Reflections on the procedural and practical ethics in researching professional social work writing

This paper explores the procedural ethics and ethics in practice involved in a multi-agency research project exploring professional social work writing. Drawing on institutional documentation and researchers’ field notes over two years, the requirements and processes involved in complying with academia-facing and agency-facing regulatory frameworks are summarised and challenges highlighted. The main part of the paper centres on ethics in practice , foregrounding the importance of ongoing dialogue between researchers and participant-stakeholders. Reflective accounts from three participant-stakeholders illustrate the interrelationship between what are often presumed to be distinct moments of a research chronology – access, representation and dissemination – in the process of knowledge making. The paper concludes by highlighting the differences between academia-and agency-facing procedural ethics, and arguing for greater institutional recognition of ethics in practice, in particular the importance of ongoing dialogue between researchers and participant-stakeholders at all stages of the research process.


Introduction
Any academic research project seeking to explore practices in professional domains needs to comply with the regulatory procedures of academic institutionsfunding agencies and the researchers' universityand of organisations or agencies where the professionals are located.Regulatory procedures and fulfilment of these are codified in written texts, setting out the terms and conditions under which research can be carried out and the researchers' agreement to comply with these.The importance of such regulatory practices cannot be underestimated: they are part of a history of seeking to protect individuals from potential researcher abuse, traceable in the 20 th century to the Nuremburg trials and Nazi abuse and torture in concentration camps, often in the name of 'science'.Subsequent statements of ethical principles governing medical research, the Nuremburg code (1947) and the Declaration of Helsinki (1964) are premised on the discourse of human rights, foregrounding key notions of risk (minimising risk to participants) and benefits (indicating potential benefits to participants and the wider community) (Annas and Michale 2008;Strohm Kitchener and Kitchener 2013).Such principles have formed the basis for the many research regulatory frameworks developed at national and international levels across academic disciplines ( in relation to Social Sciences, see https://www.acss.org.uk/developing-generic-ethicsprinciples-social-science/academy-adopts-five-ethical-principles-for-social-scienceresearch/).
However, fulfilling such requirements may be particularly challenging for researchers engaging with the regulatory frameworks of both academic and professional organisations, as is essential in the case of applied linguistics research in professional discourse.Furthermore, engaging solely or predominantly with ethics through the lens of codified regulatory frameworks is problematic for a number of reasons.Most obviously, ethics tends to be positioned as being at one point in time (around securing access and participant consent), masking the considerable ongoing ethical work (time, labour and researcher-participant discussion and negotiation) at all points in the research process (Markham and Buchanan 2015).At a theoretical and ideological level, research ethics refracted through regulatory frameworks of contemporary neoliberal higher education is often underpinned by an audit approach, whose primary function seems to be to maintain institutional reputations and ensure protection from lawsuits (see Denzin 2017;Wilson and Hodgson, 2012).Within this model, knowledge tends to be commodified as 'data' that can be abstracted from people, their specific interests and histories of production and engagement, illustrated (in part) by the requirement of funding bodies to archive all data generated so that it can be 'used' by other researchers (this is of course a much larger debate, see Hammersley 2010;Russell and Barley 2020).This commodified orientation to knowledge is at odds with the premises of much applied linguistics research in professional practice, where participants are viewed not as 'subjects' (that is, passive 'objects') of research from whom 'data' is 'obtained' but rather as stakeholders with profound interests in the goals, processes and outcomes of the research (Sarangi and Coulthard 2010).
The aim of this paper is to offer a descriptive account of some of the key ethical dimensions to researching in professional domains, using the WiSP project "Writing in Professional Social Work Practice in a Changing Communicative Landscape" (www.writinginsocialwork.com) as an illustrative case study.The value of such description is to put the spotlight on key ethical dimensions that are routinely experienced in professionally-oriented applied linguistics research but often remain invisible.The paper foregrounds the perspectives of three participant-stakeholders who offer brief reflective accounts of their involvement in the research process.
The paper adopts an intentionally broad approach in order to make visible the breadth and range of ethical dimensions across the lifespan of a project, using the categories of procedural ethics and ethics in practice (Guillemin and Gillam 2004) to structure the reflection and exploration.The key questions the paper explores are: 1. What are is the nature of the procedural ethics needing to be enacted to bring a project such as WiSP into being and what are the differences between academiafacing and agency-facing regulatory requirements?2. What is the nature of ethics in practice in a project such as WiSP and in particular the role of researcher-professional dialogue?
The paper begins by offering a brief review of the key ways in which ethics is framed in applied linguistics and discourse studies and their usefulness when applied to the particular professionally-oriented study in this paper, WiSP.This is followed by a summary of the research questions and datasets generated in the WiSP study.The main part of the paper centres on a consideration of the procedural and practical ethics involved in the study, drawing on documentation, researcher field notes of two years of negotiation and reflective accounts of three participant-stakeholders.Analysis and discussion foreground the interrelationship between what are often assumed to be distinct moments of a research chronologyaccess, representation and disseminationin the process of knowledge making.
The paper concludes by arguing for greater institutional and researcher recognition of ethics in practice and the need for a research praxis that foregrounds the importance of ongoing researcher-professional dialogue at all stages of the research process.

