Researching Social Work Practice Close Up: Using Ethnographic and Mobile Methods to Understand Encounters between Social Workers, Children and Families

Abstract

Research into child and family social work has largely stopped short of getting close enough to practice to produce understandings of what goes on between social workers and service users. This is despite the known problems in social worker engagement with children in cases where they have died. This paper outlines and analyses the methods used in a study of social work encounters with children and families on home visits where there were child protection concerns. It illustrates how mobile methods of walking and driving interviews were conducted with social workers on the way to and from home visits, and how the ethnography involved participant observation and audio-recordings of the interactions between social workers, children and parents in the home, revealing the talk, actions and experiences that occurred. Social workers often moved around the home, especially to interview children on their own in their bedrooms, and the paper shows how ways were found to stay close enough to observe these sensitive encounters within families' most intimate spaces, while ensuring the research remained ethical. Ethnographic and mobile methods produce vital data that advance new understandings of everyday social work practices and service users' experiences and of dynamics that are similar to breakdowns in practice that have occurred in child death cases.

Ethnography, mobilities, social work practice, mobile methods, children and families

Introduction

Although a large research literature now exists on child and family social work, surprisingly little of this has been applied to producing knowledge of what actually goes on when social workers and children and families are face to face. A key way such research can be done is by observing encounters between social workers and service users as they naturally occur—the method known as ethnography (Hammersley and Atkinson, 2007). At the heart of ethnography ‘is gathering data through participant-observation in a natural setting’ (Floersch et al., 2014, p. 5). The dominant approach in child and family social work has been for researchers to stay in offices and observe what social workers do there and how they talk about the work they do. ‘Natural settings’ where face-to-face practice goes on, like the family home, have been virtually ignored by researchers (Broadhurst and Mason, 2014). This contrasts with adults and mental health social work, for instance, where some researchers have got close enough to observe practice encounters (Floersch, 2002; Longhofer et al., 2010; Stanhope, 2012).

This gap in research and knowledge about the dynamics of social work encounters in child protection is all the more remarkable given that, for over forty years, reviews into cases where children have died have shown that social workers were present with children when they were injured but did not effectively engage with them. Research that gets close to practice to advance understandings of what social workers do and do not do and why is crucial to producing knowledge that can contribute to understandings of how to keep vulnerable children safe, promote their well-being and help parents. The aim of this paper is to outline and analyse the methods used in an ethnographic study of social work with children and families where there were child protection concerns. Child protection is an ‘intimate practice’ (Ferguson, 2011a) because of how it goes on in families' most intimate spaces and in how social workers must seek to get close enough to children through using talk, play, touch and listening to establish their experiences and work with parents. The aim of the research was to shadow social workers as they related to children and parents to find out what they actually do and how they do it. Such mobile research methods have also begun to be used to great effect to illuminate the lives of service users (Holland et al., 2011; Ross et al., 2009).

The paper shows how getting close to social workers' practice and staying close requires the adoption of mobile as well as ethnographic research methods. Practice encounters typically involved movement—journeying from the office to the child's home; walking and other movements within the home—and researchers must be prepared for such mobility and to stay close to the worker wherever they go. This includes into service users' most private spaces, like bedrooms, bathrooms and kitchens. The presence of researchers observing such sensitive, intimate encounters—typically in a child's bedroom—raises significant ethical issues and the paper shows how these were addressed in the study and considers the kinds of knowledge the methods can contribute to understandings of practice.

