This post develops the theoretical argument introduced in the series introduction on dashboards as policy instruments: that public-sector performance dashboards do not merely serve pre-existing publics but actively constitute them. The claim is not especially dramatic, but it turns out to be one that needs stating explicitly because it is largely absent from current public-sector dashboard strategy. Dashboard design choices are fundamentally political because they determine who is assembled as a public, around what issues that public can form, what capacities for action it possesses, and what modes of participation are available to it.
The current transparency agenda assumes that providing access to data will automatically create an "active public" capable of holding providers accountable. The theoretical literature suggests this assumption is unfounded; data access constitutes a consumer or spectator public unless additional infrastructure supports the transition to deliberative, agonistic, or active modes. What follows is an attempt to trace the theoretical foundations for that claim through Dewey's pragmatist theory of publics, Marres's concept of material participation, Le Dantec's framework for designing publics, and DiSalvo's adversarial design.
The Lippmann-Dewey Debate: Two Views of the Public's Problem
Lippmann's Challenge
The theoretical foundation for this analysis emerges from the famous Lippmann-Dewey debate of the 1920s - a dispute about democracy in technological societies that remains strikingly relevant to public-sector dashboard design.
Walter Lippmann, in Public Opinion (1922) and The Phantom Public (1927), mounted a critique of the classical democratic assumption that citizens can be adequately informed about public affairs. His central observation was that the world we must act upon is fundamentally different from the world we can directly perceive. The opening of Public Opinion illustrates the problem: on an island in 1914, English, French and German residents lived as friends for six weeks after war had been declared, because no news had reached them - "there was a moment when the picture of Europe on which men were conducting their business as usual, did not in any way correspond to the Europe which was about to make a jumble of their lives" (Lippmann, 1922, pp. 1-2).
Lippmann generalised this insight: in complex technological societies, citizens necessarily act on "pictures in their heads" rather than direct knowledge of the world. These mental pictures are shaped by stereotypes, filtered through media, constrained by the limited attention people can devote to public affairs. The "public" that democratic theory imagines - informed citizens deliberating on common concerns - is, Lippmann argued, a "phantom". His definition was correspondingly modest: "the public is not, as I see it, a fixed body of individuals. It is merely those persons who are interested in an affair and can affect it only by supporting or opposing the actors" (Lippmann, 1927, p. 67).
Crucially, Lippmann emphasised that publics are called upon precisely when problems are most intractable. "Where the facts are most obscure, where precedents are lacking, where novelty and confusion pervade everything, the public in all its unfitness is compelled to make its most important decisions" (Lippmann, 1927, p. 121). This created what he saw as an impossible situation: the public is summoned when expert knowledge fails, yet the public lacks the very expertise required to address the problems it faces.
Dewey's Response
John Dewey's The Public and Its Problems (1927) was explicitly a response to Lippmann. As Marres (2015, p. 10) notes, "Dewey himself pointed out, he first got his ideas about the material public by reading the work of the journalist Walter Lippmann". Dewey accepted much of Lippmann's diagnosis but rejected his conclusions. Modern technological society does make it difficult for publics to form and act - but this is not because the public is a phantom. It is because the conditions for public formation have been disrupted and must be reconstructed.
Publics as Constituted, Not Given
Against the assumption that "the public" exists as a natural entity waiting to be served, Dewey argued that publics must be constituted - brought into being through processes that identify shared issues and assemble affected parties. He defined the public as consisting "of all those who are affected by the indirect consequences of transactions, to such an extent that it is deemed necessary to have those consequences systematically cared for" (Dewey, 1927, pp. 15-16). Le Dantec (2016, p. 5) puts this more plainly: a public is a "unique federation of people influenced or impressed upon by a specific set of conditions".
Two features distinguish this Deweyan conception. First, publics form around issues, not identities. A public is not a demographic category or pre-existing community; it comes into being through the articulation of shared conditions and actions taken to address them. The public does not exist prior to the issue.
Second, publics require infrastructure to become aware of themselves. Dewey identified a fundamental problem: the complexity of modern society means that "consequences are felt rather than perceived; they are suffered, but they cannot be said to be known, for they are not, by those who experience them, referred to their origins" (Dewey, 1927, p. 131). People may be affected by the performance of healthcare systems without understanding that they are affected, or how, or by whom. This is Dewey's "problem of the public" - the difficulty of assembling a collective that can identify itself, understand its shared conditions, and take collective action.
Technology and Public Formation
Dewey recognised that technologies play a crucial role in the formation - or failure - of publics. Technology simultaneously creates the conditions for publics to form, by generating shared consequences, and makes it more difficult for publics to constitute themselves, by rendering those consequences opaque, distributed, and difficult to trace. As Dewey put it, the "machine age has so enormously expanded, multiplied, intensified, and complicated the scope of the indirect consequences" that "the resultant public cannot identify and distinguish itself" (Dewey, 1927, p. 126).
