Jump to ContentJump to Main Navigation
Communicative CapacityPublic Encounters in Participatory Theory and Practice$

Koen P.R. Bartels

Print publication date: 2015

Print ISBN-13: 9781447318507

Published to Policy Press Scholarship Online: May 2016

DOI: 10.1332/policypress/9781447318507.001.0001

Show Summary Details
Page of

PRINTED FROM POLICY PRESS SCHOLARSHIP ONLINE (www.policypress.universitypressscholarship.com). (c) Copyright Policy Press, 2022. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in PPSO for personal use.date: 03 July 2022

Studying narratives of participatory practice

Studying narratives of participatory practice

(p.49) Three Studying narratives of participatory practice
Communicative Capacity

Koen P.R. Bartels

Policy Press

Abstract and Keywords

This chapter provides some background to the ways in which the research was set up and carried out. It explains the interpretive analysis conducted to study communicative practices in three international cases and the narrative analysis used to make sense of the stories citizens and public professionals told about their experiences with participatory practice. It explains how the cases were selected and compared and what they add to the analysis and conclusions about communicative capacity. The chapter then explains what narratives are and how, in each of the three cases, stories were collected by qualitative interviewing of 20 local public professionals and citizens, while participant observation, document analysis, and feedback were used as additional methods to consider data about the same situations. Finally, it explains the grounded theory analysis process of identifying emergent categories, patterns and meta narratives, which formed the building blocks of the theory of communicative capacity.

Keywords:   interpretive analysis, narrative analysis, grounded theory analysis, qualitative interviewing, practices

[A]pparently innocuous storytellingcan do a great deal of work …

(John Forester, 1999, p 3)

This chapter provides some background to the ways in which the research was set up and carried out. It explains the interpretive, comparative approach used to study communicative practices in three international cases and the narrative analysis conducted to make sense of the stories citizens and public professionals told about their experiences of participatory practice. It explains how the cases were selected and compared and what they add to the analysis and conclusions about communicative capacity. The chapter then explains what narratives are and how, in each of the three cases, stories were collected through qualitative interviews with local public professionals and citizens, while observation (of participants), document analysis and feedback were used as additional methods to consider data about the same situations. Finally, it explains the grounded theory-building process of identifying emergent categories, patterns and metanarratives, which formed the building blocks of the theory of communicative capacity.

How public encounters in practice can illuminate theory

Instead of approaching public encounters from a predefined normative framework, I looked at what happens in everyday participatory practice. I did not start out with any concepts or theories to empirically confirm or enrich, but developed my own concepts and theories through an iterative process of data interpretation to theory and back again. This was done according to the principle of practice illuminating theory to ‘understand the conceptualized from the immersion in a practice that provides content to the concepts’ (Hummel, 1998, p 154). By engaging in thick description of actual practice – which, as the next section will argue, narratives can provide – we can grasp what happens in-between public professionals and citizens without relying on abstract understandings and concepts (Young, 2000; Bogason et al, 2002; Elías, 2010). It helps to grasp participatory democracy as embodied (p.50) in the experiences and encounters of public professionals and citizens. Moreover, by using an interpretive approach, we set ourselves up for surprise, letting understandings, categories and theories transpire from the data, from ‘the ones who actually and concretely embody participative practices’ (Elías, 2010, p 10). Comparing these practices as situated in diverse international contexts further illuminates the interweaving of public professionals and citizens and the communicative patterns, processes and capacity that transpire from their in-between. All this incredibly rich empirical data can then be related back to concepts and theories which surface as relevant – extending, criticising and integrating them – to develop a novel theory grounded in practice.

Looking at practice means examining the thoughts and activities with which individuals take part in the concrete situations they are implicated in. A practice is not an individual applying knowledge or contextually determined action, but is a social activity performed to participate in a meaningful and competent way in a situation. What makes your cooking taste delicious, an orchestra play in perfect harmony, or a meeting creative and integrative cannot be taught from a book or through instruction; it is an embodied experience that needs to be practised. Only through situated performances can we learn what it means to be a hairdresser, for example, as opposed to a boxer or fisherman, and what it takes to be good at it (Lave, 1988; Cook and Yanow, 1993; Wenger, 1998; Cook and Brown, 1999; Wagenaar and Cook, 2003; Wagenaar, 2004; Wagenaar and Cook, 2011). Due to their social, dynamic and evolving nature, we capture practices with action concepts – ‘all those terms that are used to describe doings as opposed to happenings’ (Fay, 1975, p 72; see also Stout, 2012b).

A practice does not reside inside the head of an individual, nor can it be read off contextual configurations. It is a transactional, or dialogical, process that exists in-between people as the rules-in-use and individual dispositions are evoked while interacting with others and the ‘push and pull’ (Wagenaar and Cook, 2011) of the situation. Practices are not mere working routines, habits of mind, standard operating procedures, technical knowledge, or specialised skills (Allison, 1969, pp 698-707), but transpire from both routine and improvised practical judgements as the situation ‘signals to the actor that certain actions are called for, but also that certain conventions, commitments, physical obstacles, normative beliefs, procedures or rules have to be taken into account’ (Wagenaar and Cook, 2003, p 150; see also Bordieu, 1972; Forester, 1993a; Wagenaar, 2004; Laws and Hajer, 2006; Healey, 2009; Freeman et al, 2011). While acting upon the situation, individuals have to make sense of what is going on, whether this is desirable, and what should (p.51) be done to change or sustain the situation (Argyris and Schön, 1976; Rein, 1983; Goffman, 1986; Rein and Schön, 1994; Laws and Rein, 2003). As this is always a social process, practicing is as much a process of learning what to do (know that and know how) as how to be (becoming and belonging), ‘always straddling the known and the unknown in a subtle dance of the self’ (Wenger, 1998, p 41).

