Introduction

Public health evaluation research is a multidisciplinary endeavour that seeks to inform decision-making for the promotion and protection of health at the population level. The aim of public health evaluation research is to build an evidence base of ‘what works’ to inform practitioner and policymaker decisions (at local and national levels) about what programmes and policies to implement to improve the health of the population. The kinds of interventions evaluated for their public health impact range from individual behaviour-change programmes, such as to promote exercise or healthy eating, to interventions targeting more social or structural determinants of health, such as housing or education. Relating particularly to the latter, recent developments in public health evaluation research in both academic and practice settings have seen increased focus on how to evaluate ‘complex interventions’. These are typically defined as interventions with multiple, intersecting components which interact dynamically with the context in which they are delivered (Medical Research Council 2008). Complex interventions are thought to bring about change to health through multiple, complicated pathways and as such, their effects cannot easily be assessed through the experimental methodologies traditionally associated with (biomedical) evaluation research, such as the randomised controlled trial. Furthermore, the rise of attention on ‘complex interventions’ has also seen a shift in the expressed aims of evaluation. No longer is it considered useful to say simply whether or not an intervention works to improve health outcomes. Increasingly, evaluators seek to understand how, for whom and under what circumstances changes to health occur through an intervention (Bonell et al. 2012). This kind of knowledge is framed as valuable for informing how an intervention might be successfully delivered elsewhere or ‘scaled up’ to be delivered at a population level. As a result, different methodological approaches are increasingly being brought into the design of public health evaluation studies, including qualitative methods that can attend to the processes and dynamics of an intervention in context, such as ethnography (Cohn et al. 2013).

In this chapter, I seek to explore how ethnography can be used as part of public health evaluation research to explore dynamics of complex interventions and how changes to health might occur and be assessed. I focus particularly on the concept of ‘scale’ as a dimension of understanding how complex interventions bring about change, and I describe the way in which ethnographic research enabled me to explore dynamics of scale in an unfolding community empowerment intervention.

Conceptualising Scale in Evaluation of Complex Interventions

The concept of ‘scale’, I suggest, is embedded within the contemporary literature on the evaluation of complex interventions for health, but is rarely explicitly interrogated or explored. Complex interventions are frequently framed in evaluation research using ‘systems’ perspectives, influenced by theories of complex adaptive systems from various disciplines including biology and computing (Hummelbrunner 2011). A typical systems perspective used in evaluation research sees interventions as unfolding within, and interacting dynamically with, a broader contextual system. This system is thought to comprise complex, dynamic structures and relationships between its constituent (social, political, economic, environmental, physical) elements (Hummelbrunner 2011). When a complex intervention, such as a community empowerment initiative, is delivered, it will prompt multiple mechanisms of change which occur at different ‘levels’ or scales of the system (Mowles 2014). This is in contrast to, for example, the simple delivery of a drug for which the mechanisms of change are thought to be confined to the individual body. Therefore, the framing of evaluating complex interventions rests on an assumption that change occurs through interactions happening across different scales of the system in which the intervention is delivered. Not only will changes to health occur because of actions taken by a few individuals, such as a programme’s implementers or its participants. Instead, due to the multiple components of the complex intervention interacting with the context in which it is delivered, rippling interactions will occur across the system as a whole (Diez Roux 2011), both closer to and more distant from the site(s) in which the intervention is actually delivered. The purpose of evaluation from this perspective is to try to identify these mechanisms of change, to provide information or evidence on how an intervention influences health and to help inform decision-making by policymakers as to whether (and how) to implement an intervention in another context.

The concept of scale, then, also relates to this purpose of public health evaluation: to produce evidence of an intervention’s effectiveness that can be transferred or ‘scaled-up’ to inform decision-making to improve health at a population level (Hanson et al. 2003). This denotes an assumption that there is a comparability of scales between different settings or contexts which enables the meaningful transference of knowledge—as ‘evidence’—between them (Cartwright 2013). A further conceptualisation of scale inherent in public health evaluation research, and arguably at the heart of public health as a field, concerns the (assumed) predictability of the scalar relationship between the individual—the level at which health outcomes are typically measured—and the population—the level at which public health decisions are made. This relationship is typically assumed to be linear and stable, such that knowledge produced at the individual level about how to improve health can be extrapolated, or scaled up, unproblematically as ‘evidence’ to the population level (Krieger 2012). Critical consideration of the different dynamics of scale at play in the delivery and evaluation of complex interventions is required for public health evaluation research to contribute productively to the improvement of population health.