Researching social work practice: getting closer

Social work researchers have adopted a variety of ethnographic methods in attempts to reach different degrees of closeness to social work practice encounters. While the term ‘ethnography’ is often regarded as synonymous with participant observation, it also encompasses a repertoire of techniques that include one-to-one interviews, discourse analysis, personal documents and vignettes (Brewer, 2012, p. 302). Interviews with social workers and sometimes with service users have long been the most popular approach to understanding social work practice, gathering their views on interventions and decision making (e.g. Platt, 2006; Spratt, 2000; Hayes and Spratt, 2009). Case vignettes have also been used as a way to get closer to typical practice episodes and as the basis for interviews with workers about their responses to the presented scenarios (Forrester et al., 2008b; Spratt, 2000; Hayes and Spratt, 2009). While vignette and interview approaches produce some valuable findings, they are limited by the fact that they produce accounts of events that the researcher did not partake in and there is no way of knowing whether or how social workers achieve what they say they do (Forrester et al., 2008b, p. 25). Forrester and colleagues (2008a) sought to overcome this by using a ‘simulated casework’ approach where actors playing parents were recorded being interviewed by real social work practitioners. For Forrester et al. (2013), using simulations based on the same case vignettes enables research to control the practice scenario under consideration and produce insights into different social workers' skills and responses to fixed cases and communicative dilemmas.

In ethnographic studies where researchers have used participant observation, the attention has been on the occupational culture, how social workers talk about what they do with their service users, make decisions and the effects of office design (Pithouse, 1998; Scourfield, 2003; Broadhurst et al., 2010; Helm, 2014; Jeyasingham, 2014). This has also been done in tandem with interviewing service users (Buckley, 2003). This work has produced important findings about how managerialism, complex computerised forms and bureaucratic tasks limit the time social workers have to do quality work with children and families (Broadhurst et al., 2010; Gillingham and Humphreys, 2010). These studies were institutional ethnographies, focused on how the social work organisation structures practice and organises everyday experience (Longhofer et al., 2013, pp. 86–7).

In contrast, Longhofer and Floersch (2012) have defined research into face-to-face interactions between social workers and service users as ‘practice ethnography’. Its primary aim ‘is to explore the context, actions, thoughts and feelings generated by the structured relationships among practitioners and clients’ (Longhofer and Floersch, 2012, p. 305). This includes collecting what these authors call ‘experience-near vignettes of actual case management’ (Longhofer and Floersch, 2012, p. 305; see also Floersch, 2002). This means researchers being participant observers in the actual face-to-face practice of social workers. Longhofer et al.'s (2010) practice ethnography of mental health social work sought to understand how workers and service users related to one another and the knowledge, skills and practice wisdom involved.

Practice ethnographies are rare in child and family social work. Dingwall et al. (1983) did accompany social workers as they visited families, but did not give systematic attention to face-to-face encounters. Broadhurst and Mason (2014) include three brief observations of practice—two in the home and one in the office environment—set within a highly persuasive theoretical analysis and argument that the focus of research needs to be on deepening understandings of what occurs when social workers and service users are ‘co-present’. Hall et al. (2006) have sought to get close to practice by studying the language and talk involved through gathering audio-recordings of social worker–service user communications. No observations are involved, as either the researcher sets up the recording equipment in a fixed room in the house and then leaves or the practitioner does it themselves. The researchers then analyse transcripts of the talk using conversation analysis and discourse analysis to examine interactional sequences (see also Slembrouck and Hall, 2011; Hall et al., 2014). This approach has a great deal to offer understandings of language and social work conversations, but my argument is that the addition of participant observation of encounters—performed on the move as well as in fixed places—can provide even deeper insights into what is done and experienced, as well as what is said in practice.

This should not be taken to imply that ethnography can give straightforward access to the whole ‘truth’. Knowledge produced through participant observation is not an exact reflection of what happened, but a researcher's interpretation and, as such, is contestable. However, the researcher getting as close as it is possible to get to social work practice enables some things to be seen and experienced that otherwise would be missed (Longhofer and Floersch, 2012). This claim is based on my experience of conducting participant observations of social worker–service user interactions on home visits, the first phase of which I began doing in the late 2000s, which provided the basis for beginning to set out what social workers do and to theorise it (Ferguson, 2011a) and developing methods for researching it (Ferguson, 2011b). The second phase of observational fieldwork and a much larger study began in 2012, which this paper draws upon. There are signs of a growing recognition of the need for researchers to get right up close to real practice encounters. A recent mixed-methods study by Donald Forrester et al. (2013) included direct observations of real social work practice with service users to make sense of what practitioners do. In research in progress, Dr Sally Holland is leading a research team doing direct observations of social workers ‘talking and listening to children’ in real practice in the four countries of the UK (www.cf.ac.uk/socsi/research/researchprojects/talkingandlisteningtochildrentlcesrc.html).

