Work-Identity in Ethnographic Research: Developing Field Roles in a Demanding Workplace Setting

Abstract

In this article we problematize our field roles as two linguistic ethnographers who aim to study the communication and documentation practices drawn upon by care workers in elderly care facilities in Sweden. Our field roles are discussed in relation to the complex nature of care workers' knowledge and competence, which results from three different aspects of their work-identities: institutional, professional, and individual. As researchers, we found ourselves in constant dialogue with the research participants, and our field roles were continuously shaped and reshaped according to the individuals and the situations in which we became involved. Even aspects of our own identities taken into the field, such as our background and personal qualities, proved to be important in establishing good relations with the care staff. Coming closer to the participants' professional identity proved to be of utmost importance for interpreting their choices and decisions in the workplace. Identity negotiation is presented here as a constructive way of discussing ethnographic field roles in the research field.

With the emergence of reflexivity and the rise of postmodernist perspectives in recent ethnographic research, we have seen increasing concern related to the researcher's field role in the generation of data (Allen, 2004; Dwyer & Buckle, 2009; Ledger, 2010; Perrone, 2010; Sarangi, 2004; Savage, 2000; Taylor, 2011; Tinney, 2008). The emphasis on the complexity of field roles has opened up a reflexive discourse, informed by increased consciousness of situational identities and the researcher's dialogical relationships in the field. Postmodern ethnographers emphasize the social practices through which the researcher's field role is negotiated. Such an approach treats the field role as an interactional and social accomplishment (Allen, 2004; Angrosino, 2005) and not as an absolute category. Negotiating a membership/field role through participating in various activities in the field is recognised as a key strategy that enables the qualitative researcher to obtain an insider perspective (Hammersley & Atkinson, 2007). Nevertheless, several ethnographers have reported their feelings of awkwardness and dread of participating in activities upon entering the field. When the research site is a workplace, these feelings can be even stronger because the responsibility of the researcher is even greater. The present article, therefore, aims at providing insight into the negotiation of field roles and identities in organizational research. The article demonstrates how two researchers' field identities were shaped in dialogue with the research participants.

Background

The data in this article are drawn from a project about immigrant care workers' communicative practices in elderly care facilities in Sweden. One of the researchers had a special focus on care workers' practices of talk in their everyday communication with the elderly. The other researcher concentrated on care workers' literacy practices. The data collection methods we have used include participant observation, qualitative interviews, and audio/video recordings. The aim of our study was to identify, in collaboration with the community, the central dilemmas in care work (tensions between institutional frameworks and professional knowledge, among others), and to find out how these constrain the care workers' documentation practices and everyday communication with the elderly. The intention to work with rather than for the community required close interaction with its members (see Roberts & Sarangi, 1999). Attaining this aim required establishing close contact with the care workers and participating (even to a small extent) in their daily tasks. The population of the study consisted of approximately 50 care workers (assistant nurses and elderly carers) employed at ten different wards in three nursing homes. The nursing homes mainly host residents with dementia and therefore with great and immediate care needs. The majority of the care workers are immigrants from Middle Asia, Latin America, and Africa, and they have Swedish as a second language. Only a minority of the care workers are Swedish-born. Being a care worker with Swedish as a second language was one of our research topics in the project. Because elderly care in Sweden is a highly multilingual work context, we were interested in how the language of immigrant workers was shaped in the workplace. The questions we asked were how different kinds of work-identity are handled in communication and how institutional frameworks meet and contradict other perspectives within care work. Such research questions require an interdisciplinary approach that emphasizes microanalysis of situated interaction and linguistic text interpreted through an ethnographic understanding of the context of the workers' everyday tasks (Creese, 2008; Rampton, 2007; Tusting & Maybin, 2007). In line with recent developments within applied linguistics, we maintain that language and social life are mutually shaping (Rampton et al., 2004, p. 2). Our point is that the methods provided by ethnography generate insights into the language practices under study, which cannot be achieved in another way.