Characterising ethics in applied linguistics research
There are a number of ways of characterising the ethical dimensions to research in the fields of applied linguistics and discourse studies.Here we foreground three.The first centres on distinguishing between institutional-level requirements (such as regulatory frameworks enacted by research and ethics governance committees and institutional review boards) and the situated practices that constitute the research process.The term 'macroethics' has been used to characterise the former, 'microethics' the latter (Guillemin and Gillam 2004;Kubanyiova 2008) with the terms procedural ethics and ethics in practice often used to capture this distinction (Guillemin and Gillam 2004;Biros et al 2010;de Costa 2015).
A second way is to characterise the ethical stance according to the particular epistemological paradigm in which particular researchers locate themselves: thus researchers working within a positivist paradigm are typified as treating ethics in much the same way as regulatory bodies, that is as a set of codified procedures governing the research, which should be robust but relate primarily to a particular moment in a research chronology, with the aim of justifying the research and securing access (de Costa 2015).In contrast, work in postpositivist traditions tends to frame ethics as being at the centre of all research activity and processes (for discussion, see Hammersley and Traianou 2012), raising fundamental questions about the positioning of researchers and participants in terms of power and the knowledges that can be generated (Copland and Creese 2015;Tagg et al 2016).
A third wayand linked to postpositivist and critical research traditionsis to specifically foreground the relationships between researchers and research participants in terms of the ideological purpose of the research: a highly influential framework in applied linguistics has distinguished between three ideological models of research; a) ethical researchresearch on subjects; 2) advocacy researchresearch on and for subjects; 3) empowering researchresearch on, for and with subjects (Cameron et al.1992; for an overview, see Mallinson

2018).
In exploring the ethics involved in the WiSP study, the aforementioned characterisations prove helpful in specific ways.Whilst beyond the scope of this paper, the foregrounding of the relationship between specific epistemologies underpinning applied linguistics research paradigms and ethics has been a central debate within the WiSP team, informing decisions around specific methodologies (Lillis et al 2017(Lillis et al , 2020)), in particular with regard to corpus linguistics (Leedham et al.2021).
In the context of this paper, the categories of procedural ethics and ethics in practice are useful in structuring the discussion around the key ethical dimensions of our research whilst also signalling the need to explore their interrelationship.The tripartite classification of researchon, with, and forserves not so much as a way of characterising the specific ideology underpinning the whole research project, but as a reflective tool for charting the different kinds of purposes and relationships driving a study at different moments in time, a point we return to in Section 6.

The WiSP project
WiSP, a UK nationally funded research project, was established in response to both academic and professional imperatives: the gap in empirical research about professional social work writing (academic); the range of concerns and criticisms routinely expressed about social work writing by social workers, service users and by those regulating the profession (professional) (see Lillis et al 2017Lillis et al , 2020)).The project grew out of a specific request to collaborate with an agency where managers had concerns about the case notes written in Adult Services (Lillis 2017) and subsequently developed into the WiSP project involving six agencies across the three core domains of social work practice-adults, children and mental health.WiSP therefore builds on previous studies carried out by some members of the research team and on the relationships established with specific social workers and social work agencies.The project set out to explore the nature of contemporary social work writing and to define the 'problematics' surrounding such writing, using an overarching ethnographic orientation involving multiple methods (see Lillis et al, 2017Lillis et al, , 2020)).A key goal was to document the range of writing taking place, the purposes of writing in driving decisionmaking and actions and the meanings attached to writing by social workers.'Writing' (often referred to as 'recording' or 'paperwork') refers to all types of writing carried out by social workers relating to any aspect of their work and includes the most visible aspects of writing, that is the texts produced as institutionally-mediated digital systems e.g.case notes, assessments, emails, as well as routine but institutionally-backgrounded practices such as handwritten notes on pieces of paper, notebooks, post-its (Lillis et al, 2017(Lillis et al, , 2020)).Table 1 outlines the main research questions and the data collected.

INSERT TABLE 1 HERE Table 1. WiSP Research questions, empirical focus and data generated
The research to date has led to publications oriented to academia and social work practice (e.g.Leedham et al 2020Leedham et al , 2021;;Lillis 2021;Lillis et al 2017Lillis et al , 2020) ) and to the generation of open access resources for educational purposes (https://wisper.writinginsocialwork.com).