In order for ethnographic observations of face-to-face social work practice to progress effectively, clarity is needed about the precise methods involved. Some researchers are developing a ‘practice-near’ research approach which they distinguish from ‘practice-distant’ research (Froggett and Briggs, 2012; see also Stake and Schwandt, 2006). Informed by psychoanalysis, ‘practice-near’ research considers the emotional and unconscious aspects of practices. Froggett (2012) used this ‘psycho-social’ approach to study the interactions between workers and young people in restorative youth justice projects. But ‘practice-near’ is also used to refer to studies that do not get near in the sense of physically close to practice encounters in real time. Ruch (2013) conducted ‘reflective case discussions’ with social workers to explore their experiences of communicating with children, without observing how the workers actually communicated with children. The next best thing for Ruch was to gather workers' accounts of what they did by getting as ‘near’ as possible to their subjective experience and emotional worlds. This produced important findings about how practitioners defend themselves against the high anxiety levels and emotional impact of the work by detaching from service users. My findings suggest that what is needed to produce unique data about the nature and lived experience of practice is an approach that observes face-to-face encounters, while staying close to the emotional experiences of workers and service users and using psychodynamic and social theories to make sense of those dynamics. The reasons for not focusing on actual practice encounters tend not to be mentioned by researchers and it is to Ruch's (2013, p. 6) credit that she provides a rare explanation: ‘Endeavouring to access more “direct” or “honest” accounts of what actually happens in encounters between social workers and vulnerable children, however, is fraught with ethical difficulties, not least determining how, and from whom, to obtain informed consent.’ This is very true: there are significant ethical and methodological issues involved. I will now show how my research study attempted to negotiate them.

The research study

The research set out to explore what social workers do and how they perform child protection, especially on home visits. Do they see children alone? If yes, where? If not, why not? How do social workers, children and parents relate to one another? What are social workers' lived experiences of the work and what enables and constrains practice that keeps children safe? The research design mirrored Longhofer and Floersch's (2012) approach of ‘practice ethnography’ and also drew on the new ‘mobilities’ paradigm in the social sciences (Sheller and Urry, 2006; Urry, 2007). Mobile research methods (Buscher et al., 2011; Fincham et al., 2010) are being developed to capture everyday movements, by, for instance, conducting interviews while walking (Clark and Emmel, 2010). In social work, Ross et al. (2009) used ‘guided’ walks and car journey interactions to research the everyday life of young people in care. ‘Guided’ walks involved a young person walking with a researcher, whom they led through places of significance to them, and the researchers also gathered data during car journeys—an approach that has been used in research with social workers (Ferguson, 2011b). Ethnography has also developed to take fuller account of the senses and the lived experience of the body on social practices (Pink, 2009). My research was also influenced by how Miller (2010) has developed the traditional anthropological approach of empathic observation into a ‘material culture’ approach that enables ethnographers to take account of the impact of the everyday objects that make up people's lives, such as houses, cars, home possessions, toys and so on. The research was designed to try and incorporate the ways that mobile and sensory ethnographies and material culture studies have opened up new ways for researchers to be present to capture the bodily and sensory experiences of research participants as they happen, while in movement as well as when stationary, and to take fuller account of the implications of where they happen, like on social work home visits.

In the mobilities approach, the researcher participates in social practices and patterns of movement, and observes and interviews people preferably during it, or afterwards, or both, about their experiences as they are having them, employing a range of observation, interviewing and recording techniques (Ferguson, 2008). Kusenbach (2003) refers to such a method as the ‘go-along’, which allows the ethnographer to observe their informants in situ while accessing their experiences and interpretations at the same time (see Stanhope (2012) for an example from mental health social work). This has similarities to what others call a ‘shadowing’ method that provides an opportunity to gain a contextualised understanding of how practices are performed (cf. Gilliat-Ray, 2011; Quinlan, 2008).

