The ethical performance of access and consent in ethnographic research on social work encounters with migrant-background service users

In this article, we contribute to an emerging body of literature concerning the often-overlooked topics of access and consent in research. We posit our understanding of access and consent as continuous ethical reflection and negotiation, conceptualised here as ethical performance, which is particularly valuable in research in institutional contexts defined by numerous power asymmetries. We draw empirically from research on street-level institutional encounters between social work practitioners and migrant-background service users in the Helsinki capital region. Access in this research was a multi-stage process including various stage-related negotiations, and the previous stages always influenced the stages that followed. Nevertheless, access and consent were always erratic and subject to revision. We describe how the need for ethical reflexivity arises in various concrete, often unpredictable, situations, and argue for the importance of paying explicit analytical attention to negotiations regarding access and consent.


Introduction
How to access the empirical research site to gain information on a certain topic of interest is a central question in any empirically driven social research (Harrington, 2003), but only recently has access become an issue of more explicit concern. Traditionally, access has been conceptualised as a tactic and something that takes place outside or before research (Cipollone and Stich, 2012), implying that researchers could in rather unproblematic ways make decisions concerning it. In this article, we discuss the ethical challenges involved in negotiating access and consent in research through an explicitly analytic approach. We draw on an ongoing research project on street-level institutional encounters between social work practitioners and migrant-background service users in the Helsinki capital region.
We contribute to this emerging body of literature by examining negotiations about access and consent as a continuous ethical endeavour in ethnographic research. We follow Crowhurst (2013) in her view of negotiations about access as dynamic and transformative processes in which the research is constantly reviewed and made sense of (also Borrelli, 2020;Kara, 2016). Indeed, Wanat (2008) has proposed a distinction between gaining access to conduct research and continuously earning the cooperation of participants to produce empirical material (also Bondy, 2012). Research can therefore be viewed as mutual negotiations on collaboration where 'maintaining access' or 'earning collaboration' becomes as crucial as achieving access in the first place (Kennedy-Macfoy, 2013;Plankey-Videla, 2012;Wanat, 2008).
We consider continuous ethical reflection and negotiation about access to be particularly valuable in ethnographic research in institutional contexts when dealing with vulnerable service users whose service encounters are defined by numerous power asymmetries. We therefore argue for the need to expand the gaze by jointly analysing the negotiations of access and consent as continuous ethical performance. Furthermore, we assert that the encounters and processes involved in negotiating access and consent can be understood as research data (Bondy, 2012;Borrelli, 2020;Brown-Saracino, 2014;Reeves, 2010). Similarly, we draw attention to how the need for ethical reflexivity actualises in various, often unpredictable, situations in research (Campbell et al., 2006;Crowhurst, 2013).
In this paper, we first discuss different approaches to access and consent proposed by earlier ethnographic research, and thereafter introduce the research project that underpinned our analysis. We then proceed to present our analytical approach and our findings. Finally, we reflect on the role of ethics in ethnographic research encounters involving access and consent.

Configuring access, consent and ethics in ethnographic research
Discussions on access often start with the figure of the gatekeeper, referring to people who have influence over who is invited to participate in research and the prior information that potential participants receive about the research and the researchers (Reeves, 2010). Gatekeepers can be viewed as facilitators and collaborators in research, and they may provide support for issues the research participation might stir up (Kara, 2016;Notko et al., 2013;Turnera and Almacka, 2017). Yet the relationship between gatekeepers and researchers can be tense, and researchers may perceive gatekeepers as barriers or even adversaries when access to participants and the research site is strongly mediated by and dependent on them (Kendall et al., 2007).
Traditionally, gatekeeping has received limited attention in research. Concomitantly, researchers' relationship and collaboration with gatekeepers is often perceived as a straightforward and finite research phase, rather than a continuous, complex and dynamic rapport-building process (for exceptions, see Aaltonen and Kivijärvi, 2019;Clark, 2011;Crowhurst, 2013;McAreavey and Das, 2013;Miller and Bell, 2012;Reeves, 2010;Turnera and Almacka, 2017). From the researchers' point of view, the relationship with gatekeepers inevitablyand sometimes constantlyshapes the researchers' understanding of the research context and their relationships with the research participants (Sanghera and Thapar-Björkert, 2008).
