Data reuse across international contexts? Reflections on new methods for International Qualitative Secondary Analysis

Abstract

This research note reports on five online workshops by an international team of scholars, the authors, with shared interests in drug (mis)use. The workshops comprise a novel form of collective international qualitative secondary analysis (iQSA) exploring the possibilities for, and value of, qualitative data reuse across international contexts. These preparatory workshops comprise the preliminary stages of a longer programme of methodological development of iQSA, and we used them to identify what challenges there may be for translating evidence across international contexts, what strategies might be best placed to support or facilitate analytical engagement in this direction, and if possible, what empirical value such exchange might have. We discuss how working across international contexts involved the authors in new 'translational' work to address the challenges of establishing and sharing meaning. Such ‘translation’ entailed a modest degree of empirical engagement, namely, the casing of empirical examples from our datasets that supported an articulation of our various research studies, a collective interrogation of how, why and which such cases could be used for best translational effect and a collective reflexive engagement with how these cases generated new and novel questions that in turn re-engaged us with our own data in new ways. Descriptions of our datasets, therefore, emerged as multifaceted assemblages of ‘expertise’ and comprised the evidential bases for new empirical insights, research questions and directions.

Introduction

Over the past two decades, there has been a growing corpus of innovation in qualitative secondary analysis (QSA) using small scale datasets in new and creative ways (e.g. Bishop, 2007; Davidson et al., 2018; Edwards et al., 2020; Hughes, et al., 2020; Irwin et al., 2012; Tarrant and Hughes, 2019; Tarrant, 2017). While this emerging scholarship has demonstrated the value and capability of QSA methods to inform on the social world in new ways, such value or utility has yet to be explored via international comparative work.

This research note reports on key methodological reflections engendered via a series of five online workshops interrogating the possibilities of, and for, international qualitative secondary analysis (iQSA) between the authors, all scholars working in the field of drug (mis)use in the Universities of Leeds (UK) and Aarhus (Denmark). These workshops, in part, comprise a mode of preparatory iQSA: namely, about the feasibility and potential empirical value of international data sharing, rather than involving data sharing. Accordingly, the workshops are built around methodological questions to establish the feasibility of data sharing as well as considering how and whether iQSA offered the potential to produce substantive insights in cross/international contexts. They are additionally preparatory as they were convened to provide an empirically driven funding application for research on iQSA. Based on methods of collective QSA (Bornat, et al., 2008; Tarrant and Hughes, 2020) each workshop was planned one at a time, and although we describe them as ‘linked’ workshops, we were not aware at the outset of what those links might be, how we might achieve them, whether the workshops would be productive or what they might produce. Therefore, they have played a key role in facilitating the ‘getting started’ on iQSA, and form part of a longer programme of work preparing for, and involving, international data sharing from thematically linked datasets ¹ .

In this note, we focus on our dialogue and reflection within and about the workshops, in which we identify a range of unexpected translation activities central to what we term collective iQSA. Here, the ‘translation’ across our international contexts refers to more than the linguistic translation of a word in one language into its cognate partner in another, but also of bringing into analytic alignment cognate theories, concepts, literatures, methodologies, fieldwork ‘sites’, study populations and so forth. In brief, as we discuss below, we identified the need to question the presumed shared character of empirically embedded meanings as they are used internationally (e.g. addiction, drug misuse, drug policy, methodology, context and so on), and consider the extent and form of ‘translation’ required to support properly shared understanding across our team. Furthermore, this ‘translation’ entailed a modest degree of empirical engagement, comprising a novel form of collective iQSA.