In my study, I conducted all the fieldwork myself. I went along with the social workers on their journeys from the office to see service users, interviewing them in the car and on foot about their plan for the encounter, expectations and feelings. I continued to interview the worker about their feelings and experience on the walk to the doorstep. I followed the social workers into the family home and observed and audio-recorded their practice encounters with children, parents and others present. The worker was then interviewed in the car immediately after on the return journey (or to the next visit) about the encounter and their feelings about it. These discussions also explored workers' general experiences and approaches to getting in, out and engaging with children and parents. I also regularly observed and recorded the interactions between workers and team managers before and following visits and interviewed managers.

The research took place in two local authorities in England, chosen to try to provide some diversity in terms of demographics and culture. Social work teams at both sites were involved who did short-term duty/investigative/assessment work and longer-term work with children and families. I based myself for three months in each local authority, where I spent on average two and a half days a week. While the most senior managers in the agencies agreed to allow access to their staff, it was up to individual workers if they wished to co-operate. It takes a lot of courage for managers and workers to allow their practice to be observed, analysed and written about, even when they know that all identifying characteristics will be changed. In total, twenty-four social workers were included in the study. Nineteen of the workers were women and five men. Three participants were from black and minority ethnic backgrounds. The social workers' length of service varied from newly qualified to some who had over twenty years' experience. The twenty-four social workers were observed and audio-recorded on a total of eighty-seven practice encounters: seventy-one on home visits, nine were interviews with children in schools and seven were office interviews. In a few cases, I did two home visits with the workers to the same families and with every worker in the study I sought to do more than one home visit to different families, which proved crucial to enabling a broader understanding of how they practised.

The research was funded by the Economic and Social Research Council (grant reference: ES/J006629/1) and granted ethical approval by my university and by the two local authorities involved. Only social workers and service users who consented were included. All the access to observations of practice was negotiated through the social workers, who asked for parents' consent for the researcher to accompany them to the visit. Some families were never asked, as their situations were regarded by social workers and managers as unsuitable for research visits. Some families who were asked did not wish to be involved and were never contacted or visited by the researcher. Where consent was given, on arrival, the social worker introduced me and I provided an oral explanation of the research. If they were happy for me to stay, permission was then sought to audio-record the interview. In none of the seventy-one home visits where prior consent was given to me visiting did a parent decline to allow me to stay and in just one case was consent to allowing audio-recording refused, but that same lone-parent mother still consented to the encounter being observed. All of them signed consent forms, a copy of which was left with them, with contact details should they wish to withdraw or make other inquiries. It was made clear to the service user that they could withdraw consent at any time, during the encounter or after it. The agreement with the agencies and the families included that, if there was the slightest indication that the researcher's presence was adversely affecting any family member or the social worker's safeguarding practice, then the researcher would leave. It was never felt necessary to do this, even though some quite dramatic and emotionally charged episodes were observed.

Observational data on the encounters were taken on the spot through hand-written notes, recording what was unfolding. Because the conversations were being audio-recorded, I did not have to hurriedly try to write down what was being said. This freed me up to draw ‘maps’ of the room(s), noting key interactions in terms of who was seated or stood where, who moved where and when, body postures, objects (toys, ornaments, furniture, etc.) in the room and workers' use of them, or not, and the presence of dogs and other pets. This was part of the method of drawing out the nature and influence of the material culture through which the practice went on (Miller, 2010). Field notes were also taken afterwards detailing further impressions of the journeys, practice encounters and what happened in the social work office involving colleagues and managers. I also tried to give names to the moods and tone of the encounters and atmospheres in the room(s). This meant practising an ethnographic approach where, rather than trying to eliminate the impact of the researcher's presence, the researcher engages reflexively with their own position, drawing on the understanding which that position offers (Shaw and Gould, 2001).