Conceptualising Scale Within an Ethnographic Approach

In this chapter, I draw primarily on perspectives from anthropology in my consideration of what ethnography might offer to public health evaluation research seeking to inform policy and practise decision-making to improve population health. My interpretation of ethnography reflects its anthropological roots in traditional ‘fieldwork’, whereby the understanding of relationships, structures and processes within (and beyond) the field of interest arises through ‘being there’ (Lewis and Russell 2011), via the researcher’s engagement with and embeddedness within these relationships (Hannerz 2003). From this perspective, ethnographic knowledge can be taken as ‘consisting of, and steeped in, social relations’ (Hastrup 2005: 141), and through the researcher-as-instrument, it enables (critical) interpretation of being and knowing in the world (Ingold 2008). In this chapter, I seek to demonstrate that these forms of positioning and interpretation unique to ethnographic research can help shed light on dimensions of scale inherent in the delivery of complex interventions, and which must therefore be acknowledged and explored through public health evaluation research.

Critical reflections on scale have emerged in applications of and debates around ethnography. Contemporary anthropological topics of interest, such as globalisation and mobilities of people, ideas and resources, have prompted reflections on how ethnography can be applied across territories to examine flows of relations (Marcus 1995; Gupta and Ferguson 1997). This contributes to disruptions of any assumed fixed scalar hierarchy between the ‘local’ and the ‘global’ (Massey 2004; Hastrup 2013), and of how the ‘field’ of inquiry might be defined in scalar terms relative to the dynamic systems in which it is situated. There have also long been debates on the process of producing ethnographic knowledge, and the inevitable scalar relationship between ‘anecdotes’ or observations from fieldwork (detailed, particular) and the interpretation of a broader social system conveyed through the ethnographic narrative (abstracted, often comparative) (Strathern 2004; Candea 2007). Hence, contemporary approaches to ethnography promote reflexive attention on relations of scale, in terms of both how the ‘object’ and field of study can be understood, and how knowledge on that object is produced and represented. As such, these perspectives hold much potential for exploring and considering scale within public health evaluation, and the role ethnography might play in interpreting how complex interventions unfold, to inform decisions on improving population health.

Ethnographic Explorations of the Enactment of ‘Community’

Overview of Ethnographic Research

To illustrate these arguments further, I draw on ethnographic research conducted around the delivery of a complex, community empowerment intervention in the UK, the Big Local (BL). The research was undertaken alongside a public health evaluation of the intervention, and here I describe ethnographic interpretations which illustrate some of the dynamics of scales and scalar relations occurring through the delivery of the intervention. The BL is a long-term empowerment intervention which was developed and funded by a third-sector organisation, BIG Lottery. Its design reflects recent policy movements in the UK (and elsewhere) to address spatially patterned determinants of health and social inequalities through local, area-based interventions (Bridgen 2004). The BL also reflects increasing shifts in public health and social policy towards engaging communities to empower citizens to identify and address local factors influencing their health and well-being (Popay et al. 2007; Reynolds et al. 2015). The BL began in 2010 and is underway (at the time of writing) in 150 areas across England, with populations ranging from 5000 to 12000 residents. Residents are given control over identifying and addressing priorities for improving their local areas to make ‘a massive difference to their communities’ (Local Trust 2015: 5). Residents are facilitated to form a committee, then conduct consultation with the wider community to develop and deliver a plan to address local priorities using allocated funding (one million pounds per area) over a period of 10 or more years.

Though not an explicitly health-oriented intervention, the BL has been identified as holding potential to improve the health of participating communities and to reduce health inequalities within and between populations. This reflects theorised pathways between increased collective control within participating communities via the empowerment mechanisms of the BL, and improved health outcomes (Whitehead et al. 2016). An evaluation study has been developed by a team from multiple universities to identify the mechanisms of the intervention which might lead to improved health and reduced inequalities. This evaluation study frames the BL as a complex intervention in a dynamic system (Orton et al. 2017).