Procedural ethics: Academia-and agency-facing
As is typical in research centring on professional domains, ethical governance procedures relating to WiSP were both academia-and agency-facing.The aim of this section is to make visible the key dimensions of procedural ethics, drawing on a two-year researcher diary and an overview of the documentation required by the institutions.
Academia-facing procedures involved fulfilling ethics and governance requirements laid down by the university and the national research funding body, providing details about the ethical principles and practices that would underpin the study and covered areas such as: research methodology, data to be collected, recruitment and consent procedures, secure storage of data, data protection, potential risks and benefits (to individuals and participating agencies) and agreement to conform with the requirement to archive all anonymised data generated with the funding body.Additionally, before agreeing to fund the project, the national funding body required provisional agreement statements from participating agencies.
Even though three agencies provided such statements, one grant reviewer considered that the ethical challenges to securing access were so great that there was a 'high risk of one or more local authorities withdrawing consent'.In order to take account of the possibility of withdrawal of consent by any single agency and to ensure that such withdrawal would not jeopardise the project, the researchers extended negotiations beyond the originally planned (and provisionally agreed) three agencies to six agencies.The procedural requirements relating to both academic bodies and agencies were codified in a substantial number of written documents (see Figure 1).

INSERT FIGURE 1 HERE Figure 1 Codification of procedural ethics: Documents required for fulfilling procedural ethical requirements
The procedural requirements, codified in written texts, were similar across the academic bodies and the agencies.However, there were additional requirements by the agencies: named 'sponsors' from each organisation were requiredpeople who were willing to advocate for the value of the research; researchers needed to produce an Enhanced DBS //full form// certificate from the Home Office to evidence any criminal record; and, with one agency, an additional legal 'data sharing' agreement was required, involving considerable time and negotiation.

1
A more fundamental distinction between the academic bodies and the agencies was that whilst fulfilling the regulatory requirements of both was essential, these were insufficient for proceeding with the research with the agencies.In order for access to be secured with the agencies, permission at two other levels was also required: the 'operational' level (a term used by the agencies to refer to the agreement from service managers confirming that participating in the research would not impact negatively on everyday work being carried out) and the individual social worker level (i.e.agreement in principle from individual social workers to participate before deciding whether to give formal consent and actually participate).The processes around securing permission at both levels were informal, in the sense of uncodified, and any success to a large extent dependent on pre-existing relationships and the networks these facilitated.For example, in agency 1 (Table 2), one researcher had previously worked extensively with a manager who was instrumental in both securing the support of his senior managers (including those involved in decision making around ethical and governance procedures) through informal discussions as well as encouraging, again through informal conversations, social workers to take part.In agency 3, one social worker had participated previously in related projects on writing and actively advocated for the value of the project to both her managers and her colleagues.Where pre-existing relationships, or ties, between the researchers and the individuals in the agency were weaker, access was harder to secure. 2   Where permission (both formal and informal) was not agreed at any one level, access was refused.In domains such as social work, where workload tends to be heavy and resources limited, decisions as to whether it is feasible or desirable to participate in a research project can change at any moment in time, at any level, overturning pre-existing agreements.Thus, agency 2, with whom the researchers had previously worked and who were very keen to participate, agreed to take part at level 1 (organisational) and at level 3 (individual), reluctantly withdrew after 7 months because of pressures at level 2 (operational level): the managers considered that the immediate workload that the social workers were carrying was too great to cope with the additional burden of participating in research (material, i.e. taking up social worker time, but also emotional, i.e. potentially generating additional stress).
Furthermore, even where permission was secured at the three levels, additional specific conditions were laid down by each agency about exactly which people/places/material the researchers could access.Thus, whilst the research intention had been to collect the same four main types of data from each participating agency (texts, interviews, field notes based on observation, social worker logs) different agencies agreed to allow researcher access to different sources of data; these ranged from what was considered by the agencies to be the least risky (interviews with social workers) to the most risky (copies of service users' written records).Considerable discussion and negotiation centred on the latter because of the need to remove all personal data. 3Some agencies considered that agreeing to the sharing of written records, even with all personal data removed, was not only problematic because of potential errors in anonymisation leading to breaches of data protection legislation, but also logistically being too complex.
Following agreements with all agencies, written records had to be anonymised before being shared with the researchers by establishing an anonymisation coding system, such that identifying information was deleted whilst relevant contextual detail was maintained (for example, replacing a name of a person with CHILD or SU for service user or the name of a particular place with a descriptor, e.g.SCHOOL) which was far from straightforward (for details of anonymisation and implications for analysis, see Lillis et al 2017;Leedham et al 2021).A pivotal role was played by the highly supportive insider participants in both advocating for the value of making texts available for research and finding logistically feasible solutions specific to their organisations: where such support was strong, as in agencies 1 and 3, it was possible to collect a wider range of data.
The outcome of the negotiations led to access and participation across a continuum of what can be described as high-medium-low, indicating the key data types being accessed.Table 2 briefly summarises the three levels of permissions required, the timescales in negotiating permissions, outcomes and final levels of access secured. 4