While ‘translation’ is involved in all cross/international qualitative comparative research (Chapple and Ziebland, 2018; Frank et al., 2021; Ritter et al., 2016), there is a difference between embarking on a cross national or comparative international study producing data together and the iQSA of existing datasets produced in different international contexts (see also, Hughes et al., 2021). In the former, ‘translation’ ideally begins as part of preparing the research proposal and continues throughout the data collection process (Haak et al., 2013; Martinus and Hedgcock, 2015). In relation to iQSA, however, we make the case below for considering how ‘translation’ may be integral to processes of ‘recontextualisation’ and ‘reconnection’ of existing data, as researchers repurpose them in new ways and in new contexts (Moore, 2007). Here, the question of how we might translate evidence across international contexts fundamentally requires us to consider how we variously apprehend data, whether and how such data can be brought into analytic alignment, and how they can ‘tell about society’ (e.g. Becker, 2007; see also Hughes, et al., 2020; Hughes et al., 2021). Thus, the analytic conversations comprising our workshops simultaneously involved questions concerning both the meanings we sought to establish within our data, as well as the meanings we held about our data, especially how they could be treated as what sorts of evidence across which international contexts. Attention to processes of ‘translation’ offers a means of analytical engagement with how we formulated new epistemic directions, as well as to how our collective iQSA leveraged a form of international analytic ‘remove’, whereby the international similarities and differences of the team members engendered new methodological questions and facilitated new empirical insights.

It is useful to note that these workshops comprise a mode of analytic engagement, an heuristic, rather than a pedagogy of techniques. While this analytical engagement is important in all cross national qualitative research, it manifests itself in particular ways in relation to iQSA. Our reflections are thus intended to provide a general ‘road map’ of the key features of our work for other researchers interested in undertaking iQSA, rather than a specific protocol. We conclude with suggestions concerning the wider applicability of our approach for international qualitative research collaboration.

Background to methods of QSA

Over the last two decades there have been large scale international investments in archives and repositories capturing a ‘tsunami’ of new data for the purposes of reuse (Bishop, 2009; Edwards et al., 2020). Technological advances have been prompted by concerns that qualitative data are expensive to produce (Holland, et al., 2006), important documents of human life (Neale, 2020) and an endlessly creative resource that connect us to the much longer social histories of which we are part.

However, qualitative data reuse is far from straightforward. Debates on appropriate methods for reusing qualitative research data suggest that the contextualised character of qualitative data production requires considerable additional theoretical work by researchers seeking variously to reuse them (Irwin et al., 2012; Mauthner et al., 1998; Moore, 2007). The theoretical complexity involved in qualitative data reuse has driven a growing body of work on the distinctive affordances and defining challenges of QSA (Bishop, 2007; Bornat et al., 2008; Davidson et al., 2018; Hughes et al., 2020, 2021; Irwin et al., 2012; Tarrant and Hughes, 2019). In addressing these questions, there is increasing diversification in the field of QSA not only in the sorts of questions or new purposes to which existing data are put, but also a reframing of notions of qualitative scale and data ‘corpus’. For example, new methods are developing in what have been termed ‘breadth and depth’ QSA involving the amalgamation of datasets, thus ‘scaling up’ what might otherwise be seen as modest qualitative sample sizes for the purposes of data reuse (see especially Davidson et al., 2018; Edwards et al., 2020).

The approach we report on here falls within a tranche of work we would describe as ‘depth to breadth’ QSA (Tarrant and Hughes, 2019), moving from work with small data samples to working across multiple datasets in a process of cumulative or ‘synthetic’ analyses (see Hughes et al., 2020). Here, ‘synthetic analyses’ refers to the synthesis of different types of data by bringing them into conversation with one another, as well as with evidence, findings, theory and developments in the social world beyond the original study contexts. ‘Depth to breadth’ methods of QSA, especially those involving empirically driven dialogic exchange (e.g. Bornat et al., 2008; Irwin and Winterton, 2011; Irwin, 2013; Tarrant and Hughes, 2020; Hughes et al., 2021) are particularly well placed for international collaborative enterprise newly beginning to explore questions of how we might conceive of, and work collectively with, data and findings from thematically linked qualitative research across international contexts.