A central issue in ethnographic methodology is the reflexivity between the researcher and the research being produced. This stems from a tradition in which ethnographic research requires the researcher to be actively involved in social action. This involvement of the researcher in the field will have an unavoidable impact on the practices being studied. Where the study of language practices is concerned, reflexivity also raises another issue, namely the difference between how participants and analysts understand these practices (Creese, 2008, p. 579). The linguist's use of structured analytical tools runs the risk of deviating from the participants' perspective. In order to address this issue we have used participant observation and ethnographic interviews as data collection strategies. Our view is that the researcher's commitment in the field to come closer to the participants' perspective informs the analysis of language. We will return to this issue in the final section of the article in our discussion of how our position in the field impacted on the generation of data.

A central tool in ethnographic research for pursuing an emic perspective is the strategy of participant observation. Nevertheless, the nature of participant observation is contested and its foundations have been widely debated in recent studies (Allen, 2004; Angrosino, 2005; Angrosino & Mays de Perez, 2000; Davies, 2008; Savage, 2000). In particular, to date there has been no homogeneous view of the meaning of the term participation in participant observation. The several attempts that have been made to map out the various field roles that researchers may adopt in settings as participant observers (Adler & Adler, 1994; Gold, 1958) are nowadays considered of restricted value (see Hammersley & Atkinson, 2007). Ashworth (1995) highlights attunement to the emotions and motivations of the group's activities, for example attunement to the shared and taken-for-granted stock of knowledge that contributes to the forms of interaction of the group being studied, as crucial constituents of participation. Interpreted within Ashworth's theoretical construct, participation has the same meaning as to participate in the role of a practitioner. The implication of this reasoning is that the researcher, by virtue of his or her outsider status, will face difficulties when entering a new social group. Allen (2004), however, has argued that there are many other modes of participation in the field, and she warns against “an exclusive focus on the meaning of participation” (p. 17). Staying in the field for a longer time inevitably entails the researcher's involvement in the lives of the communities they study. As a consequence the researcher cannot remain an outsider throughout the process. In fact, many ethnographers account for the possible ways of negotiating a feasible membership role (Angrosino, 1997; Palmer, 2010; Perrone, 2010; Woodward, 2008).

The ethnography of nursing homes is a growing field of research (Diamond, 1992; Foner, 1994; Henderson & Vesperi, 1995; Higgins, 1998; Lee-Treweek, 2000; Reed-Danahay, 2001; Tinney, 2008). As has been described in this literature, the nursing home is a challenging research site and also a stigmatised location for those who live there. Several ethnographers (e.g., Backhaus, 2008; Grainger, 1995; Lee-Treweek, 2000) have focused on the horrors of nursing home life and described them as “total institutions” (Goffman, 1961) in Goffman's sense. Reed-Danahay (2001, p. 49), however, makes it clear that recent research has sought to understand nursing home life from the inside, in a way that is attentive to the stresses and strains on those who work there, for the most part women of colour in low-paid, low-status jobs. This endeavour to pay attention to the burdens of the staff was especially relevant in our study. Care workers in Sweden are under constant scrutiny by news media, where they are often exposed to public criticism and blamed for scandals. Elderly care work is associated with low prestige and, as in other Western countries, employs the highest proportion of immigrants, many of whom are women with few opportunities to advance in the labour market. Cuban (2009, pp. 178–179) has argued that the new demands placed on carers in the market-based industry help explain why migrant women are recruited to work in elderly care. Care employers seek out workers who are willing to comply with demanding requirements and stressful workplaces.

In previous publications, Nikolaidou and Karlsson (2012) have discussed the complex nature of care workers' knowledge and competence as the result of three different aspects of work-identities. Based on Roberts and Sarangi's (1999, 2003) analysis on discourses in medical care, we have argued that care workers' decisions and actions at work depend on three different expressions of selfhood: (a) the institutional work-identity, assigned through a hierarchical position in an institution; (b) the professional work-identity, constructed together with colleagues through shared responsibilities and the use of a common discourse; and (c) the individual work-identity, based on individual character and past-life experiences. These three aspects of work-identity are similar to Gee's (2000) categorisation of identity construction. He makes reference to an Institutional-identity that foregrounds membership in various institutions (similar to the institutional work-identity, described above) and an Affinity-identity adopted by people who share “allegiance to, access to, and participation in specific practices that provide each of the group's members the requisite experiences” (p. 105) (similar to the professional work-identity, described above). The third aspect Gee identifies is Nature-identity, which suggests an essentialist view of identity. This is similar to facets of individual work-identity (e.g., gender, ethnicity, and color of skin, etc.), even though our understanding of individual work-identity is broader. Finally, Gee identifies Discourse-identity, which is the way our selfhood is constructed by the way we are talked about by others and by the way we talk about others. This final aspect can be identified in all three expressions of work-identity.