All the audio data were transcribed and analysed using thematic, grounded theory and narrative methods. A key finding concerned the relatively high number of young, pre-school children involved. In England, government guidance requires social workers to see children on their own where there are child protection concerns. In almost a third of the cases, social workers regarded the children as lacking the verbal capacity and understanding to be seen on their own (Ferguson, 2014). This made an ethnographic approach that went beyond the analysis of conversations even more important, as observation made it possible to illuminate how workers related to children of all ages through the presence or absence of touch, play, as well as talk. A case study method was also adopted by bringing together the social worker interview data from before and after the visit, the audio-recording of the interactions and the observation and field notes. Triangulation of that data provided for (seventy-one) case studies of home visits. Cross-case comparative analysis enabled the drawing-out of themes, which could be illustrated by detailed case studies. Taken together, this mixed-methods approach enabled visibility and structure to be given to the face-to-face encounters in terms of mapping patterns of casework in how social workers reached and engaged with children, or did not do so; social workers' embodied experiences of relating to children and parents and home visiting; and critical analysis of practice to provide theory and deeper understandings of how things did and did not go as planned.

Researching social work on the move

The purpose of this paper is not to present the research findings in any depth, which has begun elsewhere (Ferguson, 2014). Rather, to meet its primary aim of considering the methodological issues, I will now break down the discussion of the practical application of the methods into the ways they sought to inquire into the social workers' practices in specific contexts, focusing on the (car) journey, walks and the service user's home.

Journeys

Interviews with social workers while travelling with them to see families provided specific information about the case being visited, what they planned to do, and what they were experiencing and feeling in that moment. As workers anticipated what seeing this family meant to them, thoughts and feelings were captured:

I'm planning in my own head and thinking: right, where do I start, what do I say; will the children be there, won't they, you know; will they be at his mum's, or not. I guess, what am I, what am I going to do? Well, I want to have a look around really. I want to see, I think the state of the house is going to say a lot, so that will be, that will be something. So, I guess all this is going through my head: what do I want from today and what am I going to, what am I going to say to him?

Observing and interviewing workers as they conducted their work in the car and on foot provided fresh insights into two hitherto unresearched areas: how crucial the (car) journey is as a space within which the worker has an opportunity to prepare for the encounter; and, second, the state of mind and being workers are in when they arrive at the doorstep to see children (see also Ferguson, 2009). Some felt very stressed by getting lost; fearful because they knew the parents did not want them there; anxious at going into the unknown; while some were excited because they liked the family and took pleasure from helping them.

The decision to keep the audio recorder running after we left the car and conduct walking interviews on the way to finding the home brought into play the impact on workers of influences like strangers, the weather, darkness and the time of day, and dogs. This is typified by the following exchange during a walking interview on a first visit to a case of alleged neglect where the social worker, Jenny, went to what she thought was the front gate but was in fact the back entrance, which gave access to the garden:

[Dog barking]

SW:

I do hate visiting families with dogs.

HF:

Pardon?

SW:

I do hate visiting families with dogs. It's my biggest bugbear. Why does everybody who's got a dog think that people like them?

HF:

Yeah.

SW:

That's [house number]. I think it's down there.

[Dog barking]

SW:

Oh.

HF:

I think this is the back door.

SW:

Yeah. [Dogs barking loudly] Can we go round? [More barking] And remembering people's names is a challenge as well [small laugh]; especially if it's the first visit, like the first I've read about them.

The fact that the researcher has observed the immediacy of the experience before and during the home visit provides a rich agenda after it for exploration of what workers experienced and did. The car was found to be a productive space for these interviews because of the privacy it afforded and the absence of interruptions that often occur in offices. Crucially, workers do not have to stop what they are doing to be interviewed and the inquiry includes consideration of the very work they are in the process of doing. But, even when they did stop driving and had parked up back at the office, it was common for social workers to be happy for interviews to continue in the car, sometimes for lengthy periods.