To grapple with the nature and meaning of practices we cannot simply look at what people do, or take for granted what they say they do and what they think the situation is (Wagenaar, 2011, pp 18-19, 48-50). Instead, we need to critically examine the performance of their activities and their meaning within the social context at hand. Interpretivism is a broad tradition in the social sciences which assumes that the meaning of social activities derives from intrinsic intentions which cannot be read off social behaviour but need to be actively interpreted against the self-understandings of individuals and the functioning of this behaviour in the social context. Meaning is not the same as observable behaviour, is not a fixed entity that exists independent of actors or observations, and cannot be reduced to individual intentions or aggregate institutions. Rather, we can come to an understanding of meaning by reconstructing the life world of individuals in terms of their implicit (conscious and tacit) intentions, the ways in which these are enacted and communicated in processes of social construction, and the sociocultural rules, structures and categories within which all of this is situated (Yanow, 2000; Yanow and Schwartz-Shea, 2006; Wagenaar, 2011). Doing so requires interpretation of the ways in which physical, social, and linguistic artefacts are employed while communicating about the issues at hand (Yanow, 2003, p 242).

The methodological stream of interpretive policy analysis (IPA) interprets the language and argumentation through which policy is brought into being (Majone, 1989; Fischer and Forester, 1993). IPA provides methods for making sense of what happens in what Hajer (2003) calls the institutional void in which ‘there are no clear rules and norms according to which politics is to be conducted and policy measures are to be agreed upon’ (p 175). Due to globalisation, immigration and technological progress, we have witnessed considerable changes in the spaces in which policy making takes place, the amount of social and cultural difference in the composition of societies, and the degree of certainty about the effects of policies (Yanow, 2000; Fischer, 2003; Hajer and Wagenaar, 2003; Wagenaar, 2011). Policies are social constructs driven by particular (often conflicting) interpretations that reflect competing belief systems, sociopolitical power constellations, rhetoric, and ambiguous knowledge claims (Edelman, 1977). As no (p.52) indisputable criteria exist for settling on their nature, correctness or success (Rittel and Webber, 1973), IPA aspires to detailed empirical research of the concrete manifestations of daily policy practice, dialogical analysis of the divergent interpretations that emerge in the interactive, communicative process, and critical normative judgement of their inescapably political and practical implications (Fay, 1975; Forester, 1993a, pp 18-19; Hajer and Wagenaar, 2003; Wagenaar, 2007b; Bevir, 2010, pp 10-13).

The goal is not to develop causal relationships, generalised statements, and a definitive resolution, but to explicate the contingent meanings of a practice to aid reflection, learning and change (Wagenaar, 2011, p 309). IPA aims to aid policy actors in dealing with the ‘radical uncertainty’ (Hajer and Wagenaar, 2003, pp 9-10) they face in daily practice (Laws and Hajer, 2006). There is no Archimedean point from which they can resolve such controversial issues as multiculturalism, abortion, and biotechnology. Therefore, policy actors can never reduce the possibility that unanticipated or undesirable consequences emerge from the unforeseen interactions of the plural, complex, and contingent elements of a situation (Cook and Wagenaar, 2011, pp 14-15). As misunderstandings, tensions, and conflicts are inevitable, IPA aims at ‘enhancing the awareness of uncertainty and unawareness’ (Hajer, 2003, p 186). That means that research has to facilitate policy actors in asking intelligible questions about habitual ways of thinking and acting to explicate the meanings of a practice and, ideally, handle these more intelligibly and productively in their (inter)actions.

Despite this focus on studying and improving the contingent meanings of specific cases, IPA provides few criteria for case selection. But this is not necessarily a problem, as the cases

have a range of similarities at various levels of detail but they do not have any one essential property or set of properties in common. We do not master the new concept by discovering a rule that tells us when to apply it… Our grasp of the concept lies in our ability to provide reasons why it applies to one case but not another and our ability to draw analogies with other cases. We recognize the pattern when we can discuss whether or not it is present in other cases.

(Bevir, 2010, p 12)

According to Bevir, there is nothing wrong with a rather ad hoc approach to selecting cases, because the goal of the analysis is to make sense of contingent practices with the help of a (emergent) general (p.53) pattern or concept rather than the other way around. Defining a priori hypotheses and case selection criteria would imply deductively testing a pre-existing theoretical framework rather than an abductive exploration of a practice in order to arrive at such a theoretical framework. But that does not mean that deciding which cases to study is not guided by any logic; it is just that this is a logic of interpretive inquiry rather than of formal methods (Haverland and Yanow, 2012).11 As we are not supposed to know at the beginning what the case will be an instance of (Wagenaar, 2011, chapter 9), interpretive case selection follows the logic of the real world problem, puzzle or dilemma that motivates the research.