In a recent publication on documentation practices in elderly care homes, we have shown that the institutional work-identity is dominant, especially in relation to the individual (Nikolaidou & Karlsson, 2012). Care workers are not allowed to include personal experiences and feelings about their interactions with the residents in documentation, but are instructed instead to follow rigid guidelines on the contents of records and to use a very restricted documentation language. Care workers promoted the professional work-identity in the documentation; for example when they included information they believed to be useful to their colleagues but not in line with the institutional guidelines. Documentation practices become more meaningful to care workers when they are related to their professional identity rather than to their institutional identity, whereas the individual aspect does not go down on paper and is instead expressed orally.

Entering the world of elderly care was not an easy task. This can be seen in relation to the complex nature of the research participants' professional roles, consisting of the various aspects of work-identities accounted for above. In this article, we therefore suggest an additional perspective to the understanding of membership roles, one that views them in relation to the three different expressions of identities. We examine our participation in the nursing home field by juxtaposing the care workers' work-identities with our own field roles as researchers. We argue that our choices and decisions in the field were also defined by the aspects of identity we chose to promote each time. Our actions were in a constant dialogue with those of the care workers, and together we helped each other shape new identities throughout the research process.

Although neither of us was a complete outsider in relation to the care workers, we had shaped our identities as researchers in different ways. One of us (Nikolaidou) was a young female in her late twenties, with a second-language background in Swedish and with limited experience of social acquaintances with elderly people in Sweden. The other (Jansson) was an older native Swedish female in her sixties who had worked in the field as a trainer in a previous project dealing with staff development and second-language use at work. Our aim in this article is to demonstrate how these different positions impacted on our field relationships and how we managed to negotiate our membership roles by adjusting our field identities to those of the research participants. In the following sections, we describe the activities through which we engaged in the field setting and how the participation we undertook positioned us in relation to the care workers.

Entering the Field: Juggling Between Roles

In ethnographic research, it is not possible to predict the way the study will progress or what kind of relationships will be established with the research participants. Due to our lack of nursing background and the nature of the study, it was evident from the beginning that we could be given only the status of peripheral members. Adler and Adler (1994) argue that peripheral membership can be intentionally sought when the researchers do not wish, for a number of reasons, to participate in all the activities of the group they are studying.

Jansson's Account: Researcher, Consultant, or a Language Teacher?

My background as a linguist and a language teacher, who had in addition been the care workers' communication trainer, clearly influenced my relationships in the field. In this position, I ran the risk of reinforcing the immigrant workers' sense of inadequacy in my attempts to report back on my observations of their practices of talk. Many of them were taking literacy courses sponsored by a training project in which I had been involved. My connection to this project, where I had been working with the implementation of successful communicative practices, shaped my interactions with the second-language workers. Especially at the beginning of the fieldwork, the care workers associated me with this earlier project. They treated me as a language teacher. This was evident, for example, in the way they referred to me in their in-group communication. During one instance, when I came to the ward in order to meet an immigrant care worker, her colleague announced my arrival by telling her: “It's her, the language teacher.” Another example of this was the way they kept requesting my help with the wording of care plans and other professional texts. One care worker, here called Stella (all research participants' names have been changed for anonymity purposes), gave me the minutes she had written during a staff meeting I attended. She not only wanted me to correct the language but also asked me for advice on how to compose such a text. The management had decided that all care workers should write minutes in order to improve their language skills. I found myself caught between the observer role and the consultant/expert role, a dilemma often faced by ethnographers in professional settings when some form of feedback is expected by the participants (Sarangi, 2004). While it was difficult for me to reject Stella's request for language help, at the same time I was assigned a role that I did not want to assume. Rather than being a language consultant available when workers needed my advice, I pursued a different field role. It was the participants who were the experts in this field, because they possessed the kind of professional knowledge that I was after. In order to negotiate this insider's perspective throughout the fieldwork, my intention was to be actively involved as a member in the social action under study. I wanted to understand how they felt about the institutional frameworks that defined their work.

