history of hosting refugee populations, including 2 million Palestinians, some 500,000 Iraqis, as well as Sudanese, Somali, and Yemeni refugees (Nusair 2013, 59;UNRWA 2016). In Jordan, approximately 670,000 Syrians are registered as refugees and between 80 and 85 per cent live outside refugee camps (UNHCR 2019).
The humanitarian response to the Syria Crisis in Jordan is similar to other humanitarian emergencies in terms of the power hierarchies involved in the delivery of humanitarian assistance. The concept of 'vulnerability' is often important in defining who receives services and who does not (Agier 2011, 213) and this is particularly true for the Syria Crisis. Refugees are subject to external decision-making processes, and their lives are often characterised by uncertainty about their future in the host state and uncertainty about decisions from the aid bureaucracy about their assistance (Horst and Grabska 2015, 10).
Issues of funding are particularly critical to humanitarian power hierarchies, because of the power held by donors. Donors have, however, tried to engage with power hierarchies by encouraging localisation 4 efforts. Since the 1990s, donors have needed not just to provide funds, but are also expected to co-ordinate activities swiftly in the face of humanitarian emergencies (Macrae et al. 2002). Now, amidst greater focus on efficiency and 'value for money', donors demand that humanitarian agencies prove their interventions have impact (Anderson et al. 2012, 45). This has resulted in the streamlining and standardisation of reporting mechanisms, creating challenges for humanitarian agencies which have to fit into the frameworks that donors require (Anderson et al. 2012). These approaches sometimes prescribe narrow ways of understanding changewhich some argue comes at the expense of critical analysis (Eyben 2013).
Humanitarian assistance, although initially taking the form of the provision of lifesaving immediate assistance for displaced populations, such as shelter (camps), food, water, sanitation and hygiene, and medical services, has evolved to include a response to social issues. It has become commonplace for humanitarian responses to include activities that are seen as bringing about progress, including introducing 'modern' ideas like gender equality among the displaced (Grabska 2011). This, suggests Jennifer Fluri, is because life as experienced by those that require assistance, is viewed as 'bare life' that needs intervention (2012, 37). The 'emergency' inherent within a humanitarian response has thus come to include a more development-focused component, which can often be around promoting gender equality and women's empowerment.
The specific outworking of humanitarianism within Jordan is important to consider. In Jordan, the Syrian Crisis has created jobs for Jordanians and Palestinians within the humanitarian sector. The Jordanian government's varied treatment of refugees in Jordan (especially Palestinians, who may now find themselves working to help Syrian refugees as part of the humanitarian response) creates a complex environment where it may appear that certain refugees count more than others, and where it may be perceived that there are benefits to being Syrian refugees in a way that was denied for other groups. This has consequences for how local humanitarian staff, as well as international staff, engage with Syrian refugees.
The meaning behind 'responsible action' 5 towards refugees becomes relevant here as there may be perceived limits about what is acceptable and ethical behaviour (Brun 2016, 406). The 'hospitality codes' inherent in both Syrian and Jordanian culture also carry implications for humanitarian actors, who often transgress these codes in how they engage with refugees, especially in their intrusions into people's homes (Wagner 2016). While Jordanian and Palestinian humanitarian staff should be aware of the behaviours expected of them as 'guests', it is likely that their perspectives are less visible in the hierarchal structures of humanitarian agencies, which tend to rely on international experts (Chatty 2017, 28).

Research, M&E, and the meanings made of data
The bias towards counting Within humanitarian agencies, gathering evidence has become increasingly important. Evidence must show that interventions are making an impact. This includes the sometimes-relentless donor requirements for the so-called 'beneficiary count', which is driven by the idea of reaching as many people as possible. While data in itself is useful for enabling decision-making on programmes, the fixation on dataparticularly quantitative datacan become problematic. Voices in my research suggested that the drive to get data can divert limited staff resources that might otherwise be spent focusing on implementation. As one humanitarian worker put it: '[I]nstead of quality work, you are trying to get the numbers' (interview, Amman, 17 February 2017). Counting people is an important way that agencies demonstrate 'value for money'. Jennifer Hyndman has suggested that the act of counting people has 'highly political objectives that relate more to organisational aspirations than staff or "client" welfare' (1996,238).