In the case of my research, the real world problem is that nowadays public professionals and citizens encounter each other more frequently and intensively, but often do not manage to communicate productively about the problems they come together to resolve. Why not? What actually happens when public professionals and citizens meet in participatory practice? How do their public encounters in participatory practice take shape and meaning? How can public encounters become more productive? As I sought to come to an understanding of this real world problem without a predefined framework, the main question was: how do public encounters give shape to participatory democracy in practice? The cases had to offer as many opportunities as possible to illuminate theories of participatory democracy in practice. Therefore, I selected cases which were instances of public professionals and citizens engaged in intensive communication in challenging contexts which complicated their efforts toward mutual understanding at joint sense making. In such cases I was most likely to be confronted with new practices and forms of interaction that could challenge established political institutions and theoretical categories. I thus set myself up for surprise in order to let insights and theories transpire from the embodied experiences of public professionals and citizens with participatory encounters.

The cases were each community participation projects aimed at structurally involving citizens in decision making and problem solving in their direct living environment (community or neighbourhood) in collaboration with local public professionals working (and possibly living) in that area (Fung and Wright, 2003, p 15). The geographical setting is a crucial element: public professionals encounter residents rather than clients, consumers or citizens. Residents are usually diverse in their backgrounds and needs, and while there might be all kinds of social contacts and networks existing between them, even in the absence of these, the common denominator is that they are neighbours. Their motivations for, expectations of, and roles in community (p.54) participation are related to a specific place and space which they experience on a daily basis. Either on their own initiative or, more commonly, as a result of explicit efforts by local professionals, they make decisions, make use of budgets, or carry out initiatives as part of a participatory policy process. Depending on the nature of this policy and the participating local public organisations, community participation can involve any issue relevant to the daily lives of the residents, be it in the built environment (housing, infrastructure, greenery, for example), social dynamics (safety, cohesion, festivities), the economic sphere (poverty, entrepreneurial activity, unemployment), and the ecological domain (litter, cleanliness, sustainability).

On the one hand, community participation forms an ambitious context for bringing public professionals and citizens together. Being personally faced with the seriousness and complexity of the problems in their local area on a day to day basis, public professionals and residents can be expected to have strong motivations to improve the quality of life in their community and appreciate their interdependence and the importance of positive, workable relationships. Based on communitarianism, the assumption is that communal participation in a certain geographical location will stimulate the ability of its residents to capitalise on their shared cultural heritage, social ties, interests or experiences (Pierre and Peters, 2000, pp 137-59; Taylor, 2003; Amin, 2005). Moreover, public professionals are supposed to integrate individual goals, interests, structures and practices into a close knit policy-making community (Perri 6, 2005). As representatives of their agencies in local areas, they are expected to share a belief in the intrinsic value of collaboration for effective and democratic problem solving and maintain good social relationships (Agranoff and McGuire, 2003; Denhardt and Denhardt, 2003; Torfing, 2005; Sørensen, 2006; Stout, 2010d).

On the other hand, many contextual factors can inhibit public professionals and citizens in actually resolving the problems they face (Healey, 2007b; see also Table 2.1, p 23). Communitarianism has been criticised for its unrealistic and overly optimistic depictions and expectations of ‘communities’. Despite the weak empirical support for ‘area effects’ and theoretical objections to the glorification of closely knit communities in our modern globalised and fragmented society, it upholds the idea that ‘neighbourhoods’, especially deprived ones, are spatially bordered ‘communities’ which have local and exceptional problems that can be solved through temporary policies targeted at engaging ‘the community’ (for example, Little, 2002; Delanty, 2003; Amin, 2005; Atkinson et al, 2005). Furthermore, community (p.55) participation is broadly defined and therefore embodies a wide variety of institutions and practices12 that cannot be measured against clear standards. Its language and rhetoric might be adapted to fit local or national purposes as a ‘spray-on solution’ (Taylor, 2003, p 2) without sorting any real effects. Public authorities usually take the initiative, set the agenda and control what gets spent and done, while participants are involved based on established role patterns and have a broad variety of different motivations.

The cases selected reflect this combination of high ambition and indeterminacy. Table 3.1 shows that Glasgow, Amsterdam, and Bologna had strikingly similar ambitions aimed at achieving such communitarian norms as ‘involving’ or ‘engag[ing]’ citizens for more ‘structural and sustainable conversations’, ‘influence’, and ‘social cohesion and inclusion’. At the same time, as the next chapters show, the local contexts, designs and practices were highly divergent, involving serious challenges in addressing housing, deprivation, safety, health and sociocultural problems. Moreover, the communicative practices of citizens and public professionals in the UK, the Netherlands and Italy are embedded in highly divergent contexts including, for example, the constitutional and cultural authority of the state, the division of power and responsibilities between levels of government, the role of civil society and quangos, and the history of participation. Hence, by comparing the ways in which public professionals and residents engage in improving the quality of life in their neighbourhoods, the opportunities to illuminate the theory of participatory democracy through public encounters in participatory practice are maximised.