INSERT TABLE 2 HERE Table 2 Levels of permission, access and participation
Given the highly sensitive nature of the data the researchers were seeking to access, securing differential access across the agencies was viewed a success, enabling the researchers to work with a rich, if complex, dataset on professional social work writing.
There are two issues we wish to underline at this point: • Substantial research labour (time, knowledge, negotiation of relationships) is involved in fulfilling regulatory procedures required by both academia and agencies.Such labour is essential (research cannot happen without it) but is often invisible (i.e.funding/institutional allocation of time for such work is far lower than is actually required); • Procedural ethics involves similar processes in academia and agencies: both involved practices codified in written texts in order to secure the approval of regulatory bodies.However, a key difference is the centrality of informal relationships underpinning such codified practices in the case of agencies, where specific histories and networks of relationships, involving informal negotiation and ongoing dialogue were crucial to securing and carrying out the research project.The importance of such relationships as part of ongoing dialogue is emphasised in the following section.

Ethics in practice
In this section we focus on ethics in practice, drawing on Sarangi (2019) to outline a heuristic we consider useful in highlighting the significance of dialogue between researchers and participant-stakeholders throughout the research process.We offer reflective accounts written by three participant-stakeholders who have engaged in the WiSP project and whose contributions have helped to make the research possible and to generate research findings potentially meaningful and 'useful' in professional domains (Sarangi 2012).Each reflective account is briefly introduced and situated within the framework outlined below in 5.1 and followed by summary analytic comments.

A framework for exploring ethics in practice
We take as given the fundamentally collaborative imperative driving applied linguistics research concerned with professional practice (see Candlin and Sarangi 2004;Candlin and Candlin 2016), which necessarily places listening and dialogue at the centre of the research process.Sarangi (2019) offers a framework for making visible the communicative dimensions at the heart of such collaboration, mapping these against key research 'moments'.
Ethics of access refers to the many aspects relating to being given permission to carry out research in a particular sitethe choice of research site, methodologies, specific questions - We draw on Sarangi's four dimensions as a heuristic for articulating ethics in practice, 5 signalling the importance of ongoing listening and dialogue between researchers and participant-stakeholders, the importance of which is often signalled in professional discourse studies in terms such as collaboration and reciprocity (see Candlin and Sarangi 2004: 4;White et al 2019).Figure 2 represents Sarangi's dimensions along four axes to underline the importance of viewing conventionally marked research moments as contingent upon each other, rather than as discrete or as in a linear chronology.We have added the dimension of 'ownership' to signal the significance of both actual ownership (e.g. who legally 'owns' the data) and, what can be described as symbolic-ethical ownership.The actual ownership of data generated from WiSP resides with the university and the national funding body although in immediate practice, the researchers are those most likely to work with the data generated and decide how it should be used.However, symbolic-ethical ownership can be attributed to immediate participants in the research as well as interested stakeholders beyond the project: the former make the research possible and have a direct interest in how data might be used and the latter have interests in the dissemination of particular knowledge, as foregrounded in section 5.4. 6

INSERT FIGURE 2 HERE Figure 2 A heuristic for articulating ethics in practice
In this paper we use the heuristic to illustrate the interrelationship between what are often presumed to be distinct moments of a research chronology, underlining the central contribution of participant-stakeholders to making the research possible and for the knowledge generated.The pivotal role played by participant-stakeholders and their perspectives on the considerable work they do are rarely included in published accounts (numbers on the heuristic correspond to the three participant-stakeholder accounts).

The ethics of access (within the context of participation, ownership, representation and dissemination)
The formal, procedural challenges of securing access to social workers, the texts they write and the contexts in which writing takes place have been signalled in 4. The importance of informal relationships and of particular participants to facilitating access as well as the ongoing research is often underlined through reference to 'key informants' or 'guides' (Barton and Hamilton 1998: 62).The reflective account by one professional social worker illustrates the ways in which participants take on multiple roles throughout the research process, acting as sponsor (crucial to securing initial access), participant (crucial to securing relevant 'data') and guide (crucial to enabling the researcher to understand and negotiate specific institutional contexts).