For example, in many humanitarian emergencies, significant time is being invested in gathering data on gender-based violence (GBV). Obviously this reflects a positive shift towards recognition of GBV as a human rights abuse that needs to be ended. Yet technical guidance explicitly states that GBV does not have to be proven with data and should be assumed to be always occurring in every context (Inter-Agency Standing Committee 2015 , 2). Enabling disclosure of GBV has, therefore, become importantperhaps disproportionately so, at the expense of other aspects of gender inequality and marginalisation that require attention from donors and humanitarian actors. In addition, in line with my point in the last paragraph, when data are given this level of focus, research and M&E can morph from being merely part of the process, to being the outcome itself. In fact, data should just be the start in getting better programming to address issues that arise from it.
In Jordan, one humanitarian worker drew attention to how data within the Syria Regional Refugee Response was sex-and-age disaggregated more than any other emergency she had worked in previously; however, this somehow 'let everyone off the hook' because this became the endpoint (interview, Skype, 26 May 2017). There was little linkage between the data and improving programming.
The 'urge to represent the world through quantification' (Brun and Lund 2010, 822) is a result of the value placed on objectivity. Numbers seem to contain 'a particularly reliable form of truth' (Merry 2016, 26). This can become particularly problematic: complex social and cultural issues do not easily lend themselves to quantification. When donors and humanitarian agencies disproportionately focus on quantitative data, qualitative data may be devalued. Cathrine Brun and Ragnhild Lund describe how the agency that they were conducting research for kept asking for the 'real facts' because 'people's experiences were not considered valuable knowledge ' (2010, 822). This is similar to my experiences working and consulting for international agencies. Qualitative data may not be seen as 'real' knowledge, and rather be perceived as merely 'stories' (Malkki 1996 , 385) that cannot be verified. When this kind of data is preferred by humanitarian agencies and the donors who often drive organisational priorities, contextual aspects may be neglected. At times, refugees may feel compelled to participate in the data collection activities of humanitarian agencies, who they may perceive as having the ability to bring (or not bring) practical benefits (Kaiser 2004). After data collection, refugees do not always hear back on the research or M&E outcomes.
The meaning attached to M&E: is it 'research'?
Sometimes this may be because M&E is not perceived by humanitarian agencies as 'research' as such. With international agencies feeling they have to prove that programming has impact, it can be less important for them to learn how refugees adjust to change, or the strategies they use to cope. Instead, the approach can be extractive, focused on information that will make the case for more funding; or worse, information in which the donor has exhibited an interest. It is during this process that refugees become the 'objects' rather than the 'subjects' of humanitarian activities (Hyndman 2004, 203). Humanitarian agencies (or donors) decide what issues are important to understand, they define the scope of research, and they carry it out. In the drive for data, the participatory focus that once characterised work with communities may slip from priority (Chambers 1997). This is especially concerning from a feminist perspective. It means that what is analysed may not reflect issues of actual concern to refugees, but instead represents what others deem as relevant based on their own assumptions. One humanitarian worker commented on the 'preconceived ideas' humanitarian workers bring: 'We are the experts, we know better. So, we don't have humility to really understand' (interview, Skype, 26 February 2017). For her, the issue was that humanitarian workers don't spend enough time talking to communities. Instead, she argued: [W]e are just churning out these beautiful documents, so it catches, you know, the donor's attention, the media attention, the people's attention and we just generate the money. But it's not really learning …

Generalised assumptions and universalised remedies
The humanitarian worker quoted at the end of the last section also commented that she had never seen a gender analysis or rapid analysis that sought to 'actually truly understand the people'. Instead, the focus of analysis is problem-focused: 'It's more of, you know, what is broken?' (interview, Skype, 26 February 2017).