My former status as a trainer in the setting also impacted on my initial attempts to approach potential research participants, and especially those with immigrant backgrounds. A care worker from Africa, here called Esra, rejected my proposal by exclaiming: “Why do you ask me and not someone else?” By “someone else” she meant her Swedish-born colleagues on the ward. She was obviously discomfited by my attempt to approach her. It struck me that perhaps she associated me with my previous project and believed that I was interested in her “deficient” language skills.

These and other similar events, made me reflect on the actions I undertook in the field. The initial attempts to approach the care workers clearly illustrate how my former insider status in the field as a trainer made it difficult for me to seek out a suitable membership role. My intention to come closer to the care workers' understanding of the practices under study seemed to have failed. Interpreted within an identity framework, I did not manage to renegotiate a field identity that was closer to the care workers.

This conflict between my own and the participants' understanding of the researcher role also raised ethical issues. The employees' reluctance to let me become involved in the everyday life of the ward was not only related to their belief that I conducted normative research. By not giving me access to the residents' rooms, the care workers believed that they were protecting the residents' integrity. They argued that their residents were very sensitive, and that some of them did not allow strangers into their rooms. This was undoubtedly a very professional way of reasoning, but it also indicated that the carers were unwilling to accept me as a member of their professional community. It was clear that they did not see me as one of them. It was not only that I did not have the same understanding and experience of communication as the practitioners. The roles I was assigned in the field were influenced by my background as a researcher and as an academic. This made it difficult for me to take on a professional identity similar to theirs—my personal aim to collect data conflicted with the care workers' perspectives and their concerns about the residents.

Nikolaidou's Account: Researcher, Language Student, or a Care Assistant?

Initially, the idea of participating in the care workers' daily tasks seemed terrifying to me. Coming from a country where elder care is performed exclusively at home by relatives, I was negatively disposed to the idea of elderly care facilities. It was therefore a great revelation for me to visit a Swedish nursing home and experience this world, which had previously been connected with negative feelings.

In addition, I was a stranger to the nursing field and was not acquainted with the tasks of care workers. In Ashworth's (1995) terms, it could be argued that I did not have the same stock of knowledge as the rest of the group and, therefore, I could not initially be viewed in any way as an insider or a group participant. I could not see myself working in this environment, and I did not think I had the right skills and knowledge required to participate. When I followed the care workers to the residents' rooms, I would stand awkwardly at the entrance, uncertain about how I could help. While residents were being dressed and undressed I would always look away and when they were showered I would walk around in the residents' bedroom, trying to make myself almost invisible.

My initial decision was to actively participate in the residents' caretaking only when absolutely necessary and to give myself instead more time for observation and discussion with the staff; however, this was not always possible. Adler and Adler (1994) have argued that potential similarity to the subjects leaves researchers vulnerable to the possibility of being recruited for closer participation. This was true in my case, because aspects of my Nature-identity were very similar to those of the care workers; I was close to most of them in age and had a non-Swedish ethnic background. As a result, and because I was always available, they found it easy to ask for my help. In fact, I was so similar to the care workers that they often misinterpreted my role in the field, mistaking me for a practitioner who was there for a few days in order to practise Swedish. This was, of course, not my original intention and I was conscious to correct this misunderstanding at all times during the research. The vulnerability that Adler and Adler (1994) referred to related to my inability to refuse requests like the one described in this extract from my fieldnotes:

Katarin asked me if I could help her with one of the residents, since everyone else was busy and they had run out of staff. She explained that I would have to hold a female resident, while she would change her diaper. This was one of the most traumatic experiences so far in the nursery home, as the smell from the diaper was strong and the woman was moaning with pain and discomfort while I was holding her. The whole situation was so hard that my blood pressure dropped and I was about to faint! I had to excuse myself and go out and sit on the balcony and only after ten minutes did I feel I could walk again!

This extract demonstrates quite clearly the incongruence of my identity as a person and as a professional with the care workers. I lacked both professional knowledge and personal experience of care work and could not align my behaviour and feelings with theirs. Still, as a researcher I felt a deeper need to give something back to the people who were helping with my research, so I found myself accepting many similar requests. It soon became clear that I had to overcome many personal prejudices and fears in order to achieve the peripheral membership I had originally sought. Gradually, I started seeing myself as a person who could work more easily with elderly people and do physical work that was dirty and required a lot of mental strength. If I wanted to be accepted by the care workers, I needed to re-invent myself and thus gain a better understanding of their identity at work.