Researching practice up close in the home

Once in the home, several research challenges presented themselves. Fully observing face-to-face practice on home visits proved complicated because of the ‘private’ places the workers typically went. A key finding was that the practice was deeply investigative, with social workers routinely inspecting bedrooms and kitchens to assess standards of care. On the first two cases I shadowed, I sat down in the sitting room and remained there when the social workers moved around checking upstairs and in the kitchen. On both occasions, I could hear what went on in the kitchen because it was the adjoining room and the door was open, but I could not see what was going on and had no idea what went on upstairs. I chose not to follow the workers due to a sense of anxiety about intruding further upon family privacy. Such a concern was legitimate but, instead of making assumptions about how service users felt, I learned the importance of the researcher finding out by asking them and seeking their ongoing consent to go wherever workers do.

Subsequently, with every social worker in the study, I gained their consent to go wherever they did, unless they requested me not to. In the event, none of the social workers ever suggested I don't go anywhere with them. And, when asked for their consent for me to accompany social workers upstairs and to other parts of the home, families gave it. On the sixty-nine home visits that were done after the two initial ones just referred to, social workers went upstairs on thirty-seven and I went with them. Crucially, this was not always done just to inspect the home conditions. On fifteen visits, social workers interviewed children alone at home and nine of these took place in the children's bedrooms for all or part of the time (the others were in sitting or dining rooms and in gardens). This shows how some very important practice went on upstairs and without staying close to the social workers and following them everywhere as they moved crucial insights into what they did and children's and other family members' experiences would have been missed.

The social worker Jenny, whom we left above at the doorstep feeling anxious because of the presence of dogs, was allowed access to the home by the mother as she restrained her two dogs. Jenny managed to maintain her composure in the sitting room despite the dogs' sniffing at and snuggling up beside her. The two children aged six and ten were present and Jenny took control by initially asking to see mother on her own and for the television to be turned down. After twenty-seven minutes of discussion about her parenting and alcohol consumption, Jenny gained the mother's consent to see the children on their own. She saw them separately in their bedrooms, the ten-year-old for eleven minutes and the six-year-old for twelve minutes.

The social worker introduced herself to each child and explained her job. She worked hard to put the children at ease, telling them she was not there to take them away, but to find out more about their mother's care and drinking so she could support her to get help if this was needed. Here she is at the midway point with the ten-year-old:

SW:

And do you ever worry when mum drinks?

Ten-year-old:

[Pause] About 20% I worry.

SW:

Oh okay. And what's that 20% then? What makes you worried?

Ten-year-old:

Well, she can be like OTT like.

For the interviews with both the children, the social worker sat on the single bed, having asked them if it was okay to do so. She leaned towards the children with an open stance, making eye contact and responding to them in an animated manner. When seen alone, the six-year-old stayed much further from the bed than her brother had and leaned back against the radiator on the wall, taking support from it. Jenny adjusted her language and style to make it appropriate for the younger child. She also paid attention to the children's things—DVDs, toys, musical instruments—and incorporated them in a playful way into her communication. In Miller's (2010) terms, this was a creative practitioner who knew how to use children's material culture to help build rapport and begin to understand them. On some practice encounters in the study, practitioners spent small amounts of time with children, did not use play or any communication aids with them but relied on talk, which limited rapport building, the creation of intimacy and access to the child's experience, not least with younger children (Ferguson, 2014). In observing such encounters in intimate spaces like the children's bedrooms, the social workers introduced me and explained my role and sought the children's consent to my staying. I sat on the floor, literally shrinking myself to try and be as non-intrusive as possible.

Staying physically close to practise in families' most intimate spaces also provided important insights into how on occasions social workers do not do as they are expected to and what they intended to. In a small number of cases, they were observed not relating to children at all or achieving what they defined as effective work. On one home visit, a social worker faced by an angry mother denying child neglect completely ignored the two children, aged two and five, who were right in front of her. It was only on the way back to the office in the car through being asked open-ended research questions—such as ‘How do you feel you related to the children?’ and ‘Did you consider speaking to the five-year-old on his own?’—that the worker began to realise that she had not actually engaged in any way with the children. She clearly continued to reflect on it while on the journey as, on arrival back at the office, as she was getting out of the car and referring to the five-year-old, she said: ‘I don't know why I didn't talk to him’.