The focus on what the problems are results in agencies asking purposive and narrow questions which may be more about verifying what they already expect to be the main issues, based on other emergencies. In the case of the Syrian conflict, this includes statements about rising early marriage (Save the Children 2014). Increased domestic violence and sexual harassment also feature in humanitarian analyses (Oxfam and ABAAD 2013). Women are depicted as suddenly having to be responsible for the economic burdens of the family (International Rescue Committee 2014).
These narratives often echo analysis from other emergencies, resulting in 'one-size-fitsall universalising remedies' (Cornwall and Rivas 2015, 397) that do not always fit the context. These more sensationalist issues may take precedence in the research and M&E of agencies, rather than what refugees wish to discuss. In my experience, the latter includes issues like access to education or health services, and resettlement options. Feminist research, in contrast, is not about proving a hypothesis; rather, participants drive the knowledge that is created.
Women are often characterised in public fundraising as the 'most vulnerable' in crises. It has been said that the 'dark side' of people's experiences seems to interest humanitarians (Fluri 2012, 45). This is due to the humanitarian focus on identifying problems and solving them. It perpetuates perceptions of refugees as vulnerable and in need of assistance (Fassin 2012, 21), which in turn justifies the need for humanitarian agencies, and helps them raise funds. In the narratives of humanitarian agencies, it is sometimes the most extreme stories and exceptional cases that receive focus (Abu-Lughod 2013, 78). In contrast, feminist research is intersectionalrequiring analysis of the multiple, intersecting power hierarchies that shape people's lives. It challenges ideas of women as always and eternally 'vulnerable'.
It is important to note that early marriage, domestic violence, sexual harassment, and sexual exploitation are real challenges for some Syrians; however, they are not the only narrative. They describe the experiences of some, but not all, Syrian refugees. There is more going on beneath the surface than these dominant narratives suggest: more complexity and nuance, including around historical context (Lokot 2018). Understanding the context surrounding refugees is perhaps seen as too time-consuming (Hyndman and de Alwis 2003, 213) in an industry that seems driven by the need to make complex issues easily understood. This means agencies resort to references about Syrian 'culture' and 'tradition' terms that inevitably appear in analysis, but without further detail on what this actually means (Women's Refugee Commission 2014, 14). Invoking 'culture' is sometimes a substitute for actual analysis grounded in people's life experiences. It results in generalised assumptions about communities, instead of a feminist approach that is grounded in people's experiences.
Issues that 'trend': transforming knowledge production Another humanitarian worker reflected on the way agencies learn about the issues facing refugees. For her, this seemed to be about what was 'trending' at the time: [S]omeone will say something that no-one has heard in a meeting, and then they all become kind of obsessed with that. And then they all focus on that for a while, rather than like an organic, from the refugees' themselves, needs. (Interview, Skype, 26 May 2017) She described this as like 'an infection' that begins in meetings, from people who are distanced from refugees and who don't have a sense of what their actual experiences are. From these conversations among higher-level staff, who may not even actually speak to refugees, ideas spread about refugees. This is how knowledge may, at times, be produced.
This humanitarian worker also reflected on the way humanitarian workers talk about 'the field' and how those who were based in the capital city and who had never even met a refugee would speak about and for refugees as if they intimately knew them. For these staff, being in an office was being in the 'field' despite the fact that the organisation's activities occurred in other geographical areas (interview, Skype, 26 May 2017). Another humanitarian worker reflected on distance between decision-makers and the 'field': I remember when I was field-based, going to these working groups and I could've said anything and they would've lapped it up. Because they never went to the field. (Interview, Skype, 23 June 2017) This lack of proximity to refugees has multiple consequences. Instead of understanding the day-to-day experiences of refugees based on actual research, old knowledge from other contexts and anecdotal facts about refugees may dictate programming. Proximity is not needed because 'we know' what they need. As such, relational, participatory approaches to understanding people's lives are not seen as necessary.