Our iQSA workshops build out of this work and are illustrative of such collective dialogic exchange, albeit our work involved descriptions of datasets in the formulation of new research directions, rather than the sharing of what might be considered ‘proper’ research data. Here we make a distinction between data sharing and date reuse. Data ‘sharing’ refers to the sharing of data with researchers beyond the originating team either through the archiving of research data (Bishop, 2009), or direct sharing of partial data samples or whole datasets between research teams for the purposes of new research (see especially Bornat et al., 2008; also, Irwin et al., 2012; Thomson et al., 2012; Haynes and Jones, 2012). ‘Data reuse’, by comparison, refers simply to where researchers may revisit their own data over time (Neale, 2020), or reuse ‘found’ data (Goodwin and O’Connor, 2015) or data curated for reuse, as in archives such as the Timescapes Archive (https://timescapes-archive.leeds.ac.uk). However, methods for the sharing and reuse of qualitative research data across international contexts where archives may not yet support international sharing, and ethical concerns and protocols may differ, are as yet embryonic and our work here comprises a unique contribution to an emergent methodological field. We focus primarily on questions of data reuse and the language of ‘sharing’ in this note more frequently refers to shared analyses.

Methodology: collective iQSA workshops

Through several activities, the authors were made aware of a shared interest in each other’s work. Hughes (University of Leeds, UK) was already seeking international collaborators to explore the feasibility and value of international QSA in preparation for a funding application. Via a series of emails between Frank, Herold and Houborg (Aarhus University, Denmark), a shared commitment to this endeavour was established, and an international visit by Hughes to Aarhus organised. Via email and Skype, Hughes and Frank formulated agendas for various meetings and workshops for scholars in the Centre for Drug and Alcohol Research (CRF, Denmark), including Houborg (Aarhus University, Copenhagen division, Denmark). COVID-19 prevented the visit.

Instead, using Zoom, we convened a series of workshops in which we would be participants to use, and critically interrogate, methods of collective QSA in an international context using descriptions of five studies on drug use in which we had variously been involved. Collective QSA commonly involves bringing members of different research teams together to permit a dialogue of evidence and questioning, for example, through theoretically sampling from across datasets (e.g. Bornat, et al., 2008; Irwin, et al., 2012; Tarrant and Hughes, 2020). In collective QSA, researchers commonly explore how and whether datasets may be analytically linked, aligned, contrasted, perhaps for the purposes of empirical extension, but also as a process of familiarisation for researchers new to those datasets intending to reuse them. Alternatively, collective QSA might be developed to consider the possibilities for new analytic directions (Bornat et al., 2008; Tarrant and Hughes, 2020). Analytic ‘conversation’ is thus a conversation of evidence, theory and findings from beyond the immediate contexts of the original study (see also Hughes, et al., 2021). In work elsewhere, Hughes and colleagues have suggested that collective QSA harnesses relational epistemic and temporal ‘remove’, a defining concern of QSA, from the original contexts of data production. The language of ‘remove’ entails a shift from an over-concern with binary distinctions such as ‘primary’ and ‘secondary’ analysis, analysts or data and instead engages us in careful and precise examination and articulation of researchers’ relationships with the data with which they are engaged (see also Moore, 2007; Irwin, 2013; Hughes et al., 2020; Hughes et al., 2021).

While questions of the feasibility and analytical affordances of iQSA were central from the outset of this programme of work, we lacked any shared understanding of methods of collective QSA in any general sense, the sorts of resource or intellectual commitments it might involve, what sorts of data we might share or questions we might pursue. Our proposed work was also novel in that scholarship concerning strategies and approaches required for working across international contexts, the feasibility of international data sharing where the data have not been archived, and where approaches to the archiving of international qualitative research data for the purposes of data sharing, is as yet embryonic.

To understand how we might establish questions of data relevance across international contexts and with which data, using email and Zoom, we therefore developed preliminary questions to explore through workshop conversation:

How far is it possible to bring together thematically linked data from different international contexts?

What forms of ‘translation’ need to take place to share the data or to make it meaningful to the different team members?

What new questions might each participant bring to the various datasets (their own and others), and how far might data sharing support the interrogation of these questions?

How might we share data ethically between international teams, and what are the international differences in legislative and ethical codes for data management and sharing?