Reflections from Gillian Lucas a social worker -sponsor, guide, participant
I have been a qualified social worker for 14 years and have been involved with several related projects about writing with some members of the research team over a period of 10 years.I was excited to participate in the projects in the hope that the findings would influence future social work writing and education so when I was asked to be involved in the WiSP project I saw this as continuing with my interests.
Reflecting back about my role I am surprised at the differing tasks I managed and the challenges I faced.These included navigating my way around the service to identify the correct person to consent to the project, to facilitating a meeting with the strategic commissioning team to agree the finer detail of the research project to ensure that no breaches of confidentiality took place.I spoke to multiple senior managers to get the project underway which took months.(1) 7 Recruiting colleagues to participate was time consuming, involving many discussions: colleagues needed a lot of reassurance.Throughout the project my role was to coordinate the collection of the 'writings' and relay any difficulties, as well as coordinating (and doing much of) the anonymisation of all written data.Keeping participants on track was challenging as some colleagues needed cajoling and chasing, while others just got on with the task.Another difficulty was trying to manage the competing agendas.Social workers were doing additional work for the project, such as collecting their writings while carrying a heavy case load.I felt I had to help the research team to understand the multiple tasks social workers routinely carry out, alongside the burden of participating in the project, and while there were some tensions at times, they were open to listening and a suitable compromise of what realistically could be captured within 'real time' was reached.The particular request to spend time observing colleagues raised some concerns: colleagues found the idea of having someone with them for whole days, watching and asking questions, difficult, but after the observations took place, feedback was generally positive.I thought participants might really struggle with this part of the project but surprisingly this was not the case.I wonder if this might have been because social workers are used to working with lots of different people.(2) Some colleagues were worried about how the data would be used by the researchers and who might access and use the data in the future.Some had deep concerns that even if all personal data was removed from written records, some circumstances were so specific that people might be identifiable.A compromise was reached where it was agreed that colleagues would be able to review any text data intended to be used for publication and could veto the broader circulation of any text.Discussions about potentially identifiable parts of records fed into decisions we made about coding: devising a coding system that everyone could work with was not straightforward.(3) Social workers are fully aware that we spend far too much time on writing, however, we were all surprised at the actual time we were spending within our roles outside of our contractual hours.The research validated our concerns around the demanding role and the unrealistic demands put on us.Sadly, it has not made any difference thus far.(4) Where I feel I have made a difference is in contributing to resources based on the project where I've been able to emphasise an issue that is important to me and the authority I work for: the importance of writing to the child, rather than about the child (see https://wisper.writinginsocialwork.com).When writing in this way, wording becomes sensitive and warm, it removes jargon as the author needs to picture having a conversation with the child paying particular attention to the use of language.(5) The account by Gillian Lucas draws attention to the ongoing challenges and responsibilities held by participants who play pivotal roles in not only enabling access to a particular domain, but in sustaining and therefore making possible an entire research project: managing complex relationships with both colleagues, e.g.reiterating the value of the research and therefore its worthiness of colleagues' time and attention, and researchers e.g.drawing researchers' attention to frustrations and misunderstandings (1, 2); overcoming logistical challenges, e.g.co-designing anonymisation codings and thinking through the relevance of the researchers' analytic gaze (3).Given the considerable workload pressures faced by social workers, the mediating position of the person taking on such multiple roles is potentially onerous but crucial.
A further important aspect that the above account illustrates is the higher level of scrutiny of data use required of researchers by participating social workers than by both academia and agency-facing regulatory committees.Social workers were particularly concerned to avoid possible re-identification of anonymised material (3) through what is sometimes referred to as 'jigsaw identification' i.e. a reader identifying a particular person by connecting the details in different bits of data (http://www.transparencyproject.org.uk/jigsaw-identification).At a more fundamental level, the social worker participants actively sought to ensure that the researchers would treat written records with respect and care, that is, not as a research object, 'data', but as living documents with consequences for people's lives.For example, a condition social workers (in contrast with agencies) underlined in order to agree to share records was that only a limited level of (anonymised) detail about any one individual would be placed within the public domain (in publications and in public archiving).The ethics of access therefore involves negotiation and dialogue that goes well beyond procedural ethics.

The ethics of analysis/representation (within the context of ongoing participation, ownership and dissemination)
Once material has been collected and converted into 'data' (for discussion see Gitelman 2013), decisions have to be made about exactly what to focus on, how and why.Whilst the research process is often construed as data leading to findings, with warrants for 'findings' resting on largely methodological (and/or) theoretical justification, researchers necessarily which aspects of a phenomenon to focus on and which data to analyse.In applied linguistics studies concerned with understanding 'critical moments' of practice (Candlin 1989(Candlin , 1997)), engaging with the concerns of participant-stakeholders is a key ethical imperative.A clear example of the researchers in the WiSP project placing their analytic attention on a concern of high importance to participant-stakeholders was the decision to focus analytic energies on time, already mentioned in 5.2 and emphasised by many participants.The reflective account by another social worker further underscores professional concerns about time spent on writing and their interest in theorising the relationship between time and practice, through a focus on theories of bureaucratisation.