My ethnographic research focused on how ‘community’ was enacted through the delivery of the intervention, alongside the first phase of the evaluation study (2014–2015). I conducted fieldwork in two of the participating BL areas and I spent 13 months in and between the two areas, following the tempo of activities unfolding in each area (Reynolds 2017). I observed a range of activities occurring as the BL was delivered in each area including: regular meetings of the residents’ committees leading the planning and delivery of the intervention; events organised by the committees such as consultation activities; and the roll-out of projects funded through the intervention, such as a sports programme for young people. I also engaged in informal conversations and interactions and more formal in-depth interviews with local residents, workers and other ‘stakeholders’ closely, and less closely, involved with the intervention.

‘Craybourne’ (all places and names are pseudonyms) is an area of 6000 residents in a coastal town, formerly a popular holiday destination, but now facing relatively high levels of deprivation, unemployment and health inequalities. At the time of fieldwork, the residents’ committee in Craybourne was beginning consultation with the wider community to identify priorities for the local area and develop a plan to address these. The second area, ‘Westin Hill’, is an urban site on the outskirts of a large city with around 11000 residents, with high levels of socio-economic inequality. The committee in Westin Hill had already conducted a year of community consultation prior to my arrival and were beginning to implement their plan to address issues of promoting intergenerational relations and improving community spaces and the local economy.

Scalar Relations Emerging Through the Enactment of Community

What did and did not enter my ethnographic ‘gaze’ while exploring enactments of community around the BL intervention is a first example of the different dimensions of scale at play in this context. My gaze was particularly small-scale, given the possible scope of the BL. I selected only two out of 150 participating areas in which to conduct fieldwork, my fieldwork lasted for only around one-tenth of the expected time span of the intervention, and I considered enactments of community occurring only within my two areas, rather than at the national level at which the BL was administered. Following the premise of ‘being there’, with my interactions constituting the mechanism through which ethnographic knowledge was generated, I was necessarily restricted in my ethnographic engagements. My interactions typically involved only a small minority of the residents in each area (rarely more than the 20–30 people actively involved in delivering the BL at any one time) and a limited number of spaces (typically community centres, residential homes and leisure spaces). Yet, through my ethnographic perspective, I was able to slide between different explanatory scales. The individual interactions and activities I identified through my presence were ‘not ‘small-scale’ at all in terms of the…insights they afford[ed]’ (Strathern 2000: 66) for interpreting how community is enacted through a complex intervention and the implications of this for public health evaluation.

Furthermore, my ethnographic research enabled me to identify a diverse range of ways in which scale and scalar dynamics were being produced, in often unpredictable ways, through the enactment of community as the BL unfolded. This included practices of representation undertaken by the residents’ committees acting on behalf of the wider community; the ongoing work of asserting and negotiating boundaries around eligibility to contribute to or benefit from the intervention; and the shifting positioning of the individual relative to the collective, through the delivery of the BL. I will describe ideas of scale in relation to each of these and highlight their implications for considering the role of scale in evaluations of complex interventions for health.

The Scalar Dynamics of Representation

The practice or representation, whereby a small group of people speak and act for or on behalf of a larger group, reflects a ‘folding together’ (Hastrup 2013) of different scales. This could be seen in much of the work of delivering the BL as a community intervention in both areas. Representation was performed by the small number of committee members in each area who consulted residents on local priorities for improvement, synthesised and reformulated these opinions into a plan and then took steps to put this plan into action. These representative practices reflect negotiations of the scalar disparity between the small numbers of active residents and the broader community for, and as whom, they were acting. The tensions of this process were captured in particular in my ethnographic observations in Craybourne, as the residents’ committee commenced the phase of community consultation.