Our starting point here is that access and consent should be discussed and analysed together and revalued throughout the research process (Miller and Bell, 2012). Research projects and their outcomes may change, and the precise nature of participant consent may only gradually become clear at the end ofor even followingthe study (Miller and Bell, 2012). Therefore, taking consent for granted once participants have agreed to participate is problematic. The requirement to obtain 'informed' consent from research participants has become rather an ethical given in research, although it is arguably impossible to provide subjects with a full and fixed explanation of a study that they can use to make informed choices about participation (Plankey-Videla, 2012). As researchers cannot know the whole research process or findings in advance, participants cannot know exactly what they are consenting to (Miller and Bell, 2012;also Bergman Blix and Wettergren, 2015;Leigh et al., 2020).
Access and consent are often viewed from the perspective of identities, rapport, power and ethical reflexivity. Campbell et al. (2006) and Crowhurst (2013) both emphasise the incompleteness of ethical reflexivity, as opposed to the ability to 'fully understand and then control relationships in the field' (Campbell et al., 2006: 118). This view highlights that researchers' positions are shaped by the hierarchies attached to the social identities present, and that research ethics are contextually contingent, historically specific, and always ongoing (Renold et al., 2010). Similarly to Harrington (2003), instead of seeing research access and consent as matters of cooperation or exploitation, we view all the parties involved as having some power. Thus, 'power becomes a variable rather than fixed attribute of either researchers or participants, allowing for a much broader conception of both' (Harrington, 2003: 618).
The processes of negotiating access can involve emotional work performed by researchers as they try to find their way into the 'field' (Bergman Blix and Wettergren, 2015;Peticca-Harris et al., 2016). Borrelli (2020) has noted that, besides emotional effort, access involves gaining and negotiating trust, and the positioning of the researcher requires strategic planning. We have conceptualised this process here as ethical performance. This paves the way to addressing the consequences of negotiating access and examining how this affects the research and the results (Rantatalo et al., 2018).
In sum, gaining access to research sites has traditionally been perceived as a separate phase preceding the actual research, largely governed by the researcher and devoid of analytic or reflexive interest. More recently, it has been treated as dynamic processes involving social relations, embedded continuously in the research process, and able to shed light on the research environment and substance. Yet, although gaining access, negotiating consent, collaborating with gatekeepers and 'researcher effect' are increasingly discussed in methodological texts, they are rarely considered in empirical articles or put under explicit analysis (Leigh et al., 2020;Reeves, 2010). In this article, we direct our analytical interest to the content and meaning of access and consent and the roles and positions of different actors (i.e. the researchers, gatekeepers and participants) involved in negotiations regarding access and consent. We consider these methodological moments as points of ethical performance unfolding qualitative empirical research encounters and practices and, simultaneously, capable of offering substantive insights into, for example, the organisational hierarchies, time pressures and power dynamics inherent in the institutional encounters and relationships under study (see also Borrelli, 2020).

Data and the conducting of the analysis
Our analysis stems from ethnographic data on institutional encounters between families with migrant backgrounds and the local Finnish welfare state, produced within a research project that aims to scrutinise the 'migrant family' as a social, moral and bureaucratic category subject to specific ordering practices (Lipsky, 1980). The research sites were four social work agencies in the Helsinki capital region. Two of the agencies provide services for new migrants mainly with refugee or asylum backgrounds during the approximately three-year official period of integration, one performs adult social work, and the fourth provides child welfare services.
We draw specifically on our field notes from: 15 meetings with municipal managers and social work staff, during which we discussed and negotiated research access; 39 observed service encounters between social work practitioners and migrant-background service users; 31 short reflective discussions with practitioners after observed encounters; and five turndowns of research participation by service users.