Another humanitarian worker highlighted that part of the challenge around understanding gender norms is the fact that it takes time to understand these norms in a humanitarian emergency, yet there is pressure to respond quickly. In the case of the response to the Syrian conflict, she said that they did not have Syrian colleagues who could advise on cultural and other issues, so they had to go with what they 'knew' from similar contexts and from rapid assessments (interview, Amman, 1 May 2017).
In some cases, as other humanitarian workers explained, knowledge about Syrians comes from Jordanian staff. This can be problematic because it assumes Jordanians know about Syrians, or that the differences between Jordanians and Syrians are minor. In my research, I found that Syrians felt strongly that this was not the case (interview, Amman, 17 November 2016). One humanitarian worker also emphasised that this approach limits the knowledge production to a narrow subset of Jordanians: to those of a certain educational level and perhaps economic status, who are in a position to be asked for input by agencies (interview, Skype, 1 August 2017). How higher-class Jordanians interpret the lives of Syrians may be different to how Jordanians from other socio-economic statuses might reflect on this.
The humanitarian worker who reflected on the pressures of implementing also commented: 'It's better than not doing anything. We have to deliver' (interview, Amman, 1 May 2017). For her, the problem was that agencies failed to 'dig' after the initial response. This idea of reflecting later is, however, difficult in practice. In a protracted humanitarian crisis, the point at which the situation is sufficiently under control in order to 'reflect' is difficult to gauge. This is particularly true for an industry that has often consigned 'reflection' to the point when the activities are subject to a formal evaluation, which could be years later. One humanitarian worker felt the problem around lack of reflection was systemic: '[T]he culture of humanitarianism doesn't suit itself well to thinking, reflecting, or analysing. It's all about just doing' (interview, Skype, 26 May 2017).
That a deeper analysis has been substituted with stereotypes and assumptions is, I suggest, strongly linked to the humanitarian bureaucracy itself. In the increasingly bureaucratic humanitarian machinery, 'humanity' may slip out of focus (Waters 2001, 44). Bureaucracies by nature are focused on efficiency and professionalism, which means that other aspects linked to the 'humanity' of a humanitarian response, e.g. ensuring people's dignity or respecting the perspectives of people being served, may not be prioritised. Within narratives that are based on being as efficient and professional as possible, results become more important than relationships: '[T]he desire to measure places a premium on numbersfor instance, lives lost and saved, people fed, children inoculatedto the neglect of non-quantifiable goals such as witnessing, being present, conferring dignity, and demonstrating solidarity' (Barnett 2011, 16). Whilst the latter descriptions in this quote reflect feminist values, the focus on generating data and being efficient may cause these values to slip from focus.
Unlike other businesses or entities that involve serving, '[a]n aid agency does not need to receive the approval of aid recipients to continue to receive donor funding ' (Anderson et al. 2012, 37). Refugees and humanitarian workers are 'trapped' within 'asymmetrical relationships', where accountability is 'skewed' towards donors instead of refugees themselves (Harrell-Bond 2002, 53). The limited opportunities for refugees to provide feedback to agencies serving them is, itself, disempowering. They are the 'beneficiaries', but the decisions are made by others.
The lack of respect and care for refugees may also extend to how their time and space is valued. Calling a refugee at short notice to say that you would like to conduct a quick 'home visit' with donors in a few days is perhaps less appropriate within the Middle Eastern context of hospitality, building relationships, and exercising reciprocity. But it reflects a lack of awareness of the power held by agenciessomething that feminist approaches try to address. During my research, refugees shared experiences of being visited by various researchers for international and local agencies, and their confusion at the assessments in which they were asked to participate (interview, Amman, 14 November 2016; interviews, Irbid, 10 and 11 April 2017). Many never heard back after answering numerous questions designed to gauge their eligibility for interventions like cash assistance, yet remained hopeful that they would be contacted, even several months after the visit.