Negotiating Field Identities

The benefits of enacting a field identity are frequently discussed in the ethnographic literature (e.g., Dwyer & Buckle, 2009; Hammersley & Atkinson, 2007). Many ethnographers explain how a membership role allows them more rapid and more complete acceptance by their participants. In the following section, we show how we gradually, and often unconsciously, modified our actions and behaviour in the field in order to come closer to the care workers. We illustrate how we negotiated our field identities, not only to be accepted by them but also to attain a better understanding of their working knowledge and experiences and the reasons for the choices they make concerning communication.

Jansson's Account: Strategies for Being Accepted by the Staff and Limits for Participation

Self-presentation and personal appearance are described as salient considerations in gaining respect (e.g., Hammersley & Atkinson, 2007). I acquired the staff's individual qualities and learned a lot from them. I sought to understand and come closer to the care workers' personalities, their individual work-identity, and the way they behaved and interacted with the elderly.

One professional mode of behaviour in the nursing home community that I quickly adopted was the exchange of compliments with the residents at the nursing home. This was a way of demonstrating social affiliation and reducing the institutional character of the interaction during task completion. Such exchanges also made it easier to create situations of confidence with the care workers. In this respect, compliments can be interpreted as an expression of the professional aspect of the care workers' work-identity, as well as the individual aspect of their work-identity. I therefore used compliments as a strategy for opening up a small talk sequence with the residents, for example by saying, “Oh what a nice sweater you have today.” Such an approach could, for instance, lead to a discussion about the resident's relatives who had bought the sweater. The residents also showed me appreciation for coming to the ward by saying, for example, “It's always nice to have you around here. That's what we all think” or “Oh how nice to see you around here again” (fieldnotes). I was flattered by this praise and I felt a form of joy that arose from having found a niche for negotiating an identity that was close to the care workers. The care workers often participated in my conversations with the elderly, for example by suggesting a topic, “Have you told Gunilla about our excursion this morning?” In this way, I was given a role as a social communicator and a helper with the residents, a role that in a sense legitimized my presence on the ward.

Examples like these illustrate how I started to adopt aspects of the care workers' professional identity as I began to understand how to approach them and talk with them. When it came to social talk with the residents I had, at least to a certain degree, the same professional knowledge. There are certainly also examples from the field that reveal that there were differences between my participation in social talk and the care workers', for example in the ability to be sensitive to the resident's interpretation of current events, an ability particularly important in interactions with patients suffering from dementia. It happened sometimes that care workers intervened in my communication with the elderly. In retrospect, this action can be interpreted as a professional concern for the elderly and a mode of steering the conversation into a frame that was more attentive to theirs. It also illustrates how participation in language practices as a researcher is different from that of the people involved.

Gradually, I realized that my good social relationships with the residents earned me a degree of acceptance from the staff. The carers noticed that the elderly residents trusted me and did not hesitate to request my help. In some cases, as in the following extract, it was difficult to determine the limits of the field identity as a helper I sought to assume:

While they are bathing the resident Wanda says that Ada [a resident] is sitting outside and requests help. She sounds a bit worried. “I can follow her to her room and help her while you are taking care of Maja,” I suggest. Wanda accepts my proposal and says, “Yes please, that would be great of you.”… I follow the resident to her room. She wants a glass of milk and I go and fetch one in the kitchen where I help another resident who is blind and cannot find her wheelchair. When I come back to Ada's room I follow her to her bed. (fieldnotes)

The difficulties in determining what was an appropriate behavior in my efforts to come close to the care workers' personalities in settings like the one described above might be because some of the care workers' professional skills were based on individual experience. This obscured differences between the individual aspect and the professional aspect of work-identity (Nikolaidou & Karlsson, 2012; Törnquist, 2004). One question that could be raised is whether I offered too much in this setting—to the detriment of the research. Later in the evening, Eva and Wanda made comments that indicated I offered more than was expected from me, and that what I was doing was part of the care workers' tasks and responsibilities.