For the research, we negotiated access and consent in stages (Grant, 2017;Reeves, 2010). These included: (1) municipal research permit, (2) municipal managers, (3) senior social workers/local team leaders, (4) social work practitioners and (5) service users. We saw these stages, and the roles of gatekeepers and research participants, as overlapping. In the research project as a whole, the municipal managers, local social work team leaders and social work practitioners all acted as both gatekeepers and participants. Also, all participants were, in a sense, gatekeepers, at least of their own thoughts and experiences, since they decided what to reveal (or not) in the research encounters (Harrington, 2003;Kara, 2016;Wahab, 2003).
The first and second authors conducted the empirical research, although at some initial meetings concerning access to the field, all four researchers were present. The first and second authors also carried out the first stages of the analysis and prepared the initial drafts of the manuscript, which were discussed by all four authors at regular intervals, critically evaluated and developed further. The first author, then, conducted the final stage of the analysis and was responsible for writing the article, while the others continued contributing by editing drafts and clarifying the contribution.
Our analysis began by identifying analytically interesting, recurrent, or isolated points and events from the data and from our experiences as actors in these processes and encounters, as we read and discussed the material. We then gathered from these events a selection of critical moments regarding access and consent. We further developed our understanding of these issues by reviewing earlier research, and, by using an abductive approach (Tavory and Timmermans, 2014), we engaged in a back-and forth process of interpretation and dialogue between our data and the relevant methodological and empirical research literature.
The following sections present and discuss the results of this process. The extracts we have chosen to elaborate more in detail serve as illustrative examples of our interpretations. The extracts have been translated into English by the researchers. We have made some changes in the extract texts when we viewed this necessary in order to ensure anonymity or increase readability.

Access as pre-negotiated, multi-stage, extended and under-revision
We negotiated access at local team meetings with potential practitioner participants, after the municipalities and the municipal management had given us access. We had also discussed the research with senior social workers in meetings in which some expressed a strong desire for their teams of practitioners to participate. Hence, many managerial gatekeepers paved the way for our study. This raised concerns regarding the prenegotiated nature of the practitioners' participation, and we wondered whether they participated out of actual interest, a sense of duty, or on direct or perceived orders from the management.
Although these concerns were never entirely eradicated, many meetings with the local teams of practitioners were meetings where access was intensely discussed and the research proposal considered and challenged, as the following example from our research diaries shows: Even though the management level has been enthusiastic about the whole idea of considering migrant families, people here clearly challenge more: 'What's the point really, and how do we make this work in practice?' This example from a social work practitioner team meeting highlights the main difference between previous stages and the fourth stage of negotiation with practitioners. Although practical matters were considered at earlier stages, the emphasis here shifted from the research idea to the concrete practicalities of actually carrying it out, and this opened a demanding stage of obtaining and maintaining the practitioners' attention during their daily work (also Aaltonen and Kivijärvi, 2019).
Immediately after the team meeting, I received two emails suggesting service user meetings I potentially could come and observe. I had to turn them both down due to double-booking by the other research site. Since then, I have not had any invitations in a couple of weeks, and I have not received answers to my emails.
This extract is a telling example of the dynamics of timing in empirical research. The researcher notes here a positive response to the first presentation of the research project to practitioners at one research site, which quickly developed into two invitations to potential service encounter observations, both of which she unfortunately had to decline. The extract shows how researchers should always respond immediately when access is granted (Grant, 2017), but it also shows how changeable this access can be. Although the practitioners expressed interest and seemed to support the research, it resulted challenging to negotiate their continuous commitment (also Rantatalo et al., 2018). Perhaps also due to our research design, which meant that we were not constantly physically present in the agencies, we found ourselves battling for the practitioners' attention amid various initiatives, projects, student theses, staff training sessions, and the general time pressures of their work. We were also advised to more explicitly motivate potential practitioners and service users to participate.
The importance of gatekeepers and their relationships with potential study participants has been discussed in earlier literature (e.g. Clark, 2011;Crowhurst, 2013;McAreavey and Das, 2013;Reeves, 2010). When social work practitioners act as gatekeepers for the research participation of service users, a good rapport may create a trusting attitude towards the researchers, and willingness to participate, which may subsequently turn into a sense of obligation. On the other hand, a lack of trust between practitioners and service users may also result in willingness to participate in research, in order for the service users to reveal reasons for their mistrust.