The work of planning and doing consultation—of identifying groups of people, spaces, events and mechanisms to elicit residents’ opinions on the local area—involved a continuous interplay between scales. A range of discrete and targeted consultation activities was carried out that constructed the community as segmented and localised, but also as a broader, holistic ‘picture’ that was being painted through the collation of different views. There was constant movement between the community as a ‘whole’ and as a series of socially and spatially demarcated segments or groups, performed through the different types of consultation activities undertaken in Craybourne. Shortly after I began my fieldwork in Craybourne, the residents’ committee appointed a part-time ‘community worker’, Katy, to manage the consultation process. Katy identified her role and workload to me in terms of the need to ‘get round’ as many groups in the community as possible. She arranged weekly drop-in sessions at a couple of centres in Craybourne, as a mode of connecting directly with different types of people, and would feedback on her progress to the rest of the committee at meetings, illustrated in my notes from one such meeting:

Katy said she’d been holding her drop-in sessions and although they’d been slow to start with they’d picked up and she now has a queue of people… She also talked about a few groups of people who wanted to talk to her, but in a group setting and in a ‘safe’ place. She said there’s a rehab group who meet at Craybourne Action who’d like to have a session with her, and there’s a group of older women who meet at the Aroma café who would like to meet with her, but in an environment they feel comfortable in…. (Craybourne, observation, July 2014)

My ethnographic notes highlight the processes of segmentation of the community occurring through Katy’s accounts of her consultation work, into different groups and spaces defined by particular characteristics and needs.

However, later on, the limitations of this approach to ‘covering’ the whole community were identified, as Katy expressed frustration that her sessions were not ‘bringing in enough people’ and were too ‘ad hoc’. This exemplified a disruption to the assumed scalar flow between the series of small-scale engagements with discrete groups and spaces, and mapping the community as a ‘whole’ through the consultation process, thus indicating the unpredictable scalar dynamics of doing representation.

My ethnographic work highlighted a further example of the dynamic relation between community as a whole and a series of discrete segments which occurred towards the end of my time in Craybourne, when the committee were coming to the end of their year’s worth of consultation activities. I observed efforts in committee meetings to gather together all the information they had produced through the various activities to feed into the development of the plan to guide the delivery of the next stage of the BL. These conversations involved attempts to identify where different pieces of information were held and the format they were in, including data on various personal computers and copies of completed questionnaires stored in homes and local workplaces. This process of locating and amalgamating the fragmented outputs of consultation indicated how representation, as a core mechanism of the unfolding BL intervention, engaged, traversed and sometimes disrupted flows between the partial and the whole, as different scales at which the community was enacted. This suggests that evaluations must attend closely to how specific mechanisms of interventions shape and produce the scales at which interactions occur, which may disrupt expected notions of how groups of people, spaces and structures are positioned relative to one another and the wider system in which they are situated.

Shifting and Negotiated Boundaries

Another dimension of the enactment of community that I identified through my ethnographic fieldwork was the processes of asserting and negotiating boundaries around who and what was eligible to contribute to and benefit from the BL intervention in each area (Reynolds 2018). The shifting of boundaries (spatial, social, conceptual) indicated an ongoing fluidity of the scale(s) at which the intervention was being delivered and at which its impacts might be experienced. An example of this can be seen in Westin Hill, where my fieldwork began as the residents’ committee were preparing to deliver activities to meet the priorities specified in the plan they had developed, following a period of consultation. Yet, while this plan had been agreed, the steps needed to begin delivering against priorities of improving the local economy, community spaces and intergenerational relations appeared unclear to the committee. In the first few months of my fieldwork, I observed the committee develop a pilot process for commissioning projects to help meet the priorities. They decided to focus first on improving intergenerational relations and, within that, invited proposals for projects to improve resources for young people in Westin Hill. As such, this pilot process served to assert a boundary around the community that was very focused in its scale, funding a few activities that targeted only a specific group of the population.

Yet, even within this small-scale enactment of the community intervention, other negotiations of boundaries were performed, indicating the dynamic flow of the mechanisms of the BL across and between different sites, groups and levels of activity. One of the first projects chosen to be funded by the Westin Hill residents’ committee was a weekly youth sports programme located at the communal sports pitch on the Palmer Grove housing estate in Westin Hill. In some ways, situating the sports programme at Palmer Grove appeared to assert a rather limited spatial boundary, wherein the community focus of the BL was enacted only within the confines of the sports pitch and through the intended beneficiaries of the programme (young people). Yet, the activities unfolding here also carried connections that extended beyond the scalar limitations of this small space. The programme itself had been run successfully before at Palmer Grove and elsewhere in the wider borough over the previous couple of years, and a range of flyers and posters arranged on the registration desk at the entrance to the sports pitch publicised a number of similar upcoming events beyond Westin Hill.