Reflections from Michael Balkow, a social worker -participant, author
I decided to take part in the research project because I have often found there is a distance between social work practiceestablishing relationships with service users, their support network, then problem solving within these complex networksto writing about the work undertaken.The two domains, if we refer to them as fieldwork and office work are of course inextricably linked but the importance placed upon the two seems to be heavily weighted in competing and different directions.For example, a family may feel that a social worker really understood or empathised with them on a home visit, leading to a significant change in their life, however a social work manager may be more interested in how quickly and densely detailed this was documented.The relationship between the twosocial work writing and social workalthough interdependent can also be antagonistic, as many social workers have felt that one stifles the other, spending so much time writing about service users, that there is little time left spent supporting them.Therefore, the focus on time and writing is a highly relevant place to begin an enquiry into the nature of social work writing.( 1) The anxiety generated around the need to spend time on recording starts in social work training.I recently supervised a student who was panicking upon the completion of her placement, that all her records were not up to date before she left.She said that another worker had told her: "if it's not written down, it didn't happen," a phrase which is used a lot and I know the WiSP researchers have quoted.It reminds me that the opposite is true: as Tom Grimwood states in his book Key Debates In Social Work and Philosophy, (2016) "if it happened, there is writing" (p118) which I think exposes some of the fallibility of written recordsthey are almost always lacking and defined just as much by what is included in them, as by what is excluded, left unsaid, or missed out.(2) Of course the amount of time spent on writing in contexts of care is not only a concern for social workers.Experienced foster carers who work with the more complex, or challenging children ranging from unaccompanied asylum seeking children, to children mixed up in drugs and gangs through criminal exploitation, can feel that the 1.5 to 2 hours per night they have to spend filling out incident reports, is unnecessary and detracting from the day-to-day work they do as carers.The pressure comes in part from foster care regulations and pressure from their foster care agencies, to record everything in a trail of accountability "in case something happens", by which they mean 'goes wrong.'I'd been thinking about the ways in which the bureaucratisation of social work was being theorised and enjoyed coauthoring a paper drawing on my own reading and the WiSP study (Balkow and Lillis 2019).( 3) Further down the line, consideration could be given to how social workers develop and improve their writing skills, as part of their continuous professional development.
The majority of ongoing training that social workers undertake tends to focus on specific service user groups or issues concurrent to those groups.If training does focus on record keeping, it is often in response to updated forms or software packages and how these should be correctly completed.A question could be: Is there a useful reflective dimension to social work writing, other than the procedural and the statutory uses?Perhaps if we write about service users and our interactions with them, then we begin to learn about their lives, constructing a narrative, story or snapshot of their lived experiences.( 4) Michael Balkow's account illustrates how decisions to participate in a research project may be driven by existing professional and theoretical concerns, a key concern in this instance being the problematic gap between 'fieldwork' and 'office work', indexing the time spent with families in contrast to the time spent on writing (1).A focus on time was built into the WiSP research design (the interview schedule, the observation, logging) based on a review of concerns typically raised by the social workers, so the topic of time was pre-scripted and therefore to a certain extent already in alignment with a key professional concern.What could not be pre-scripted were the specific findings and understandings generated from the range of data sources which evidenced in specific ways the central problematics of time in social work practice (for a full discussion, see Lillis et al 2020).As such it is an example of what might be described as shared 'motivated looking' (Sarangi and Candlin 2001).The existence of some shared motivated looking is important for the very existence of the project -as illustrated in this account by echoing and validating existing professional concerns and thus participants' decisions to take partas well as for the potential usefulness of the project and continuing collaborations.Ethics of analysis and representation (decisions about what to focus on, why and how) are therefore bound up with the ethics of access and participation.Furthermore, future directions for research and analysis may be driven by calls by some participants for researchers to extend their analytic gaze ; in this instance, to extend the empirical focus on time beyond social workers to all involved in social care (3), and to offer theorisations which reconfigure writing as a valuable reflective rather than administrative tool in social care (4).

The ethics of dissemination (within the context of representation and ownership)
'Dissemination' tends to be treated as the end point in a research project's chronology yet this presumes that there is only one end point (often framed as such by funders) rather than multiple, and numerous iterations.Any one particular end pointmarked by a publication or a presentationis potentially also a starting point for further investigation of different aspects of the data generated.Furthermore, there are many stakeholders beyond the immediate project participantsother professionals and agencieswho may value the findings and insights presented but want to ask different questions of the data generated, arising from what they see as particularly pressing needs.The reflections by a development officer, currently leading a project aimed at transforming professional recording practices, raises questions about whether researchers (should or do) re-examine their priorities for analysis and dissemination, in light of specific professional and policy interests.