These case examples illustrate another key finding about the value added by participant observation and practising sensory ethnography (Pink, 2009) in how it enables the researcher not only to see and hear what is happening, but to sense atmospheres in rooms and moods in how people are relating. This provides for what Froggett and Hollway (2010) call ‘scenic reconstructions’ where the ‘gestalt’, the intrinsic emotional texture of the encounter, can be brought to mind and articulated. Their notion of gestalt has some similarities to what elsewhere I have tried to evoke in terms of ‘atmospheres’ (Ferguson, 2010, 2011a). Reconstruction of the ‘scene’ in the example just given shows a worker who had arrived at the home feeling stressed due to the organisational requirement to get cases assessed quickly, which was compounded by the fear she felt when experiencing an angry parent. The resulting atmosphere of tension caused the worker to become emotionally defended and so anxious that she was unable to think about the children. Here, we see how the kinds of complex psychodynamic processes skilfully drawn out by Ruch (2013) after the event are manifested in face-to-face practice. This is an example of how the creation of ethnographic case studies provides vital data that can illuminate experiences and processes that are similar to the dynamics of breakdowns in practice that have occurred in high-profile child death cases. However, the research methods and underpinning theoretical perspective adopted here allow for a much deeper analysis and level of understanding than is typically found in such reports.

Conclusion

In this paper, I have tried to show how observations and audio-recordings of practice encounters wherever they occurred in the home and walking and driving interviews can produce original knowledge about how social work is done, or not done effectively. I have also tried to show how such research can be done ethically. In terms of helping service users to understand their right to deny me access and consent, I was mindful of power dynamics and that families could perceive me as an extension of the social work department. Following Miller and Bell (2002, p. 65), I tried to behave in ways that signalled my difference and independence from the workers. I made it clear that I was there to observe, so never spoke or joined in, unless spoken to and, even then, kept my replies politely short. My silence on one occasion throughout a forty-five-minute home visit was probably why one six-year-old girl came up to me as we were leaving and asked ‘Are you a grump?’. I smiled and happily repeated what I had said at the beginning of the visit in preparing them for my silent presence.

What Guillemin and Gillam (2004) call ‘ethically important moments’ inevitably arise in such sensitive research. Staying close to social workers meant walking into whatever they did—like the occasion when a social worker entered the bedroom of a seventeen-year-old young man and he was in bed at 2 pm. Although the social worker stayed and talked to him, out of respect for his privacy, I left the room. It is very easy for research ethics committees to be risk-averse and refuse to support such sensitive research without properly assessing the risks (Murphy and Dingwall, 2007). It is essential for research to be scrutinised for the presence of ethical procedures which try to protect participants from harm, that ensure they can give informed consent and have exit routes out of studies. But my study suggests that a researcher can be present at child protection encounters without having a negative impact on practice or service users. Some social workers even told me they felt their practice was improved by my presence. The work is difficult and complex, is usually done alone and they felt more secure having the company. Most said they found participation useful because it made them think, reflect and helped them to learn.

The consequences of preventing such research go well beyond its implications for researchers and agency staff. As Sarah Ashencaen Crabtree (2013) shows, practice ethnographies are crucial to unearthing service users' experiences of services, producing learning about what makes a positive difference to vulnerable people's lives and vital knowledge about problematic and even abusive practices that otherwise may remain hidden. By its very nature, this kind of close, shadowing ethnographic research cannot control, erase or avoid stepping into potentially compromising situations. The critical issue is for researchers to be clear about what constitutes ethical research behaviour, to reflect very hard on it on the spot and to have a repertoire of responses at the ready to respond appropriately to ethically important moments when they arise.