I also identified dimensions of scale playing out through the registration procedure for the sports programme which involved giving an address and postcode to identify whether participants were residents of Westin Hill or not. This had been a point of contention and negotiation during the earlier residents’ committee meetings where members of the committee were concerned that programmes they chose to fund as part of the BL should be for the benefit of residents of Westin Hill only. However, the sports programme organisers had pointed out the challenges of managing this and their reluctance to turn away anybody who was not a resident. Alex, the organiser of the sports programme, conveyed to me during an interview some uncertainty around managing the ‘official’ Westin Hill boundary:

… and they can maybe dot, maybe have a map for a project, so when people come, can you put a dot or a pin on your road or something? … Just so we can see and we can build a picture of the [sports] project within Westin Hill… Where everyone’s come from… And did we get anybody outside? Because I know there was like priority is Westin Hill but I don’t want to turn away anybody, you know, that sort of … and then that’s ways for people to kind of say, look [the programme’s] here from Westin Hill, look this is what we’re doing to encourage building of knowledge … Because we’ve, obviously we’ve got the register and we can kind of allocate the roads and the postcode to where they are. Because when we asked them, are you Westin Hill? they didn’t know.

This example illustrates how the asserting and negotiation of boundaries through the delivery of the BL performed the intervention, and the community, at multiple scales. Asserting a boundary through the narrow social and spatial focus of the sports programme was a mechanism through which the residents’ committee felt they could take steps to begin delivering against their plan for the whole community. Yet, the enactment of this programme revealed connections with broader spatial contexts, and the negotiation of eligibility to participate revealed the porous nature of boundaries that undermine their apparent fixed scale. This ongoing boundary work (Reynolds 2018), identified through my ethnographic interpretations, indicates that the levels at which an intervention’s mechanisms of effect occur are not neatly bounded, but porous and negotiated, and that the scale of the target population ‘receiving’ the intervention is not fixed, but constantly in production.

Disrupting Aggregation from the Individual to the Collective

A third example of the kinds of scalar dynamics at play in the enactment of community through the BL was the positioning of the individual relative to the collective. My ethnographic observations and reflections identified the shifting positions of individuals relative to the collective, challenging assumptions from a public health evaluation perspective that community reflects a neat scalar hierarchy of individuals aggregated into a collective. During my fieldwork, I found myself following in detail the shifting positions and pathways taken by several individual residents that diverged from and intersected with the collective activity and spaces of the delivery of the BL. This indicated the limitations of interpreting community as an entity of greater scalar magnitude than the individual for understanding the plurality of ways in which the mechanisms of the intervention unfolded and prompted interactions.

An example of one such trajectory involved Magda, a long-time local resident and very active member of the residents’ committee in Craybourne, and someone I came to spend quite a lot of time with during my fieldwork. During my time in Craybourne, Magda became involved in ‘Star People’, a social enterprise scheme offered alongside the BL to participating areas, to enable individual residents to develop social entrepreneurship ideas. Shortly after I first met Magda, she began to tell me about her desire to find a new source of income and to get back into catering, something she had trained in some years before. As I got to know her better, chatting before and after committee meetings and at other events, Magda told me about her plan to apply for a small fund from the scheme, to develop a lunch club that she would host in local venues around Craybourne.

My presence in Craybourne and developing connection with Magda enabled me to follow her progress with applying for—and receiving—the funding and setting up her first lunch club, and to identify this as a trajectory that at times distanced Magda from her continuing involvement with the (collective) work of delivering the BL. Much of the preparation work for Magda’s new venture was located in her private home (in the kitchen), and her interactions with other people and spaces to facilitate setting up the venture tended to involve those unconnected with the BL, for example the local council environmental health officer. Furthermore, Magda’s discursive construction and narration of her trajectory to me served to highlight its distinction from the collective nature of the main intervention, despite the Star People funding only being available to residents of BL areas. Our conversations about her venture tended to occur on the spatial and temporal ‘fringes’ of the ongoing BL activity, for example as we were leaving committee meeting venues, or at a local café when Magda and I had arranged to meet socially. Following a rather difficult few meetings of the BL residents’ committee a few months into my fieldwork, Magda emphasised firmly to me that the lunch club was, in her mind, ‘completely separate’ from the BL.