This disconnect from the issues facing refugees, the disregard for contextual and relational factors, and the inability to empathise and show respect, does not occur in a vacuum, and may be a consequence of the way power operates within large bureaucracies. This distance flows into how research and M&E occurs: power hierarchies within agencies affect how humanitarian staff engage with communities.
I have myself been in meetings with 'gender specialists', where power hierarchies were so entrenched that the senior manager interrupted and literally shouted over her subordinate staff who were trying to share their opinions. The lack of self-awareness and using power in such negative ways, while attempting to implement complex programmes that seek to address unequal power, reveal a startling layer of hypocrisy. It helps to explain the ways in which refugees are treated by humanitarian staff, as well as the distance between humanitarians and the populations they serve.

Moving forward: proximity, reflexivity, and greater openness in research
What can help unravel these power dynamics? My research points to the presence of systemic barriers that affect research and M&E processes in humanitarian agencies, including the drive for data generation, pressures to respond quickly, as well as the humanitarian bureaucracy itself. These barriers result in assumptions and stereotypes about refugees being relied upon, and feeds into the distance between refugees and humanitarian workers. How might we move forward from here? What can feminist values offer? Proximity First, there is a need for humanitarian workers to be more proximate to refugees. Reflecting broadly on how refugees are treated by humanitarians, one humanitarian worker I interviewed said the problem was that she felt humanitarian workers failed to put themselves in the shoes of the refugees they are serving. She reflected on the principle of 'proximity', or being 'close' to refugees and showing solidarity through spending time with them. For her, proximity meant that 'you saw them as human beings'. This broke down the 'divide of us versus them' and resulted in a more 'organic' humanitarian response. She added that, in Jordan, this kind of approach was mostly used by small solidarity organisations, laughing as she added, 'They weren't very good at writing proposals, you know, like their monitoring and evaluation was all over the place, but they did seem to care and like the refugees!' (interview, Skype, 26 May 2017).
Daniel Wordsworth, CEO of the American Refugee Committee, writing about this idea of proximity, relates it to the participatory approaches to development and humanitarianism in the 1990s inspired in part by the writings of Chambers (1983): 'Concepts like immersion and closeness to the poor were considered badges of honor; you had to live in villages, walk the same steps to collect water, and sit by the campfire at night' (Wordsworth 2017). He contrasts this with how agencies work with refugees today, arguing that they are led by 'technocrats' who 'laud distance, objectivity, regional compacts, and systems-wide approaches' (ibid.), arguing that the original spirit behind humanitarianism has been lost. These insights are all relevant to researchers in humanitarian settings.
As a potential solution, Cathrine Brun has suggested that an 'ethics of care' approach might be useful for humanitarian agencies (2016,404). This approach shifts the humanitarian worker (including M&E teams and other researchers) from 'a distant, detached observer who creates knowledge based on the application of standardised models rather than on experience in the particular local context' (Olivius 2016, 280), to someone whose actions are grounded in trusted relationships with the populations being served.