Reflections from Miriam Smith, a planning and development officer -an interested stakeholder
The Write Right About Me (WRAM) is a multi-agency records improvement project aimed at supporting a change in writing practice for children and young people supported by the care system.The project is approaching writing from a children's rights perspective, seeking to make improvement in relation to the ethics around 'voice' in writing and use of language.I came across the WiSP project in a basic Google search -'research about writing in social work'and it generated great excitement not only because of the ease with which I was able to find it but also because of its focus on the act of writing itself in a social work context.(1) The ethical basis for a child's voice in their record already exists and there is an increasing amount of literature about the absence of child's voice in their records (for example, Bruce 2014; O'Reilly and Dolan 2016) which adds weight to this as an area for improvement along with the outcome of the Scottish Independent Care Review ( https://www.carereview.scot/).However, even where research makes reference to the way a child's voice is written, it does not offer the findings in a way that provides constructive tools for the improvement of writing.(2) Although I was initially disappointed that the WiSP project's focus was from the perspective of social workers, rather than the perspectives of children and young people, its findings are key in understanding and creating a positive learning culture that can begin to support a change in writing practice that can benefit both social worker and child.The open access resources generated on the basis of the research, https://wisper.writinginsocialwork.com and the focus on writing and use of language itself transcends specific professional boundaries and is providing us with reflective tools to explore writing across professions including Health, Police, Education and agencies responsible for legal processes.(3)I also wonder whether the WiSP data can throw light on the child's voice on written records, even though I know that was not a key focus of the research .
The real improvement test for writing care records, from my perspective and the WRAM project though is feedback from the children and families who we write about-what helps them and what does not.Working with the PI of WiSP to shed light on this and to build constructive tools, based on those already developed, will be an invaluable resource for those who write about, for and with the children and families supported by the agencies that make up the care system.(4) Miriam Smith's account illustrates how 'dissemination' is not one final point in a research project's chronology but is rather best viewed as a series of specific opportunities for participation and engagement that go well beyond the immediate research context, which in turn feed back into decisions about analytic work.In the case of WiSP, the specific concerns raised by WRAM (1) have led to a WiSP researcher being involved in multidisciplinary work aimed at changing policy practice (3), as well as discussions about how the data and the insights generated from the WiSP project might engage with the specific questions raised by WRAM.In research projects setting out not only to explore, but to engage with, professional concerns, there is an ethical imperative to view the data generated as a resource for ongoing engagement with pressing professional concerns that go beyond the immediate parameters of the research project (4).Whether researchers (should or do) decide to re-examine their priorities for analysis and dissemination depends on how they view the larger community's rights to symbolic-ethical ownership (5.1).
The three brief researcher/participant-stakeholder reflections on their engagement with the WiSP project illustrate the ways in which ethics cannot be viewed as simply procedural or at one point in a research chronology, but rather as permeating the research trajectory.They also give some indication of what we describe as collaborative focusing that is, researchers and participant-stakeholders (implicitly and explicitly) working towards reaching some agreement about what should be the primary focus of analytic attention at any one moment in time.Some of this collaborative focusing maps on to what are understood as conventional academic research priorities: i.e. the importance of carrying out careful analysis to generate robust research findings and arguments warranted by data and (re)presented in peer-reviewed publications.However, some of this collaborative focusing in a professionally-oriented project necessarily leads to the setting of different priorities for researchers working with participant-stakeholders: examples in the WiSP project include building resources with and for practitioners, as signalled by Gillian Lucas (5) and Michael Balkow (4) setting priorities for specific analyses of data, such as focusing on the voice of the child in written texts, either as primary addressee in any writing (Gillian Lucas 5 ) or as being the primary perspective (Miriam Smith 1, 4).