The practitioners were quite positive, and some said they could immediately think of a couple of potential participants. Many stated that participation depended on trust between the practitioner and the service user. It did not seem reasonable to the practitioners to suggest observations when, for example, the researchers met service users for the first time or in multi-professional meetings with many people present. I said that I trusted their judgement.
In this example, practitioners bring forward the idea of service users' trust in the practitioners and the extension of that trust to an unknown researcher. Research participation may involve many reconfigurations and recontextualisations of roles regarding the relationships between practitioners and service users. Whenever research participation is suggested to service users by their social workers, there is a need to consider the mixed dynamics of dependency and reciprocity (Kara, 2016;Williams, 2017).
The extract also provides an example of different preconditions for participation. In practice, observations took place in the situations referred to in the extract, because the parameters of access varied according to the social work agency or individual practitioner. Different plausible or non-plausible conditions for meetings that researchers could attend were discussed throughout the active empirical research phase. The limits might include telephone interpretation or the presence of children at meetings, with the latter being impossible to predict in practice. We were therefore operating within porous preconditions for access. While this made access unpredictable, it could also be read as a sign of practitioner gatekeepers' engagement and commitment.
Many practitioners expressed that they saw the research important and wanted to offer support by reaching out to potential service-user participants, but this was not easy to arrange in practice. Practitioners generally met service users at quite lengthy intervals between meetings, and they did not consider it feasible to contact the service users individually merely to discuss research participation, due to the practitioners' workloads and the lack of a common language with service users, which would make an interpreter necessary. Their access to potential service-user participants, therefore, had its own limitations and conditions.
Our final point here concerns situations where, during the research encounters, access surfaced, as in the following example: I leave the home visit together with the practitioner X, and in the courtyard we stop to reflect a little on the meeting. X brings up the moment when one of the family members suddenly became quite upset, explaining that this did not have anything to do with the research; it was related to the complex and difficult situation from which the service users' relationship with the agency had arisen. I don't ask about this any further because, at the meeting, I got the impression that the parents might not want these events to be part of the data.
This extract highlights the need for ongoing in-situ attentiveness and decision-making over access. It first refers to the possibility of losing access that has already been granted by the service users, either because the research turns out to be upsetting in some way, or due to the varying dynamics in the relationships between service users, practitioners, and the services overall. The example also shows how we as researchers were constantly regulating access as we drew boundaries around the issues discussed when, for example, reflecting on the observed service encounters with practitioners. We self-regulated our access, making decisions over what was appropriate to ask or discuss when service-user participants were not themselves present. Also practitioners drew these lines, explicitly or implicitly, and sometimes openly discussed them.
This same practice of regulating access took place, in another sense, in the interviews with both practitioners and service users, when we decided whether to ask about certain issues or continue with a certain topic, since we constantly sensed the participants' willingness or reluctance to discuss them. At times, participants also clearly stated that they did not want to discuss specific topics. Access thus became intertwined with consent and was constantly weighed, reflected upon, and complexly regulated, and not least in the analysis and writing phases of the research, when the researchers are still confronted with decision-making over the issues to be included in the analysis and publications.

Power and positionalities in negotiating research engagement and collaboration
Although there are always potential vulnerabilities in disclosure, participants in social research often find the opportunity to tell their stories motivating, and this presumably worked as an important driver for participation with respect to both practitioners and service users. In the negotiations regarding access and consent, service-user participants sometimes accompanied their consent with the phrase 'I have nothing to hide'. We wondered, whether this referred to an assumption that the researcher might be snooping or fishing for secrets, or that the participants, once having agreed to research participation, no longer viewed they could hold back information they might wish to keep to themselves. It also places the perceived confidentiality and anonymity of research in a questionable light. Finally, the phrase may be interpreted as a display or assertion directed at the social work practitioner.