In some cases, I was involved in care situations that required both institutional knowledge and professional skills like those of the care workers. The extract below illustrates a situation where my participation was clearly limited, even though I was present when something went wrong. I became a sort of witness to the event. A resident had fallen on the floor and I happened to find her:

When I enter the television room I find Elsa half-lying on the floor. I become frightened and I don't know what to do so I go and tell Carla who is in the kitchen. She gets upset and asks the resident if she has pain. We put a pillow under the resident's head and ensure that she is lying in a comfortable position…. After lunch Carla writes an incident report about the fall. “We find her on the floor,” she writes. “It was you who found her. Shall we write that?” she says to me. We begin to discuss the event, how we found her, what we did, etc. At the report meeting, Carla tells about the event, and says that “it was Gunilla who found her.” (fieldnotes)

Since I found the resident on the floor, I became involved in the discussion of the details characterizing the event. My knowledge became relevant to the staff's institutional task of filling out an incident report for the nurse. In an incident report, the care worker has to describe the sequence of events and explain the cause. In this case, the person responsible for writing the report should be the person who found the resident on the floor. Carla pointed out that I had found Elsa on the floor and asked whether she should write that. Instead of writing the name of the reporter, she wrote “we” when she described the event. This example clearly illustrates the limits of my participation and my involvement in the social action under study. First, I did not manage to adopt a professional role to enable the situation to be handled in an appropriate way. A practitioner would probably have helped the resident before rushing away to the kitchen looking for the others. Second, I lacked an institutional role as a reporter of the event. This became obvious in the writing of the incident report, in which Carla avoided including my name.

Nikolaidou's Account: Uncovering the Care Workers' Work-Identities

As time went by in the field, it became clear to me that the best strategy for establishing a close relationship with the care workers was to emphasize our similar Nature-identities, or in other words our common background. Most of the care workers were women who had immigrated to Sweden for work or family reasons. The similarities between the courses of their lives and mine were striking. I had come to Sweden because of my family and for work, and I had spoken Swedish for only a couple of years. Like them, I was using a language other than my mother tongue in my workplace, and I was constantly aware of language use in both oral and written situations. I consciously referenced my limited knowledge of Swedish in order to avoid the danger of being perceived by the care workers as someone who was evaluating their language skills.

This shared Nature-identity played an even more important role because the research participants trusted me more easily and opened up to me in a way they would probably not have done if I had been a Swedish researcher. It made sense to them to talk to me about the stress experienced when working in a language other than their mother tongue. Some of them even discussed their status as immigrants in Swedish society and the efforts they made to be accepted as equals. The following extract comes from an interview with Maya, a care worker who had lived in Sweden for 11 years. The extract is indicative of the personal nature of the discussions I had with some of the care workers:

M. It is very important that we immigrants help each other because we are a part of society, if we don't help no one will help us. [pause] It is very difficult, especially in Swedish society, it is very difficult to come in.

Z. What do you mean “to come in?”

M. Like integration [pause] and it is very difficult to accept things that we accept more easily, very difficult for them, so to come in that circle they are sitting is very difficult for me.

The interview with Maya ended in a friendly discussion about our efforts to adapt to Swedish society and culture. During the interview, it became clear to me that when she said “we,” she also included me in the group of immigrants working in Sweden. Similar discussions often took place, as the care workers confided their disappointment at not being able to get a job that matched their previous education and their frustration at not being able to express themselves clearly, which therefore meant they were viewed as less clever and capable persons. In addition, we shared anecdotes about our first days in Sweden and found out that we were sometimes equally impressed, surprised, or amused by Swedish culture. It can be argued that I was “making a role” for myself, in line with Angrosino and Mays de Perez's (2000) advice that a researcher should react to any preassigned positions. I was therefore rejecting the role of an external academic and becoming someone they could relate to, a colleague and even sometimes a friend. This role was directly relevant to my identity as an individual and had a positive impact on my field relations.