Workshop descriptions

There was a cumulative character to our workshop conversations, whereby emerging questions, preoccupations and substantive themes in one workshop were taken forward in the next. Workshop 1 involved sharing methodological understandings of what is QSA and how it might be done, reviewing examples of QSA and identifying what could be included as a relevant dataset. This workshop established the appetite amongst the authors for undertaking QSA. Workshop 2 involved Frank and Herold describing the contexts of their various studies, their study aims, the key characteristics of each study sample, the key stakeholders, the types of qualitative interviews conducted, how much data still exists and in what format (e.g. paper and digital) and the broader fieldwork undertaken. Based on these descriptions, all four authors discussed what sorts of questions the data might support, and the workshop concluded with agenda setting for the following workshop. Workshop 3 involved Houborg and Hughes presenting on their studies, a discussion of what connections, if any, these studies had to those in Workshop 2, and the thematic continuities and connections that emerged between the studies so far discussed. The conversations in Workshops 2 and 3 generated new empirical insights and new intellectual directions for iQSA, briefly outlined below. Workshop 4 explored the ethical concerns and next steps for sharing data across international contexts – where data sharing is not simply a case of sharing data – and how we might usefully report on these. It was in Workshop 4 that we formulated the idea for this Research Note as a ‘roadmap’ for other researchers working in this area. Workshop 5, discussing the ‘findings’ section of this Research Note, involved reflection on our previous workshop conversations, establishing agreements on attribution in publications, discussing and agreeing approaches to co-writing, making and receiving comments and the character of dialogic exchange so far. We additionally considered what are treated as data – for example, our conversations, the descriptions of data and the datasets – and definitional language around methodology. A reflective log from each workshop was circulated for additional comments, enabling us to track conversations, as well as specific contributions and suggestions. As the workshops progressed, we identified resources that would support our discussions and set up a Sharepoint folder accessible for all team members. Besides storing the reflective log, we also stored our reflections on ethical concerns (including how to anonymise, apply general data protection regulation and informed consent rules, etc.) important in the next steps of preparing and data sharing internationally. We also used Sharepoint to formulate and store a data/intellectual property agreement about strategies for team members’ attribution of ideas and findings from our discussions in forthcoming outputs ² .

These workshops therefore entailed a close blend of epistemic, methodological and pragmatic concerns that developed as we sought to bring our various understandings, expertise and experience in questions concerning data reuse, alongside our substantive and theoretical expertise, into ‘conversation’. However, as we endeavoured to establish the ‘contextual embeddedness’ of our datasets (Irwin and Winterton, 2011), in terms of the broader contexts (academic discipline, policy) in which they were formulated, the funding schemes and purposes for which they had been produced, and researchers’ specific epistemic questions and interests during the formative contexts of the studies, we encountered challenges for establishing and sharing meaning.

‘Translational’ work in collective iQSA

From the outset, we encountered points of surprise – moments when explanation unexpectedly became necessary – where we were variously required to engage in different sorts of explanation, more usefully described as ‘translation’ work, to support discussions of what, to each of us, had been self-evident (see also Simovska et al., 2019) ³ . These points of surprise were critical to questions pertaining to the value and feasibility of iQSA and the formulation of new research directions. For example, essential for our developing conversations was elucidation of: the policy histories shaping the formulation of each study, alongside policy-driven treatment imperatives and resources; of the geographical location of each study and the socio-economic histories of these places; and of the research histories shaping each study, their funding contexts and study aims.

Understanding these various histories was indispensable for developing a shared conceptual language in ways that moved beyond straightforwardly drawing on specific theoretical frameworks. In illustration, a long history of policy development in Denmark has meant that the language of addiction, so commonly used in the United Kingdom, has been replaced by that of substance/drug dependency or (mis)use. Were we, then, comparing our studies based on what the participants did, their practices, (e.g. drug use) or how they discussed their drug use (discourses, narrative conventions and policy language)? How far is it possible to separate discourse and practice (Hughes, 2007)? Indeed, was it possible to bring concepts to bear on data which had been deliberately formulated in opposition to such language? Consequently, how far could our iQSA be comparative, or comprise a shared epistemic endeavour, with such differences at play and, if we laid claims to the latter, to what degree was this possible when our conceptual framing differed so markedly – not solely because of differing policy framings, but through longer, disparate, intellectual histories of engagement? ⁴