In their roles as gatekeepers, practitioners at times showed great interest in choosing the 'right' participants (also Aaltonen and Kivijärvi, 2019). This often prompted the researchers to consider the roles of different actors in the research, in terms of their involvement, commitment, and collaboration on the one hand, and ownership of and responsibility for the research and its results on the other.
The practitioner starts to speak about the service user who has already given preliminary consent to research participation, but the observation has not yet taken place. I ask whether they will meet in the near future. It seems that meetings are frequent, but the practitioner worries about the service user's mental health, which seems to have worsened. The practitioner is unsure whether the service user would properly understand what the research is about and believes that an interview could become chaotic. Before I say anything, the senior social worker takes charge and states that, because of changes in the service user's mental health, participation in the study would no longer be ethical.
This excerpt shows again how, due to changes in service users' everyday life trajectories, negotiations concerning access were continuing, situated, and context dependent. It also highlights the power of gatekeepers, since they one-sidedly decided on the service user's withdrawal from the research (Turnera and Almacka, 2017). Interestingly, besides evaluating participation from an ethical perspective, practitioners extended their consideration here to the quality of the research data potentially obtained from an interview with the service user in the current situation. This may be read as their attempt to balance the needs and interests of both the service-user participant and the researcher, as they exercised their power as gatekeepers.
A [the service user] and the interpreter arrive. B [the practitioner] introduces me to A, and A says that it's OK if I stay and explain the research. I start giving the presentation, sensing that I'm doing this in a complicated way and taking too much time. A seems surprised by the informed consent form to be signed and perhaps also by the individual interview. A says that it's OK for me to stay and observe the meeting, but all the rest of it … B intervenes and starts explaining the research design. B repeats some things I have already said, but also adds or emphasises some issues, such as the fact that this research is being conducted in many different social work agencies in the capital region, and that both the service users and the professionals are being interviewed.
In this example, the practitioner gatekeeper (B) takes on a strong role in the conversation about research participation and informed consent. When it seemed that the service user (A) was having doubts over research participation, B took charge. The meeting took place in B's office, and, although extra time was appointed to service meetings where the research was presented and discussed, we were potentially taking up both B's working time and A's service-use time, so B operated here from a position of power and intervened to guide the conversation. Interestingly, B also exercised power by choosing to stress certain aspects of the research design to alleviate any threat the research participation might present by emphasising anonymity and equality, as she explained that the research was taking place in many agencies and that also the practitioners were being interviewed.
Our approach of presenting the research for service users at the beginning of service meetings we were hoping to observe was not ideal, since it did not offer the service users ample time to reflect on the aims and methods of the research and the consequences of participating. We worried that this might be perceived as pressure to participate or cause confusion of roles, despite us emphasising the separateness of the research from the services and explaining the voluntary nature of participation.
The moments of presenting the research and negotiating consent were intense in-situ trust-building situations. Partial participation was also taken up in these discussions, which meant participating only by the service meeting being observed, with or without the use of tape recorder, and not being interviewed, and some opted for this. Although the possibility of partial participation may be understood as flexibility in the research approach, it can also be read as persuasion: I offer the info sheet to Z [the service user] and talk about the research and its independent and voluntary nature. Z folds the sheet without looking at it and puts it away. Z says that there are so many things going on right now in their lives, and declines participation because of this. I'm about to get up when Y [the practitioner] says that perhaps the problem is the individual interview meeting with all the extra arrangements and asks whether it would be possible to participate only by consenting to the researcher observing the service meeting. I say that this is also possible and ask whether this would be OK for Z. Z agrees. I hand out the informed consent sheet, which Z again folds and puts away. We talk a little more … I again mention the informed consent form and suggest that we go through it together before signing. Z declines again participation because there is so much going on right now. I express understanding and confirm that it is completely OK to refuse to participate. Y asks whether it might be possible to participate without signing the written consent form. I say that it seems to me that Z is not willing to participate even if we agree upon the research consent verbally. I gather my things and leave Z, Y and the interpreter to continue the meeting. As we say goodbye, Y says that we'll be in touch about other potential meetings.