Discussion
This paper set out to explore and reflect on the ethical dimensions in professionally-oriented research projects, focusing on WiSP as an illustrative case study, and using the categories of procedural ethics and ethics in practice to structure the discussion.
The paper illustrates the nature of the procedural ethics involved, documenting the codified requirements (Figure 1) as well as the processes involved in meeting the requirements of academic and agency governance bodies and the significant amount of invisible labour involved.The historical and immediate imperatives for fulfilling such requirements in order to protect participantsand in the case of social workers, the vulnerable children and adults about whom social workers are writingwere underlined.
Key differences between the procedural ethics as enacted by the academic institutions and the professional agencies are evident.Most obviously, whilst fulfilling the procedural requirements set out by governance bodies was essential in both cases, fulfilling the agency regulatory requirements was insufficient for proceeding with the research.In addition to 'organisational' level agreement, additional formal and informal permissions were required at 'operational' and 'individual' levels (Table 2): whilst the organisational and individual levels are routinely recognised in applied linguistics research processes, the operational level is not, and the importance of working to secure permission at all levels is usually backgrounded.Furthermore, decisions made in relation to procedural ethics by the participant-stakeholders at all levels were informed by: a) existing relationships and previous work with the researchers, b) the immediate material contextual conditions, e.g.workload, levels of stress, existing priorities.Procedural ethics therefore cannot be separated from ethics in practice, with ethics in practice permeating the complete research chronology (pre, during and post the official end period).In contrast to academic-facing bodies, where decisions made were fundamentally decontextualised, that is abstracted from the site of research (see Hammersley 2009 for a critique of university committees), decisions made by the agency-facing bodies were embedded in the contemporaneous contextual conditions in which the research was intended to take place.Ethical procedural judgements made by the agencies (organisational level) and the participants (operational and individual levels) were therefore fundamentally informed by contextual as well as procedural considerations.
In seeking to illustrate the key dimensions of ethics in practice, we have foregrounded the perspectives of three participant-stakeholders.Their accounts underscore the importance of ongoing dialogue between the researchers and the participant-stakeholders, in particular: a) exploring alignment between professional concerns and the researchers' stated interests in order to sustain a project (participation and involvement of any kind involves substantial investment by participants in terms of time, intellectual and emotional engagement with researchers); b) the researchers listening and deepening understanding of apparently procedural practices-notably in the case of WiSP these relate to text anonymisation practices and potential consequences (see discussion in Leedham et al , 2021); c) collaborative focusing in establishing priorities for action e.g. the foci of analysis and representation (for example, outputs may be both academia-facing.such as academic articles, and professionally-oriented, such as resource building and feeding into existing professional projects).
The importance of generating multiple types of useful/usable knowledge is of key concern in research projects exploring professional discourse studies (Candlin 2009;Sarangi and Candlin 2010;Lillis 2021b).This position signals the need to unsettle familiar categorisations about the nature of applied linguistics work being undertaken.For example, one characterisation of research in professional domains is to set the 'rigour' (of research) against the 'relevance' (of practice) (Schön 1983; see discussion in Sarangi 2019), as if these were necessarily at odds.Yet as is illustrated in 5 relevance and rigour are more likely to be contingent categories, negotiated at specific momentsbetween researchers, between researchers and participant-stakeholdersin projects such as WiSP which sit at the juncture of different kinds of expertise, interests and concerns for action.
The discussion of ethics in practice in the WiSP project problematizes what has become a conventional tripartite distinction in discussing ethical imperatives underpinning applied linguistics research: research on, with, for (for discussion see Mallinson 2018).In some sense, WiSP is a conventional 'research on' project: it is framed around specific research questions that a research team set out to explore.However, it would be more accurate to describe WiSP as involving aspects of on, with, for at different moments of the research chronology.In exploring the problematics of writing with managersthe initial impetus for the research (see 2)the researchers necessarily also explored issues around writing with social workers, seeking to explore their perspectives and practices.In the conceptualisation and the enactment of the project, some agencies, managers and social workers actively collaborated with the researchers in the direction of the research and the articulation of the problematics, whilst others were keen to participate on the basis of helping the researchers to generate useful findings.
As to who the research is for there are similarly multiple purposes: as a research project it seeks to contribute to academic fieldssuch as professional discourse studies and literacy studiesand is therefore for furthering academic knowledge.But it also seeks to be useful and usable to professional social work, in terms of both policy and pedagogy/professional development and therefore is working with different individuals, agencies and professional bodies to that end, as illustrated by WiSP collaboration with the WRAM project (see also https://www.iriss.org.uk/resources/multimedia-learning-materials/writing-analysis-socialcare).Professionally oriented applied linguistics research demands an ethics of practice that cuts through binary framings (i.e. is this research on or for?) and draws on insights from a range of (apparently opposing) notions, such as Hammersley and Traianou's (2014) emphasis on the production of 'sound knowledge', alongside alternative frameworks, such as Noddings' (2003) 'ethics of care', in this case showing care for participant-stakeholder concerns.

Concluding remarks
The importance of procedural ethics cannot be underestimated in research in professional domains.Such research typically involves the bringing together of outsiders ('intellectuals') and insiders ('professionals') (for discussions, see Shils 1968;Sarangi 2012), who seek to explore key communicative aspects to people's professional lives, often with the lives of 'others' often centrally implicated.In the case of the WiSP project, which seeks to explore the writing practices of social workers, the 'others' are some of the most vulnerable children and adults in society.The regulatory frameworks constituting procedural ethics offer substantial protection for their identities, as well as the identities of participating social workers.
However, in the contemporary context of higher education where procedural ethics occupies such a large institutional space, there is a danger that an overreliance by researchers on TABLES AND FIGURES but also concerns more fundamental questions about exactly what is being 'accessed' e.g.what participants do share about their perspectives.Ethics of participation refers to the processes and meanings of gaining informed consent for research participation as well as the particular 'participation frameworks' embedded within research methods, e.g., what kinds of participation a particular interview schedule or interview interaction enable.Ethics of interpretation/representation refers to decisions around which analytic tools are used when interpreting and representing data, e.g. which particular applied linguistic/discourse frameworks should be used for analysing written texts.Ethics of dissemination/intervention refers to decisions around what understandings to disseminate from the research, e.g. which specific areas researchers should publish about, in which order and on what basis.