Table 3.1: Policies for community participation (my translations)


to ‘support the development of active and informed communities that can engage with and have an influence on the community planning process’ (Glasgow Community Planning Partnership, 2004, p 6)


involving residents … both in setting goals on the neighbourhood level and in implementation and evaluation … [with] extra attention for difficult to reach groups … [as] in order to have structural and sustainable conversations’ (Gemeente Amsterdam, 2008b, p 5)


‘valuing the active citizenry, stimulating/promoting an increase of social cohesion and inclusion, involving persons who are usually disadvantaged or less inclined to participate’ (Comune di Bologna, 2008, art 40.1)

(p.56) Narrative analysis: examining what happens in-between

A narrative is a story people tell about a real or imagined situation or range of events that wittingly or unwittingly enables them to pinpoint what happened, make sense of these happenings, and express their evaluation of the situation. We do it all the time, for example when we tell our friends or family about how our day was. Likewise, public professionals and citizens make sense of their encounters through narrative, that is by giving detailed accounts of concrete experiences. A range of studies have shown that we can learn a great deal about modern governance by examining the stories policy actors tell to make sense of their personal experiences of the messy, conflict-ridden and complex nature of everyday practice (Kaplan, 1993; Roe, 1994; Vinzant and Crothers, 1998; Abma, 1999; Forester, 1999; Bevir and Rhodes, 2003; Maynard-Moody and Musheno, 2003; Wagenaar, 2004; Hendriks, 2005; Hampton, 2009; Van Hulst, 2013). As a result, narrative analysis has gained wide acceptance in public administration, policy analysis and political science (Hummel, 1991; White, 1992; Forester, 1993b; Patterson and Monroe, 1998; Fischer, 2003, chapter 8).

Narratives are not just random stories, but form a distinct mode of knowing, a form of linguistic expression – different from logico-deductive or informational statements – through which we go about making sense of the world and communicating our understandings and experiences. The study of narratives stems from the ancient Greek tradition of storytelling to convey wisdom and the religious practice of studying the meaning of sacred texts. From the 1960s, the disciplines of history, literary criticism and psychology developed it as a scientific method by analysing personal and social histories, myths, fairy tales and novels as primary devices/carriers of sense making through which we can access contextualised meanings (Bruner, 1986; Polkinghorne, 1988; Czarniawska, 2004; Elliot, 2005). In the social sciences, narratives illuminate the complexity of daily practice by revealing the multiplicity of values, knowledge, and identities being performed in the communication between policy actors (Wagenaar, 1997; Abma, 1999; Forester, 1999; Wagenaar, 2007b, pp 26-7, 42). As such, narrative analysis helps us to grasp what happens in-between public professionals and citizens as they talk about the problems they face together.

It is possible to take a hermeneutic, discursive or dialogical approach to narrative analysis (Wagenaar, 2011). In a hermeneutic approach, narratives help to make systems of meaning (traditions, routines, cultures) intelligible by accessing the subjective experiences of (p.57) ordinary individuals and, from their hidden assumptions, values and emotions, reconstructing how (a part of) social reality actually looks to them (see Durose, 2009 for an example). Such thick description (Geertz, 1973/1993) of what people actually do and think can be enlightening, but often comes at the cost of a more critical stance towards possible flaws in the self-understanding of social actors, the assumption that meanings exist independent of the observer and can be readily discovered, and the role of overarching structures and power configurations (Wagenaar, 2011, pp 46-50). In a discursive approach, narrative analysis reveals how linguistic structures and power configurations condition our knowledge and actions (see Dubois, 2009 for an example). This helps to reveal that daily practices inevitably depend on the historically contingent, taken-for-granted and dispersed institutions in which social categories, everyday relationships and governance systems are encapsulated. But, as a result, the role of active agency often remains ambiguously compromised by the pervasive force of overarching discursive structures (McAnulla, 2006; Wagenaar, 2012).

In a dialogical approach, narrative analysis helps to grapple with the in-between, or the ‘constantly evolving process of interchange’ (Fay, 1996 quoted in Wagenaar, 2011, p 55) through which policy actors try to come to an understanding of complex, contingent and changeable situations. These understandings are always partial and tentative, as they depend on the specific physical, social and temporal ways in which individuals are positioned in the world. In consequence, meaning is not fixed in the consciousness of individuals and does not reside exclusively in overarching institutional arrangements. Instead, Wagenaar (2011, p 57) explains, meaning is relational as it emerges from ‘our interactions with others and with the world’, and can be made insightful by capturing the ‘give-and-take’ that ‘emerges from the patterned activities we engage in when we grapple with concrete situations that present themselves to us as in need of being resolved’.

Narratives do dialogical, or relational, work for policy actors (Forester, 1993b): through storytelling, policy actors engage in an ‘active, never-ending activity of sense making in a world of action’ (Wagenaar, 2011, p 219) that enables them to make practical judgements and influence the course of policy processes. By revealing the purposes particular stories fulfil in their policy context, we can uncover how storytellers communicate with their direct and wider audiences. Narratives thus enable, and follow from, everyday practices. A crucial point here is that the shape and meaning of particular practices are not pre-determined and do not rest on a rational, objective assessment of their strengths (p.58) and weaknesses for the situation at hand, but, instead, derive from the relational work that narratives do.