Approaches like this draw on feminist values around building relationships and showing solidarity. They require spending time with refugees to actually understand the issues they face, instead of assuming. They require dismantling the power hierarchies that consign refugees to being called 'beneficiaries' (Hyndman and de Alwis 2003, 218). One humanitarian worker described how she tried to challenge such power dynamics: I remember being in Jordan and telling the team there that these people that you're going [to] in the camp, and I'm looking down upon, and you find that the questions are so irritating at times … Without them, you would not be in this job … [T]hey are paying your salary … And I want you to feel the fact that you have been given this privileged opportunity to speak for them, to advocate for them. They are paying you for it. (Interview, Skype, 26 February 2017) This humanitarian worker felt it was important for her team to see power differently; to recognise to whom humanitarian staff were actually accountable. She explained what this means in very practical terms: It's being in a very eye-level with each other, sitting on the floor, going into their home, inviting the others. And sitting in a round, the same way that they … until they feel comfortable with you. (Interview, Skype, 26 February 2017) Sitting at 'eye-level', being proximate, is uncomfortable. It is not about going to see the refugees as if seeing a spectacle, but it means going to understand and learn. It means 'bearing witness' to the suffering of others despite our own discomfort (Feitlowitz 2011, 58). It means showing empathy and compassionwhich may at times be at odds with humanitarian agency language around efficiency and 'value for money'. It may be easier to focus on the mechanics of implementation instead of the fact that these are real people's lives. One humanitarian worker discussed the way refugees would protest at the UNHCR office, trying to get UNHCR to take action on their cases. She commented on the behaviour of humanitarians towards these refugees, saying: [I]t becomes almost like the homelessyou just push past them without looking at them … this walling off and … lack of empathy is also like a protective mechanism. (Interview, Skype, 26 May 2017) Being in proximity to the pain, uncertainty, and injustice experienced by others is not comfortable, but it is a critical starting point to unravelling power dynamics between humanitarian agencies and refugees.

Reflexivity
Practising reflexivity is part of drawing upon feminist values in conducting research and M&E. This includes how our education, previous work experience, race, age, gender, and economic status influence the way we think about communities that we serveand the way we do research and M&E. Taking on feminist values encourages humanitarian workers themselves to engage in greater reflection on their own values, motivations, and backgrounds, and how these affect research and M&E processes. While I have encountered many humanitarian workers in the sector who do just thistogether with others informed by commitments to participatory methods and a commitment to putting 'the last first' (Chambers 1997, 1) there are many complex reasons that make these commitments challenging to enact.
This need for reflexivity is perhaps particularly relevant for humanitarian agencies working with refugee populations; forced migration itself is 'neither passive nor apolitical' (Hyndman 2000, xv), rather it 'cries out for moral positioning' (Chatty 2010, 1). Research and M&E among displaced groups cannot continue to be viewed as neutral or objective. It is shaped by who researchers and humanitarian agencies are, including the pressures they face to advocate and raise funding. The feminist approach to understanding how knowledge is produced, and why it is produced, may enable shifts in how data analysis by humanitarian agencies is conducted and valued.

Conducting and presenting research and M&E with greater openness
Third, I suggest that there is a need for humanitarian agencies to bring a greater openness to research and M&E processes; from the development of questions, to the kinds of methods used, to the way analysis is conducted and presented. This requires recognising (again) the value of qualitative, participatory research that prioritises relationships with refugees (Rodgers 2004, 49). It means interviews shifting from being 'too fast, too purposive, or much too short' (Ghorashi 2008, 118) to being more open. It means taking time to listen, allowing refugees to discuss what is important to them, and then presenting their stories with care. Trinh Minh-ha reminds us of the importance of this unrushed feminist engagement: Never does one open the discussion by coming right to the heart of the matter … To allow it to emerge, people approach it indirectly by postponing until it matures, by letting it come when it is ready to come. There is no catching, no pushing, no directing, no breaking through, no need for a linear progression which gives the comforting illusion that one knows where one goes. (Trinh Minh-ha 1989, 1) Towards the end of my fieldwork in Jordan, my research assistant and I were speaking to one Syrian woman. She speaks some Englishshe understands a lot but is less confident in speakingso we always did the interviews in Arabic. While my research assistant was translating, this participant suddenly burst into tears. She said to my research assistant, 'I want to thank you … Because you are really translating word for word' (interview, 27 March 2017). She explained that while at an embassy applying for reunification with her husband, she was giving an account of her experiences to the embassy officials during an interview, but the translator was only partially communicating what she said. She understood English enough to know her story was not being fully translated. The power hierarchies she was a part of meant that her experiences were treated in a specific way, as something to be summarised and simplified. For her, the fact that my research assistant carefully translated every word meant something significant; it meant the representation I received in English was not a summarised, simplified version.