This is precisely the situation of public service performance data. The consequences of public service performance - waiting times, quality variations, resource allocation - affect millions of people. But those consequences are mediated through systems that make it difficult for affected parties to identify that they are affected, understand how they are affected, trace effects to causes, or identify others similarly affected with whom collective action might be possible. Dashboards enter this context as potential infrastructure for public formation - or as devices that foreclose such formation.
The Problem of Relevance
Marres (2015, p. 10) identifies a crucial distinction between Lippmann and Dewey's formulations. Both defined the public in terms of being affected by issues, but they understood the resulting problem differently. For Lippmann, the more broadly we define the public, the less relevant the issue becomes to any individual member; stakeholders care intensely while the general public cares diffusely, which leads toward elite governance. For Dewey, the public's combination of intimate affectedness and outsider status is not a flaw to be overcome by deferring to experts but a distinctive political condition requiring infrastructure for its articulation.
As Marres summarises, "the public consists of actors who are intimately affected by issues, yet are not participants in the networks, platforms and vocabularies of issue articulation" (Marres, 2015, p. 10). This describes the situation of public service performance data exactly. Citizens are affected by public service performance, but they are outsiders to the professional networks, technical vocabularies, and institutional platforms where performance is defined, measured, and acted upon. The question is whether dashboard design addresses or ignores this problem of relevance.
Material Participation: How Objects Constitute Political Subjects
Marres and the Material Public
Noortje Marres's Material Participation (2012/2015) extends the Lippmann-Dewey debate to examine how material objects and technologies specifically enable or disable public participation. Her central argument is that participation is not simply a matter of discourse or deliberation; it is accomplished through material practices and devices.
Marres notes that Lippmann himself attended to how material settings shape the possibilities for public engagement - the commuter reading headlines on the train has limited attention; the industrial worker's "thought goes on in a bath of noise". These material conditions are not merely background; they constitute what kinds of public engagement are possible. Applied to public-sector dashboards, this observation has direct force: how executives consume performance data - in board meetings, on mobile devices, through printed packs - shapes what kind of public engagement with performance is possible. The five-minute window between meetings is not just a constraint to design around; it constitutes a particular mode of engagement that forecloses certain forms of deliberation while enabling others.
Marres introduces the distinction between nominal and effective participation. The conventional view treats technology as merely instrumental - providing tools that enable participation which is fundamentally discursive. Marres argues instead that material devices constitute forms of participation that would not otherwise exist: "to ask about the role of things in political participation is to disrupt a long-held assumption of political theory: the idea that material environments and technological mediators present no more than 'preconditions' for public participation" (Marres, 2015, p. 10).
Three Versions of the Community of the Affected
Marres identifies three versions of "the community of the affected" in political theory. The procedural (liberal) version holds that those affected by an issue have a right to participate in decision-making about it; for dashboard design, this implies identifying who is affected and providing access. The materialist version proposes that political communities form through factual relations of material harm; dashboards should make visible the material consequences of public service performance. The post-instrumentalist version - drawing on Latour, Stengers, and others - goes further, arguing that objects themselves have capacities to provoke, challenge, and organise publics; dashboards are active agents in constituting publics, not neutral conduits.
The post-instrumentalist position is most relevant to this analysis. It suggests that dashboards are not merely representations of performance data but devices that actively shape who is assembled as a public and what capacities for action they possess.
Marres's formulation of the problem of relevance captures the core tension: "the problem of the public is that it is at once intimately affected by issues, but also finds itself at a remove from the platforms that are in place to address them" (Marres, 2015, p. 10). The design of dashboards determines whether affected parties can actually constitute themselves as a public capable of action, or whether they remain dispersed, passive, and unable to connect their experiences to systemic causes.
Designing Publics: Issues, Attachments, and Infrastructuring
Le Dantec's Framework
Christopher Le Dantec's Designing Publics (2016) operationalises Dewey's theory for design practice. He identifies three entangled components through which publics are constituted: issues, attachments, and infrastructuring. Issues are the shared conditions around which a public might form - not objectively given but requiring articulation, being made visible, named, and connected to affected parties. Attachments are the internal relations that bind a public together, including commitments to shared goals, dependencies on shared resources, and affective bonds sustaining collective action. Infrastructuring is the ongoing work of building and maintaining the sociotechnical systems that support a public's capacity for action. As Le Dantec (2016, p. 9) summarises, "issues, attachments, and the work of infrastructuring are the entangled components that constitute a public".
Le Dantec and DiSalvo (2013) develop the concept of infrastructuring specifically in the context of participatory design, arguing that the formation of publics is not a one-off design outcome but an ongoing process of building and maintaining the sociotechnical arrangements through which collective action becomes possible.