Four characteristics enable this relational work: narratives are open ended, subjective, value laden, and action oriented (Wagenaar, 2011, pp 210-16). First, the course and meaning of stories are open to change: they are provisional representations of situations that can still take unexpected turns. What appeared to be the closure of a chapter can turn out to be merely a short passage. Second, stories are about the idiosyncrasies of specific people who face doubts and certainties about everyday issues. The details of the person and situation inevitably invite empathy, side taking, and judgement. Third, narratives are not just random stories about people and situations, but are intelligible moral constructs of characters operating in a specific setting. The story represents beliefs and values about bigger underlying issues and how things ought to work. Fourth, narratives function as warrants that provide justification for a particular course of action. The setting and plot of a story endow actions in an indeterminate and complex context with certainty and legitimacy. Hence, narratives transmit information, emotions, moral values, beliefs, visions and norms by weaving together these different bits and pieces of everyday human conduct into a meaningful whole.

Narrative analysis can be conducted in a variety of ways, with narratives, discourses and frames13 often used in combination (see Laws and Rein, 2003). For example, Hajer’s (1995) discourse analysis of environmental politics relies extensively on the analytical concept of storyline. Depending on the goals of the research and the nature of the data, a narrative analysis will tend towards either a holistic approach, providing detail-heavy descriptions of significant themes in actors’ lives, or a categorical approach focused on the identification of linguistic structures (see Chase, 2005; Elliot, 2005, p 38). Forester’s (1999) approach is characteristic of the former: he provides long excerpts from unstructured interviews to illuminate how planners’ daily practice is a matter of practical judgement, anticipation, imagination, emotional responsiveness, empathy and political sensibility. He then takes these skilfully apart to reveal the work their narratives do to enable practices within social, political and economic constraints. Alternatively, Gold and Hamblett (1999) take the latter approach by uncovering in short quotes how policy actors use linguistic devices such as intensifiers, markers, qualifiers and metaphors to structure their accounts of situations. In this book, I use a mix of holistic and categorical approaches, depending on the need for a particular kind of analysis emerging from the data (see Table 3.2 for an overview). (p.59)

Table 3.2: Ways to structure narratives

Narrative element



Frame (actions, values, causal beliefs)

Rein and Schön, 1993; Laws and Rein, 2003

Reconstruct organisation of thoughts

Plot line (for example, ‘story of decline’) Beginnings, middles, and ends Characters (heroes and villains)

Stone, 1989; Kaplan, 1933; Roe, 1994; Stone, 2002

Reveal how causal and temporal order is imposed on a range of events

Diagnostic-prescriptive stories

Rein and Schön, 1994

Reveal normative leap from description to prescription

Signifiers (empty and floating) Metaphors (for example, health, tools, war)

Laclau, 1996; Jeffares, 2007 Stone, 2002

Categorise the vocabulary in use

For instance, when public professionals and citizens in Bologna are telling different stories about the same meetings, I analyse these as diagnostic prescriptive stories that ‘describe what is wrong with the present situation in such a way as to set the direction for its future transformation’ (Rein and Schön, 1994, p 26). When policy actors reflect on the current state of affairs, they select and label certain salient features of what is a complex and ambitguous situation and organise these features in a seemingly compelling and coherent way based on a ‘normative leap’ (Rein and Schön, 1994, p 26) from description to prescription. Alternatively, I analyse narratives as causal stories evolving around plotlines (Stone, 1989, 2002). For example, a ‘story of decline’ can legitimise immediate intervention in a situation to prevent some kind of disaster: ‘In the beginning, things were pretty good. But they got worse. In fact, right now, they are nearly intolerable, Something must be done’ (Stone, 2002, p 138). By comparing conflicting diagnostic prescriptive stories and plotlines, I can reveal the deeper tensions of seemingly innocent disagreements, as well as the underlying communicative pattern that sustains them.

As a third example, when public professionals and citizens in Glasgow constantly refer to the same ambiguous concepts in many different, and often conflicting, ways, I use signifiers to reveal how their vocabulary is a source of miscommunication. Signifiers are ambiguous, evocative and enticing concepts which provide a common language. Their abstract, broad brush, evocative and symbolic nature can have a binding and integrative function, because it creates the impression of equivalent interpretations. But it can also be a source of conflict because misunderstandings, tensions and conflicts are left implicit (Laclau, 1996; (p.60) Stone, 2002). A concept can either be a ‘floating signifier’, when it embodies multiple meanings that policy actors seek to enforce in a hegemonic discursive struggle, or an ‘empty signifier’ when critical mass has accumulated for a particular interpretation to represent the ‘true’ meaning of the idea or phenomenon (Jeffares, 2007). In each case, the vocabulary in use is analysed by identifying recurrent signifiers, counting how often they are mentioned during the interviews, and comparing the quantitative and qualitative differences in their usage.

The analysis goes through four levels of abstraction in total: first-order narratives, second-order narratives, metanarratives, and theoretical narratives. First-order narratives are the ‘bare’ stories that public professionals and citizens tell. This ‘raw data’ about their daily practices derives from qualitative interviews (explained in the next section). Second-order narratives are the analytical interpretations of firstorder narratives. I construct an account of how public professionals and citizens structure their narratives to render their experiences meaningful to themselves (conveying particular identities, values, beliefs, feelings, and so on) and to the broader context of which they are a part (why a specific audience has to appreciate the unusual or unexpected qualities and causal sequence of these events). Second-order narratives, then, focus on the interplay of the content, structure and performance of the stories (Elliot, 2005).