This extract indicates a discrepancy between the practitioner's (Y's) keen efforts to make participation feasible for the service user (Z) and the researcher's more cautious approach. Y's efforts are understandable because Y had organised the meeting and perhaps felt bad for the researcher or did not want to go to the trouble of finding another potential service-user participant for the research. Nevertheless, the extract again shows the strong role played by the gatekeeper, not only in terms of granting access, but also in the negotiations regarding consent.
Service users who were, for whatever reason, living under stress or in particularly difficult situations in their lives may have found the research participation challenging because of the time it required and because it involved extra arrangements. Earlier research has also pointed out that official ethical standards requiring written consent at the outset can prove detrimental for research on sensitive issues or interested in people and groups who can be considered 'difficult to access' (e.g. Miller and Bell, 2012;Miller and Boulton, 2007). Participating without signing a written consent should be possible, especially if the participant is not fully literate. Signing official documents for research participation might also cause anxiety and suspicion in participants who distrust officials and the system. The extract reveals a tension between wanting to eliminate potential obstacles to participation and undue persuasion to participate. At the same time, expecting an immediate positive and uncomplicated attitude towards research participation from prospective participants can potentially bare influence on the characteristics of the group of participants, the research data, and the results (Miller and Bell, 2012).
Difficult moments in the consent processes could also be linked to power relations between family members. With this we refer to situations where the gatekeeperpractitioner might have already obtained initial consent from one adult family member, but it became clear at the meeting to be observed by the researcher that this information had not been shared with others. Perhaps they had not considered it necessary to discuss the participation beforehand; the research and participation were unclear, meaning that they did not know how to discuss it; or they had merely forgotten to tell other family members. These situations nevertheless further exemplify the ways these negotiations developed rapidly and were complex, surprising, confusing and even messy. Researchers are often left with only the opportunity, and obligation, to reflect afterwards on their decisions and actions and the ethical and other consequences these might have for the research and the people engaged in it.
As the last point here, we wish to draw attention to the positionalities of the people engaged in the research encounters. In negotiations regarding access and consent, researchers do not fully decide their positions since, to a certain degree, roles are assigned to them (Reeves, 2010). If we were to locate ourselves at 'the scene' (Martín Pérez, 2006: 10;also Kara, 2016: 84-86) and list gatekeepers' and participants' possible perceptions of us, these might include, at least, 'academic', 'expert', 'ignorant of the real world', 'potential threat or critic', 'ally', 'colleague', 'partner in reflection', 'politician', 'social work teacher' and 'yet another person asking personal questions.' From the outset, our 'intersectional ' (e.g. Hill Collins, 2015;Mattsson, 2014) locations positioned us as part of the institutional context. We were introduced to the service-user participants by the practitioners, and we were, like many of the latter, Finnish-speaking white women with university degrees. At times, however, commonalities with service-user participants became apparent, for example, when sharing service-user experiences of the welfare services or the national social security system in general, and in some cases we shared a common language other than Finnish (or English) in which we communicated.
Our 'insider position' (Kara, 2016: 84-86;Kusow, 2003;Markova, 2009) with regard to the practitioner participants was not to be taken for granted. Some practitioners expressed doubt over our understanding of the realities of the institutional settings and demands (such as a lack of time and other resources) framing their work. Other examples of doubts included challenging our use of the term 'migrant' because the service users of the particular agency were mainly people with refugee or asylum backgrounds, or questioning the translation of the research material into twelve languages by pointing out that many of the service users were in fact illiterate (cf. Roesch-Marsh et al., 2011). However, the practitioner participants also, at times, approached the researchers as social work teachers able or prepared to evaluate their work performance or give guidance. As we were positioned and we positioned ourselves, we continuously assessed, modified and tested our roles in these moments (Borrelli, 2020).
Researchers have found that personal contacts with the study population or site may result beneficial for the negotiations concerning access (Duke, 2002;Reeves, 2010). Our earlier experience of social work practice, education and research, our positions in the university, and some network memberships might have eased access especially at the institutional level and at the individual level with the practitioners (cf. Grant, 2017;Roesch-Marsh et al., 2011). Some practitioners who were familiar with us personally (e.g. through their studies or through serving as social work practice teachers) acted as 'respected colleagues' (Bergman Blix and Wettergren, 2015) who participated in the research and encouraged others' participation. However, previous connections or teacherstudent relationships may also have made it difficult to decline participation (Turnera and Almacka, 2017). Relational configurations of trust and power are not straightforward or one-dimensional, but always multiple, fluid and complex (Edwards, 2013).