We found the language of ‘translation’ useful to describe our efforts towards clarification and explanation because it includes both linguistic translation as well as the ‘translation’ of cognate theories, concepts, literatures, methodologies, fieldwork ‘sites’, study populations and so forth, through explanation and critical interrogation (see also, Irwin et al., 2012). Translation also operated as a form of analytic ‘remove’, namely, international remove, through which we identified the analytical affordances (what becomes possible) and deficits (the limitations of possibility) of bringing differently constituted research data into analytic alignment across international contexts, rather than relying on the ‘naïve’ treatment of research data as ‘data’ (Silverman, 2017). Such ‘remove’ was critical for understanding how we might analytically apprehend those datasets – how we recontextualised them for the purposes of our collective analyses – and how to recast them as particular orders of data – how we rendered them able to speak to the new questions we sought to bring to them (Hughes et al., 2021). In this way, these points of surprise were critical to clarifying the forms of methodological engagement involved in iQSA and the character of ‘evidence’ necessary for translating across international contexts. Such translation, however, was not achievable solely through direct explanation. As our workshops progressed, it became clear that clarification, explanation and so forth relied on the strategic use of empirical examples which, through their very complexity, allowed us to address complex nexuses of meanings, practice and so on in the studies we sought to describe.

Strategies for discussion: casing from the data

It was at this juncture that the shared use of interview-based methods across the five studies under discussion became critically important for our discussions concerning the possibility of iQSA. To facilitate comparison across the datasets, we used either our own experiences as researchers about particular interviewees, or interviewee’s accounts about themselves, as ‘emblematic cases’ (Thomson, et al., 2012) for the purposes of explaining or illustrating particular aspects of our research or conceptual thinking. Such emblematic cases took the form of extensive vignettes (summaries of participants’ life situations) and were an especially productive form of epistemic reflexivity. They helped to render the unfamiliar familiar, serving to support translation across contexts, and to bring the epistemic ‘silences’ in our own thinking, or of the formative team, into analytic ‘view’ ⁵ . Further, detailing complex situations through these examples helped us collectively to engage with complexity in our thinking. While we were confronted with significant differences of understanding across international contexts, the repeated empirical embedding of our discussion helped us avoid overly abstract conceptualisation and identify potentially fruitful connections across the five studies. In this way, ‘translation’ entailed a modest degree of empirical engagement, namely, the casing of empirical examples from our datasets that supported an articulation of our various research studies, a collective interrogation of how, why and which such cases could be used for best translational effect and a collective reflexive engagement with how these cases generated new and novel questions that in turn re-engaged us with our own data in new ways. Descriptions of our datasets, therefore, emerged as multifaceted assemblages of ‘expertise’ and comprised the evidential bases for new empirical insights. Thus, there was a blurring between what we had considered ‘proper’ data – for example, fieldwork data, and the broader ‘synthesis’ of expertise, scholarship, knowledge and experience on which we drew in the formulation of our collective iQSA.

In the third workshop, through critical attention to our methodological language, we acknowledged how the ‘first-person’ point of view in interview data was integral to such empirical embedding, enabling us to draw on participants’ own explanations and accounts of their lives as they related to drug use. In other words, similarity of research methods in the original studies supported a modest degree of cross-study comparison. Further, we found that protracted and immersive fieldwork experience enabled us to use extensive and detailed knowledge of participants’ personal circumstances and elaborate our empirical examples or illustrations. However, this was only the case for four of the studies as a fifth study involved access to inmates of secure penal institutions, where possibilities for familiarisation with participants prior to interview was nonexistent, and there was a lack of gatekeeper engagement – often a rich source of contextual insight in qualitative fieldwork. This fifth case helped clarify the extensive use to which we put our experiential knowledge in translating our studies (see also Mauthner and Parry, 2010) and served to illustrate how proximate knowledge of study contexts might be critical in processes of ‘getting to grips with’ the embedded contexts of datasets (Irwin, 2013). In other words, this helped identify potential limitations for iQSA. It is important to distinguish this idea of proximate knowledge from ‘first-hand’ or ‘co-presence’, however. Hughes, for example, was not the fieldworker in the study she described, but had had access to extensive fieldnotes for each interview and still retained notes and email exchanges between herself and the interviewer about specific interviews and participants on which she drew in workshop discussions. Thus here, as elsewhere (Hughes et al., 2021), ‘degrees’ of proximity and remove varied for each author in relation to their different studies, as well as between the authors.

iQSA: the potential for new insights?