Metanarratives are the overarching stories of cases, covering a set of narratives, which reveal the significance of the situation in time and institutional context as well as how individual intentions and perceptions are related to this context and to each other (Kohler Riessman, 2002; Fischer, 2003, chapter 8). Metanarratives are collections of second-order narratives that reveal how the same range of events, activity, or phenomena is interpreted, acted upon and valued in different ways, and how public professionals and citizens structure their narratives to organise these different interpretations. By comparing various second-order narratives, we can make out the habitual communicative patterns that together characterise the three cases. An important strategy in this process is to identify the ‘dominant narrative’ that supports the most common mode of communication and compare this to the ‘counter narrative(s)’ that emphasise(s) contrasting or conflicting practices (Roe, 1994).

From there, I develop the theoretical narrative: the overall story of the research which synthesises the empirical results, presents the theoretical framework that emerges, and clarifies the links and contribution to the literature. Doing so involves an iterative process of reviewing literature and redrafting in order to further develop the emergent theories and (p.61) concepts by drawing on, extending and integrating existing theories and concepts. It is a tricky balancing act to relate and present the emergent theories and concepts to relevant literature without losing their originality, validity and nuance or giving the reader the impression that they are theoretical categories which were merely confirmed by the empirical findings. One way of going about this is the method of ‘ideal typing’, in which the emergent categories are abstracted into analytically ‘pure’ forms that do not empirically exist. Ideal typical constructs provide an analytical yardstick for interpreting and evaluating the divergent practices, patterns and outcomes in specific cases (Weber, 1949; Rutgers, 2001; Stout, 2010c). Existing theories and concepts can be used as ideal typical constructs, as long as their relevance derives from the emergent categories and their meaning is grounded in the everyday practices illuminating theory (the next section will discuss how this can be done through ‘memo writing’ and ‘theoretical sampling’). As such, narrative analysis helps to develop a theoretical understanding of how public professionals and citizens sustain and overcome patterns of communication through their narratives.

Grounded theory: dialogical analysis of practice

Analysing narratives is a highly sensitive, subtle and intricate process. A variety of qualitative and quantitative approaches can be used in this process (Abma, 1999; Elliot, 2005; Jones and McBeth, 2010). For this book, the stories public professionals and citizens tell about participatory practice have been interpreted through grounded analysis of qualitative data (Wagenaar, 2011, p 10). This is not the same as qualitative research (see Ritchie and Lewis, 2003; Denzin and Lincoln, 2005); it means that the meanings of narratives are actively interpreted by going back and forth between different practices, concrete situations and local context. It is an iterative dialogical process of confronting initial assumptions with empirical data, generating analytical categories from this process, comparing these categories again with more empirical data and so on. This means constantly moving back and forth between pre-existing assumptions, newly acquired information and potentially relevant theoretical explanations. It is an extensive, iterative and dialogical process of observing, questioning and reflecting on the pluralism of meanings produced by a multitude of persons as they talk about and move around in daily practice (Yanow, 2000; Yanow and Schwartz-Shea, 2006; Wagenaar, 2011).

Grounded theory does not provide any strict rules for dialogical analysis, but offers heuristics – sensitising concepts and strategies of (p.62) inquiry – that guide the researcher in being reflective, systematic and grounded (Dey, 1999) while gradually building a theory grounded in the data. The goal is to gather high quality data (‘deep’ or ‘thick’ descriptions of actors’ actual thoughts, experiences and actions) and analyse these data to formulate original and cogent theories abstracted from and illustrated by the data. Rather than proposing one best way of analysis, the main concern is to monitor the ‘quality of qualitative research’. Glaser and Strauss, who first developed grounded theory in 1967, considered that much of the research in the social sciences labelled as qualitative forced findings into speculative theoretical categories which were not firmly grounded in the empirical material (Charmaz, 2005). In this section I explain how I conducted 59 qualitative interviews and analysed them through a grounded theory process of coding, memo writing and theoretical sampling (Charmaz, 1990, 2002, 2006).

In each case study, I first examined all the relevant policy documents and research that I could obtain in order to have a decent background in the local governance system, community participation policies, and the neighbourhood. To find my way in each city, I got in touch with well embedded contact persons at the neighbourhood and/or city level who helped me to draft an initial list of interview candidates, facilitated my access to important meetings during my research period, and provided me with relevant policy documents. Over the course of about three months, I then conducted 20 interviews with a variety of public professionals and residents. Their selection developed organically according to emergent themes and suggestions made by other participants, but in general I tried to talk to people with varying types of experiences and expertise. I tried to maximise variation in backgrounds by contacting public professionals working in different organisations and functions, and residents living in different parts of the neighbourhood and for varying periods of time (see Table 3.3).

Intensive interviewing formed the main thrust of the data collection process, while observation of participants at meetings and document

Table 3.3: Summary of data















Policy documents





Transcribed text (pages)





Memos (pages)

9 (64)

8 (59)

8 (88)

25 (211)

(p.63) analysis were also used to check the reliability (Fischer, 2003, pp 154-5) of the first-order narratives provided by the respondents and the second-order narratives that came out of my analysis. The analysis of policy documents provided me with background information about the history and characteristics of the neighbourhood and project necessary for being a knowledgeable interviewer and becoming aware of underlying values, beliefs, goals and issues. By observing participants at meetings, I obtained an impression of the atmosphere, issues and practices that residents and public professionals were engaged in. I often noticed that themes from the interviews also emerged during the meetings as well as how various participants reconstructed these in different ways. The notes taken during meetings, together with general fieldwork notes, formed an important source of reflection on the course of the research (Fielding, 2001).