Within humanitarian agencies, research and M&E processes can affect the analysis and presentation of data. Instead of people's experiences being presented with complexity, their contributions may be summarised, simplified, and generalised for ease of communicating an advocacy or funding message. Humanitarian agencies often focus on the negative: on the key challenges faced by communities and the solutions they plan to implement to address them. This approach to understanding people's lives may cause communities to think that 'problems' are what people want to hear about. In my research, a Syrian woman was speaking to me in a room with a few other women, prior to a research feedback session. She paused during our interchange, pointed to her friend sitting nearby and said, 'You should interview her. Her life is even worse than mine' (Irbid, 10 May 2017). Her assumption was that I wanted to hear a terrible story.
Challenging this approach to research, Lila Abu Lughod warns: 'Superficial vignettes and extreme cases tell us little about the variety of ways women experience their lives and the contexts we must appreciate in order to make sense of their suffering ' (2013, 78). Anecdotal, stand-alone examples that are presented without context may 'flatten the three-dimensional lives' of people, shortening their stories into 'tidy case studies' devoid of complexity (Lindisfarne 2000, 124). The use of simplistic analysis, although attractive (Rosling et al. 2018), may be a product of the marketing/public-facing language that presents interventions that are able to resolve problems quickly. When it comes to issues related to gender equality, this approach makes gender inequality seem like a simple issue, merely 'an expression of under-development' that can be 'remedied by the introduction of international human rights norms' (Olivius 2016, 272). In actuality, 'it is not so easy to talk about "patriarchy" or to put one's finger on how power works' (Abu Lughod 2013, 6).
Engaging in research and M&E that is informed by feminist values may mean unravelling the common narratives to understand what occurs beneath the stereotypes and simplified summaries. Generalisations do make things easier when designing interventions, but they miss complexity. Our work must reflect the complexities involved in talking about power. We must be transparent in communicating the limits of our analysis, including what we do not know.

Conclusion
During humanitarian responses, research and M&E processes do not always benefit refugees themselves, and are not always doing as much as they could to support the empowerment of refugees and present their own views and analyses as actors with important knowledge to share. This article argues that feminist values may offer rich and useful insights for engaging with refugee communities within research and M&E processes. These feminist values also need to inform the research methodologies and the decisions about how research is presented and used.
Drawing on the experiences of humanitarian workers in the context of the Syria Crisis, this article argues that the way power operates within humanitarian bureaucracies affects how humanitarian workers engage with refugees. It urges a return to the principles of proximity and reflexivity, as well as greater openness in research and M&E processes. Now, perhaps more than ever before, there is a need for critical reflexivity on power, humanitarianism, and the space between refugees and humanitarian staff. When research and M&E are less extractive and less purposive, it may be possible to uncover different, unexpected narratives. Notes 1. I have managed GBV programmes, worked in technical advisory roles on gender equality and adolescent girls, and also conducted research as a consultant since 2008. My work experience has primarily been with international NGOs, including longer-term field postings in Jordan, Nigeria, and Burundi. 2. This includes work by Liisa Malkki (1996), Graeme Rodgers (2004) Chambers (1997). 4. 'Localisation' refers to the process of ensuring a more locally led humanitarian response, where power and funding is not solely concentrated in international agencies. As an agenda it achieved particular prominence at the World Humanitarian Summit in 2016. For further information, the 'Shifting the Power' project has a useful resource: 'Localisation of aid: are INGOs walking the talk?', which is available at https://startnetwork.org/resource/localisation-aid-are-ingos-walkingtalk (last checked 28 August 2019). 5. Cathrine Brun suggests that 'responsible action' is viewed narrowly within humanitarian agencies, urging an expansion of what is considered ethical behaviour and a greater focus on interpersonal relationships between humanitarian staff and refugee communities (2016,406).