Throughout the workshops, we identified thematic continuities across all five datasets and were additionally surprised at how the simple process of describing datasets supported the development of new research questions, as well as new insights. Workshop conversations and the questions others brought to the various datasets supported their recontextualisation, producing new ways of (re)apprehending the data, as well as original findings and outputs from each study. Through this form of dialogic synthesis, new research questions concerning heroin users were developed. Briefly, these can be summarised as: the possibilities for sustaining intimacy through different drug-taking phases, friendships, family networks and circumstances; the ‘trade-offs’ involved in sustaining or moving away from a drug taking life; and how the temporalities shaped through policy contexts, services and rhetoric constrain or enable drug-users in formulating new ways of living. The empirically situated dialogic synthesis comprising our iQSA therefore moved beyond the preparatory work establishing the value and feasibility of data reuse across international contexts, to support the development of new research questions, and new empirical insights. Furthermore, these conversations worked to identify the pragmatic, ethical and legal challenges of data sharing across international contexts. These two areas, substantive developments and new questions concerning th possibilities or otherwise of international data sharing, form the basis of two forthcoming outputs from the authors.

Conclusion

We describe our research endeavours sustained over a series of five workshops as entailing a form of collective iQSA. Our workshops resembled, and indeed built out of, existing methods of collective QSA in that they involved formulating new research relationships, working across disciplinary boundaries and at an often significant (up to 20 years) remove from the formative contexts of the various studies. However, working across international contexts introduced novel forms of epistemic remove requiring new translational work to develop the bases for international comparison and shared analysis. Such ‘translation’ comprised a modest form of empirical engagement, in the recasting of the datasets, the production of new narrative cases and the analytic production of new and novel empirical insights, research questions and directions. The use of first-person point of view generated via interview-based methods across the five studies was especially important in facilitating dialogic synthesis across different forms of evidence (Hughes et al., 2020) ⁶ . Significantly, workshop conversation and the questions we variously brought to the different datasets produced new ways of (re)apprehending the data, findings and outputs from each of the studies. The conversations constituted a form of synthetic engagement across evidence, findings, theory, as well as developments in the social world (Hughes et al., 2020), as well as a form of expert evidence, suggesting a consideration of what constitutes ‘data’ in QSA more generally.

Frank and Herold have considerable experience of comparative international research and suggest that the work undertaken in our iQSA has application in these more general contexts (Frank et al., 2021; Giertsen et al., 2015; Houborg et al., 2016; Thom et al., 2013). Often, when undertaking comparative international research, the tendency is towards reducing complexity to achieve pragmatic bases for producing and comparing datasets. On this basis, researchers also call for enhanced methodological reflections in cross national qualitative research (Haak et al., 2013). This is certainly the case in research on drug dependency where the layered and multi-directional character of drug policy presents significant challenges for bringing policy frameworks into direct comparison (Frank et al. 2021; Ritter et al., 2016). Foregrounding the pragmatics of comparison in collective workshops such as those described here, rather than assuming the complexities can be erased at subsequent stages of fieldwork, might be a productive means for developing the bases, as well as strategies, for international comparative research. Thus, the ‘depth to breadth’ approach we outline may support scholars in developing depth explanations and solutions relating to human behaviour in complex processes and provide a new space for cross/international interdisciplinary research that supports cross-disciplinary dialogue through epistemic questions concerning the (re)contextualisation of data.

Finally, we conclude by suggesting these workshops demonstrate the value of unfunded research (Edwards, 2020) especially for the essential work involved in the development of grant applications, especially for highly experimental and exploratory research, where outcomes are uncertain. This has been the case for our iQSA development, which has been necessary to establish methodological feasibility, ascertain empirical value and identify any challenges – pragmatic or ethical – to be overcome.