For the qualitative interviews, an in depth and intensive approach was taken (Weiss, 1994; Holstein and Gubrium, 1995; Gubrium and Holstein, 2002; Johnson, 2002; Wagenaar, 2011, chapter 9). The goal of the interviews was to develop the narrative of the participants, to provide a window onto their experiences. To do that, I tried to build a ‘working relationship’ with the participants (Wagenaar, 2011, pp 252-3). The key to high quality narratives, or good qualitative data, is to make sure that the participants understand the goals of the research and what is expected of them, as well as that they feel comfortable enough for self-disclosure. If such a working relationship is absent, participants will provide short and generalised statements, good intentions and opinions, instead of the detailed, open and extensive descriptions of personal experiences that narratives consist of. Therefore, I interviewed participants in a setting they were familiar with (their office or a public meeting space) and explained in advance the purposes of the research and the interview. Most importantly, I always stressed that I was looking for detailed and concrete experiences, and that I was there to learn from them because they were the expert, not me. Although people often modestly denied the significance of their experiences, I made a sincere effort to convince them of the opposite.

During the interviews, my most important task was to monitor the quality of the data. This is a difficult task, because the interview should develop as a directed conversation or unfolding story, which on the one hand is informal, open ended, and guided by what the respondent is saying, while, on the other hand, follows a general list of themes and questions to keep the participants on topic (Charmaz, 2002; Wagenaar, 2011, chapter 9). I always let the wording and order of the questions vary according to what a participant was telling me, thereby (p.64) qualifying as an unstructured interview (Fielding and Thomas, 2001). But unstructured does not mean unsystematic: each interview had a beginning, middle, and end to ensure a natural build-up of questions and expectations and flow of information and disclosure. Therefore, I always started with the same opening question (‘You are a resident of [neighbourhood]/working here as [function], could you please tell me in which circumstances you first came to live/work here?’) and closing question (‘Based on everything that you’ve told me, what lessons do you draw for the future?’). The opening question is formulated relatively open, but nevertheless focuses attention on their personal history and actual experiences with regards to community participation by focusing on the concrete conditions under which they first came to live or work in the area. From their initial answer, many themes emerged that we could then explore further. The closing question induced participants to reflect and get a sense of closure.

In the middle of the interviews, I always formulated the questions in an active and open way to stimulate participants to provide detailed answers that did not sum up or skim over issues and experiences. I often had to ask for more elaboration or details of a statement, with probably the most often used question being: ‘Can you give me an example of a concrete situation in which you experienced this?’ Asking for examples guides a participant away from giving generalised answers or justifications that summarise rather than describe experiences. When determining the next question, I tried to create a natural transition from what a participant had just said by either continuing on the same topic or coming back to a topic that had been mentioned previously (which I noted on my note pad with so called ‘markers’). I further monitored the natural flow of the interview by being alert to emotional signals, encouraging them to tell more by verbal and nonverbal cues, not intruding on the participant’s story, and talking about myself only briefly and without disclosure when the participant asked me something (Weiss, 1994; Charmaz, 2002; Legard et al, 2003).

Perhaps the best illustration of this interviewing approach comes from an interview with a police officer in Amsterdam. After talking for around 20 minutes about her daily activities, the neighbourhood, and contact with residents, I asked how the introduction of the Neighbourhood Approach – the local community participation project – had affected her daily practice. She responded by asking me what I meant by ‘the Neighbourhood Approach’, because for her it was completely unclear what this project had actually changed to her contact with residents. I responded that I thought it was very interesting that for her the change was unclear and encouraged her to tell me more (p.65) about this. In this way I not only diverted attention away from myself, but more fundamentally invited her to develop her narrative about struggling with the nature and added value of the project. It also led me to focus in other interviews on what participants understood to be the nature and added value of the project, discovering that it did not have very clear rules and structures and was only an overarching label for a broad array of diverse and ongoing activities, and identifying the flexibility of the policy framework as an important conceptual issue.

As such, the data gathering of my research followed a grounded theory process to reveal ‘constructions or competing definitions of the situation as given in action, not merely stated in reconstructed accounts’ (Charmaz, 2006, p 180; emphasis added). The quality of the data is in the level of concreteness and detail of the respondents’ narratives. This enabled me to raise analytical issues out of diverging stories, and translate these into conceptual categories and, then, theoretical explanations. I recorded and transcribed the interviews word for word in order to get a firm grounding in what a respondent had actually said. By transcribing verbatim, not omitting ‘ums’, repetitions, errors, and incoherent sentences, I could present utterances in their actual, real world, narrative form and tease out doubts and inconsistencies in the thinking of a participant – even though some of this has undoubtedly been lost in translation.

The analysis of the transcribed interviews started with coding: ‘the process of defining what the data are about’ by ‘naming segments of data with a label that simultaneously categorizes, summarizes, and accounts for each piece of data’ (Charmaz, 2006, p 43). The most important question to ask when looking at interview data is: What is this an instance of? This is the first step from description to conceptualisation: what is actually going on in the data? Codes have to be active and evocative words that are narrow enough to describe detailed parts of data and broad enough to represent underlying assumptions and feelings or broader patterns and tensions. In other words, codes are sensitising concepts that stick closely to the view of the participants as well as lead to theoretical categories. I always tried to stay close to the data by describing participants’ thoughts and activities on their own terms, rather than resorting to already existing theories or acquired insights, to see things that would be overlooked from a priori assumptions.