Conclusions and discussion
In this article, we have analysed negotiations of access and consent in research. The challenges in gaining access, and doubts concerning consent, are not often discussed openly in research accounts, although many researchers experience and may struggle with them. We argue that interactions between researchers, gatekeepers and participants are importantly relational, generative, transformative, and ongoing, and they have a continuous impact on research (Crowhurst, 2013). Rather than being treated as something that 'happens before' the actual research, the processes for gaining access and consent should be accorded the same importance and analytical interest as other research encounters (e.g. Bondy, 2012;Rantatalo et al., 2018).
Access in this research was a multi-stage process including various stage-related negotiations, and the earlier stages always bore some influence on the stages that followed. We started with the municipal administration and municipal managers and worked our way towards social work practitioners and service users. This may have caused confusion about the independence of the research (Plankey-Videla, 2012), and we were concerned about the voluntary nature of participation for both practitioners and service users.
Yet access and consent could never be taken for granted but was always erratic and subject to revision. Our analysis revealed active, explicit and implicit testing and reconfiguring of the research on behalf of both practitioners and service users. Gatekeepers and research participants might pose explicit questions about the research design, exercise power over its conduct, resist full collaboration and seek professional gain. Besides revealing dynamic, yet asymmetric, relationships in the research encounters, the analysis shed light on organisational manager-practitioner roles and practitioner-service user relationships in the context of the study (Borrelli, 2020). Furthermore, it highlighted the necessarily collaborative nature of any empirical research and the need to foster long-term research-practice partnerships.
We see access here as extended, through the manifold negotiations and regulations of access to information that take place in research encounters, and we understand access as intertwined with questions of consent. Similarly, we found that the roles of gatekeepers and participants overlapped. Furthermore, gatekeepers actively engaged, not only in procuring access, but also in the negotiations regarding consent. Our analysis also revealed some limitations regarding gatekeepers' own access, the porosity of the preconditions gatekeepers impose on access, and practitioner gatekeepers' balancing of the needs and interests of both service-user participants and researchers as they exercised their power as gatekeepers.
Our positionalities as 'insiders' or 'outsiders' (Kara, 2016: 84-86;Kusow, 2003;Markova, 2009;also Merton, 1972) with respect to the managers, practitioners and service users influenced the negotiations around access and trust building. Yet, these positionalities were multiple, complex and shifting (Kennedy-Macfoy, 2013). Our statuses and positions were negotiated in relation to our own self-perceptions, the perceptions of others, and the demands of the research (Sherif, 2001).
It is not our aim to argue that turndowns or reluctance to participate are necessarily problematic as such, as we consider them inevitable, on the one hand, and view them as a sign of the potential participants' decision-making power and possibilities, on the other. It is important, however, to be reflexive about who turns down participation and why, and what consequences this might have for the overall group of participants, the data formed and produced, and the results.
Professional and academic research guidelines have the potential to obscure the need to continually reflect on and analyse the ethical challenges present in research (Miller and Bell, 2012). Consent is not merely a tick in a box or a signature on a piece of paper, but must be genuinely negotiated and renegotiated throughout the research, and ethical considerations regarding access and consent should be integrated into research throughout the process (Kennedy-Macfoy, 2013).
Importantly, ethical reflexivity is hands-on and concrete. It involves continuous in-themoment decision-making on difficult situations rather than simply sticking to the formal research ethics expressed in the research plan and research permits (Jäppinen, 2015;Kara, 2016). It therefore requires a great deal of continuous self-awareness from researchers. Although not extensively discussed in research accounts, failure is as well part of research (Harrowell et al., 2018). We have attempted in this paper to demystify the research process by offering a critical examination and analysis of it and of our own roles as researchers.