My epistemological background as a sociolinguist and a workplace ethnographer, combined with the specific research aims for this study, were the last tools I had at my disposal for becoming a participant in the group of care workers. During the research, it became clear that the use of Swedish as a second language was not the primary reason for the problems occurring in the nursing homes. My research aims helped me achieve with the research participants what Ashworth (1995) has called “emotional and motivational attunement” (p. 375). I had achieved a level of participation whereby I could see reality at the nursing home from the care workers' perspective. I was finally able to interpret the decisions they made during documentation and the reasons for the difficulties they encountered. For example, at the beginning of the research I was always puzzled when I observed that a care worker opted not to write about the aggressive behaviour of a resident. Later I realised this decision was based on the institutional instructions that did not allow them to write in a way that was threatening to the residents' integrity. In private, however, the care workers informally discussed these incidents between themselves, and I came to understand that this served the professional aspect of their identity, because they needed to inform each other on the actual state of each resident. This realisation was made clear to me only when I started working with them for real, in other words, when I became a kind of colleague they could trust.

The more my research progressed and the more data I collected, the better I came to understand how they felt about the texts they were reading and writing, and the literacy demands placed upon them. At the end of the study, it was no longer clear whose motives we were promoting, theirs or mine? It had somehow become a shared aim to prove that other factors besides the workers' language skills shaped their documentation practices. This overlap indicated that I had succeeded in establishing a kind of insider status, maybe not in the art of caring for the elderly, but in being accepted also as an immigrant working with texts in a language that I did not fully master.

Discussion

In this article, we present a reflective account of the way we as ethnographic linguists approached a research field that was almost unknown to us. Our aim has been to show how we managed to approach the research participants by examining their knowledge and work-identities and by making an often unconscious effort to align our identities to theirs. We discuss the care workers' work-identity as a complicated phenomenon composed of three different aspects: institutional, professional, and individual; and we show how they came to be relevant in our efforts to create a membership role for us in the field.

We first entered the field with the initial dichotomy between insider and outsider researcher in mind. We viewed ourselves as complete outsiders to the field of elderly care and wished to gradually negotiate an insider role. We soon came to realise that this dichotomy was too barren and did not accurately describe the situation we were in. Even though neither of us had ever worked in a nursing home before, we felt we could not be called complete outsiders, because we both had some kind of previous relation to the field and to the research participants. According to Christensen and Dahl (1997), the insider-outsider dialectical roles constitute a continuum rather than a dichotomy “with the extreme points in the continuum existing in conceptualization rather than fact” (p. 274). The authors have questioned the extent to which outsiders can really be outsiders in relation to what they are researching and suggest that all researchers are close to their research participants in one way or another. Similarly, Allen (2004) has argued that the insider-outsider status in the field “is something to be discovered and renegotiated as part of the research process” (p. 22). We found that this role negotiation was performed through a constant shaping and reshaping of our identities, in relation to those of the research participants.

Early in the research, we realised that the fastest way to come closer to the care workers was to make use of the shared features in our personalities and backgrounds, or in our Nature-identities (Gee, 2000). For example, Jansson's ease in communicating with the residents made the care workers appreciate her. Nikolaidou had a shared immigrant background with the research participants, and this enabled them to open up to her in many different ways. Interestingly, neither of us had initially thought that these individual features would play an important role in the research. Rather, they gradually became prominent during our efforts to establish good relations with the participants and an insider role in the field.

The professional aspect of the care workers' identity was harder to interpret and potentially imitate. As academics, we had different professional backgrounds and different aims from the care workers. We wanted to collect good and informative data, whereas the research participants focused on the residents and also on protecting their own professional integrity. Some of the examples we have discussed in this article indicate this incongruence of professional identities; for instance, the care workers initially refused to give Jansson access to the residents' everyday lives and some of them resisted the idea of participating in the project, while others asked Nikolaidou to help with tasks for which she was not prepared.

The more the research progressed, the easier it became for us to decode the care workers' professional identity, and the more we tried to align our identities to theirs. In other words, we made an effort to prioritize their goals in the activity over our own and, in this way, to be accepted as something more than just observers. The care workers gradually started placing trust in us. They allowed, and sometimes encouraged, Jansson to be involved in certain forms of care, while they talked more freely to Nikolaidou about their problems with documentation at work. Jansson was no longer merely a trainer and Nikolaidou merely a researcher. We had embraced their own priorities and this was made clear to them. They could now talk in front of us freely at lunchtime and they could share their opinions about the manager and the residents. We would argue that we had come as close to being colleagues as is possible for someone who is actually in the field as a researcher.