While the use of sophisticated software packages such as NVivo are popular for grounded theory, I analysed interviews in a perhaps more crude and time consuming, but also more creative and engaged way by using the comment function of Microsoft Word to label segments of data, or ‘meaning units’ (Wagenaar, 2011, p 262). A meaning unit can (p.66) vary from a sentence to a chunk of text as large as a page, depending on what that piece of data is an instance of. The codes emerged and developed from an iterative interpretive process, with each interview generating some new codes – even though this happened less towards the end of the research process. Initial coding led to a variety of detailed codes in the beginning and gradually turned into focused coding, in which I only started to use the most significant and frequent codes. In each case I tried to start afresh rather than importing codes from the other cases, but several codes proved to work in making sense of specific practices in each case. While analysing an interview, I noted the codes and the numbers of the comments in a separate document together with some initial explanations of their meaning. This document then formed the basis for a memo, which I wrote according to the themes that emerged from the comparison of several interviews.

Memo writing was a pivotal step between initial analysis and writing drafts as it helped to evaluate the data and analysis; to explain codes, link them to each other, develop ideas, and fine tune subsequent data search (Charmaz, 2006). Several techniques that helped to focus memos were to give them a title, define categories, and discuss where the categories and data were leading the research. Furthermore, numerical and graphical representations of codes, actors and connections aided insight into the relationships between codes and categories (Roe, 1994, pp 155-62). Initial memos were detailed reconstructions of first-order narratives, which I interpreted by analysing how the narratives were structured in terms of plotlines, signifiers, frames, and so on. In more advanced memos I developed metanarratives by defining categories, identifying gaps, and looking for patterns. Gradually, the focus changed from comparing data with data towards creating dialogue between theory and data. Indeed, while my first memos were analyses of individual interviews of about two to three pages, final memos were theoretically driven treatises of about 20 pages. As such, memo writing furthered the grounded theory process from analytical categories as descriptive and synthesising tools to conceptual categories that serve theoretical definition and the production of a metanarrative.

Finally, the process turned to perhaps the trickiest part: theoretical sampling. This refers to the stage in which one should be more confident in judging what the most relevant parts of the analysis are and how the data could be rearranged and ordered. Theoretical sampling means turning the inductive process on its head, by taking the emerged theory and returning to the collected data and/or to the field for additional data. Data analysis is now supposed to be guided by the emerged theoretical categories in order to reconfirm their (p.67) presence, refine their properties and links to the broader context, and spot possible flaws and unforeseen insights (Glaser and Strauss, 1967; Charmaz, 2006). Grounded theory leaves unclear, though, when exactly this salto mortale should be taken or when it is complete (Dey, 1999). In this process of theorising, I wrote several theoretical memos as well as feedback reports for each of the three cases. The latter forced me to explain in clear terms what I had found and what that meant; the former helped me to write freely about my ideas and dilemmas in formulating a theory. Through this process of going back and forth between theories and data, I started to see that my findings were substantially about communication and in the end came up with the theoretical concept of communicative capacity for rearranging and making sense of the data. Continuous redrafting and iterative literature reviews led me to develop and fine tune the emergent theoretical framework.

Final remarks

Taking a comparative, interpretive, theory-building approach shows that, as John Forester (1999) states, ‘apparently innocuous storytellingcan do a great deal of work’. Narrative analysis helps to examine how public encounters give shape to participatory democracy in practice. It illuminates what happens in-between public professionals and citizens through a dialogical process of grounded theorising. By telling stories rather than articulating their views in abstract arguments, professionals and citizens help us to imagine what it is like to be there and experience the situation in all its overwhelming messiness and indeterminacy. This enables us to appreciate the subtleties in activities and expressions, discover the hidden meanings of words and behaviours, see the practical ways in which theories and policies manifest themselves in the real world, and feel the difficulties of practical judgement and emotional struggles (Forester, 1999; Young, 2000; Schein, 2003; Ryfe, 2006; Petts, 2007). By comparing across cases and theorising from these micro practices, we can get a better understanding of what communicative capacity is, why it is so challenging, and how we can exercise it. (p.68)


(11) In fact, Wagenaar (2011) argues that heuristics, rather than methods, is the correct term for the kind of rules and strategies of inquiry that guide interpretive research, as the inductive process harbours so many unexpected turns and dilemmas that there is no real method in managing the research.

(12) In the cases reported here, the terms in use were citizen participation (partecipazione dei cittadini – Bologna), community engagement (Glasgow), and resident participation (bewonersparticipatie – Amsterdam).

(p.238) (13) Frames are cognitive definitions of problematical situations which organise thoughts to facilitate, or legitimate, action by linking events or actions (what has been done or will be done) to values (what should be done) and causal beliefs (what has brought about this situation or will bring about a desired situation) (Rein and Schön, 1993)

(14) Final interview in Glasgow was cancelled at the last moment and I was therefore unable to arrange an alternative before the end of the fieldwork period.