Yet, despite this conceptual and spatial separation of Magda’s venture, and of her narration of it to me (as someone ostensibly not part of the BL or the community it represented), her trajectory intersected the collective nature and practices of the BL. On several occasions, Magda brought to committee meetings samples of new recipes she was trying out for the lunch club for feedback from attendees, and once invited me and several committee members back to her house after a meeting to try her food and see her kitchen, newly renovated in anticipation of her catering venture. Furthermore, people with whom Magda had become friendly through being involved in the BL were among those who attended her first lunch club event. Thus, Magda’s pursuit of a personal, occupational goal was individual but also collective, such that it emerged from, and at times was performed through the collective endeavour of the BL as a community intervention.

Thus, these ethnographic encounters serve to disrupt assumptions of simple aggregation of scale from the individual to the collective (or population) level, which underpin much evaluation research and its methods of measuring and interpreting the health impacts of an intervention. Instead, my ethnographic practice highlights the need to attend to the more dynamic and unpredictable flows between scales, and how they constitute each other through an unfolding complex intervention.

Discussion

Through this chapter, I have sought to highlight some of the ways the concept of ‘scale’ and scalar relations are embedded in current public health framings of the evaluation of complex interventions. Drawing on empirical examples from fieldwork exploring the enactment of ‘community’ through a complex, community empowerment intervention, I highlighted how ethnographic approaches enable the identification of flows of relations and interactions across scales as the intervention is delivered within a dynamic contextual system (Mowles 2014). Ethnographic research showed that enactment of community through the BL intervention engaged and occurred across different social, spatial and conceptual scales that were not distinct from one another, but instead frequently ‘enfolded in each other’ (Hastrup 2013: 148).

These ethnographic interpretations, produced through engaging in situated relations in two areas participating in the intervention, indicate the need for evaluation approaches to pursue a more nuanced understanding of how interventions unfold within a series of contextual structures and relations. While ethnography has been upheld as valuable for evaluation research in terms of the ‘holistic’ perspective it offers (Cohn et al. 2013) and for its attention to ‘context’ (Morgan-Trimmer and Wood 2016), recent debate has pushed for recognising further and potentially more critical roles for ethnography in public health evaluation (Reynolds 2017). The example described here shows a valuable role for ethnography in illuminating flows of relations across the different scales at which an intervention might bring about change, and how the processes of the intervention itself might disrupt those very scales and scalar relations, illustrated, for example, through the practices of representation within the BL. As such, ethnography may facilitate interpretation of the complex, dynamic pathways of interaction between intervention and context that existing evaluation approaches have failed to capture fully (Shoveller et al. 2015).

The ethnographic process in itself is also an example of the dynamic scalar complexities of producing more abstracted and generalised accounts from personal, situated interactions. In switching between levels of focus, from following individual trajectories to interpreting collective processes, my ethnographic approach highlights the potentially problematic basis of traditional analytical approaches within evaluation which rest on aggregating information from the individual to the population level. The fluidity of the ethnographic perspective across these scales creates an opportunity to consider the ‘incommensurability’ (Lambert 2013) of the analytical basis of public health evaluation, where health is measured (and felt) at an individual level but extrapolated to function as ‘evidence’ for decision-making at the population level. Rather than taking the target population of a complex intervention as a stable, aggregate unit of analysis, an ethnographic approach enables greater reflection on how the scales of the categories employed in an intervention and evaluation are varyingly performed, shaping and shaped by the mechanisms of effect that bring about change to health. Therefore, ethnographic approaches offer much to the field of public health evaluation: highlighting the dynamic scalar relations in the processes through which complex interventions bring about change to health, and prompting reflection on how knowledge produced through evaluation might usefully inform public health evidence and decision-making at a population level.