The institutional aspect of the care workers' identity was of great interest for our study and in some cases easier to adapt to. Some aspects of care workers' institutional behaviour were imposed on us from the very beginning; for example, we had to wear the same clothes they did for sanitary reasons. Similarly, we were committed to treating the residents with respect and discretion. But this is where the similarities ended. As researchers, we were not obliged to provide care to any resident or to participate in the documentation processes. This often placed the care workers in an awkward position because they did not know where to position us in the institutional structure. This is made clear in Jansson's example when she found a resident on the floor. The care workers were not sure if they were supposed to include her in the documentation practice because she did not have a formal role. Nikolaidou often encountered mistrust when she asked to make copies of the residents' files, and she had to reassure the care workers more than once that she had permission for this. Our institutional identities could hardly be aligned, therefore, with that of the care workers because we belonged to a different institution and were subject to different rules and frameworks.

Despite the strategies we invented in order to acquire a membership role, issues of power were always at stake in the research field. Power relationships between researchers and research participants are a constant concern in many ethnographic studies (e.g., Beoku-Bets, 1994; Labaree, 2002; Tinney, 2008). Unequal distribution of power in our study was manifest in many different ways, often experienced as an obstacle in the negotiation process between the care workers and us. Because of her age and professional status, the participants often saw Jansson as a language expert who was there to help them and even correct them. In addition, she was associated with an earlier project where she had been working as a communication trainer. Jansson found it difficult to convince the care workers that she was there only to observe and learn and that, in most cases, they were better judges than her about what should be written. Nikolaidou, however, struggled with presenting herself as an immigrant who is also an academic. Even though she managed to make successful use of her identity as an immigrant who has difficulties with Swedish, her professional status contradicted the similarities she claimed to have with the care workers.

Finally, the field roles we negotiated for ourselves had a great impact on the kind of data we gathered. Our involvement in the field puts into focus the way in which the researcher's involvement alters the language practices under study, and hence the data collected (Tusting & Maybin, 2007). Because we came so close to decoding and understanding the components of the care workers' work-identity, it helped us get a complete picture of the communication and documentation practices in the nursing home. Our analysis demonstrates how we moved from a horrified perception of the nursing home to a more nuanced picture that is more sensitive to the care workers' dilemmas. This tension between perspectives is reflected in earlier research within the field of nursing home ethnography. Recent researchers are more sensitive to the strains and stresses of those who work there (Reed-Danahay, 2001), while others (e.g., Grainger, 1995) use Goffman's description of homes as “total institutions” (Goffman, 1961). What we have done in this article is to describe the strategies through which we revised our understanding of life in the nursing home and the language practices under study.

Jansson's ease with the elderly gained her the trust of both employees and residents, and she managed to get permission to video record the interactions between them. In this way, she captured the subtleties that characterized everyday staff-resident interactions in the nursing home, which made the dilemmas described by the care workers more understandable. Had it not been for her close field relationships with the care workers, the carers would probably not have shared their negative feelings about being forced to implement good communication practices in problematic and sometimes hostile environments. Nikolaidou's identification with the immigrant and second language status of the employees made it possible for her to ignore their claimed language difficulties and locate problems in the documentation structure of the nursing home. In other words, it was our negotiated research roles in the field and the continuous interaction with the research participants that determined the quality of the data we collected.

By “making a role” for ourselves we were able to negotiate access and finally collect useful data. It has been claimed that such information can be more easily gained by insiders (Haniff, 1985; Merton, 1972), but we tend to agree with more recent studies that describe field roles as contingent and shifting between contexts (Allen, 2004; Christensen & Dahl, 1997; Dwyer & Buckle, 2009; Labaree, 2002; Ledger, 2010). Even when not participating fully, the ethnographer is likely to align herself with the identities of the research participants. While we were influenced by our own identities, institutions, and research aims, the research process brought us closer to the care workers' professional world and their identities in the workplace. It was through unravelling the institutional, professional, and individual aspects of their identities, but also through opening up our own selves and our own identities that we gained their acceptance. In other words, it can be argued that ethnographic work at its best took place only when we started negotiating who we were in relation to the research participants and vice versa. Negotiation of this kind seems to represent ethnographic reality more realistically than arguing for, or aiming to adopt, fixed and robust roles in the field.