Application to Public-Sector Dashboards
Applying Le Dantec's framework to public-sector dashboard design reveals how design choices determine what kinds of publics can form. On issue articulation: what issues does the dashboard make visible, what issues remain invisible through what is not measured or excluded from summary scores, and can users trace aggregate scores to specific impacts on their communities? On attachment formation: does the dashboard support users in identifying others similarly affected, are there mechanisms for collective interpretation and action, or does it position users as isolated individual consumers of information? On infrastructuring: does the dashboard support ongoing collective action on identified issues, or does it treat data access as the end point of the transparency obligation?
The current performance scoring framework dashboard, analysed through this framework, presents aggregate scores without connecting them to specific communities or impacts; offers no features for collective interpretation or user connection; and treats data access as the terminus, with no support for action beyond individual consumption. The gap between what these components require and what the current design provides is substantial.
The Dual Nature of Designing Publics
Le Dantec (2016, p. 9) identifies a crucial duality: publics are both designed - constituted through intentional choices about infrastructure - and designing, actively constituting themselves through ongoing practices. The public sector transparency agenda has focused almost exclusively on the first dimension, designing dashboards that constitute passive publics as recipients of information. The second dimension, supporting publics in designing their own capacities for collective action, is largely absent.
Adversarial Design and Agonistic Publics
DiSalvo's Agonistic Framework
Carl DiSalvo's Adversarial Design (2012) introduces the concept of agonism to design theory. Drawing on Chantal Mouffe's (2000) political theory, DiSalvo argues that design can do the "work of agonism" - maintaining the productive contestation that democratic politics requires. As DiSalvo puts it, "agonism is a condition of forever looping contestation. The ongoing disagreement and confrontation are not detrimental to the endeavor of democracy or the project of governance. Rather, agonism is foundational to democracy, providing the basis for robust debate and decision making" (DiSalvo, 2012, p. 3).
This contrasts with the consensus model implicit in most public-sector dashboard design, which assumes that providing "objective" data will lead to shared understanding and agreement. The agonistic perspective suggests instead that contestation about the meaning, validity, and implications of data is constitutive of democratic engagement with performance information.
Revealing Hegemony
DiSalvo identifies "revealing hegemony" as a key tactic of adversarial design: "a tactic of exposing and documenting the forces of influence in society and the means by which social manipulation occurs" (DiSalvo, 2012, p. 5). Applied to public-sector dashboards, this prompts uncomfortable questions. Who decided which metrics would be included in the oversight framework, and what interests do those metrics serve? The methodology for calculating composite scores embeds value judgements about what matters - who made those judgements? The decision to publish at Trust level rather than service level, or to aggregate quarterly rather than monthly, reflects choices about what kinds of accountability are enabled.
A dashboard designed to reveal hegemony would make these choices visible and contestable, inviting publics to engage not just with the data but with the infrastructure that constitutes the data as meaningful.
Political Design Versus Design for Politics
DiSalvo (2012, p. 10) draws a distinction between design for politics - design that serves existing political processes, such as a dashboard that helps executives monitor performance - and political design, which does the work of agonism by constituting publics, articulating issues, and enabling contestation. Most public-sector dashboard work is design for politics, serving the existing transparency agenda without questioning its assumptions. Political design would involve dashboards that make visible the choices embedded in data infrastructure, enable contestation about those choices, support the formation of publics capable of collective action, and maintain productive agonism rather than foreclosing debate through the rhetoric of "objective data".
DiSalvo's agonistic framework and Ehn's (1988) conception of design as emancipatory practice work the same seam from different directions. Ehn (1988), rooted in Scandinavian workplace democracy, asks who gets to shape the tools and systems that structure work; DiSalvo asks whether the resulting artefacts maintain or foreclose political contestation. The dashboard question sits at the intersection: who is involved in shaping the data infrastructure, and whether the resulting artefact supports or forecloses the constitution of publics capable of collective action.
How Public-Sector Dashboards Constitute or Foreclose Publics
The Current Situation
Analysing public-sector performance dashboards through this theoretical framework reveals that they currently constitute a particular kind of public. Users are positioned as passive consumers of information rather than participants in its constitution or contestation. No infrastructure supports users in identifying others similarly affected or coordinating collective interpretation and action. Data access is treated as the terminus of the transparency obligation, with no support for translating data into collective action. And the rhetoric of "objective data" forecloses contestation about the choices embedded in data infrastructure.
What Kinds of Publics Are Being Constituted?
Different dashboard designs constitute different kinds of publics. A consumer public consists of individual users accessing information for private purposes; this is the current self-service model with no collective features. A spectator public consists of passive observers of performance information, produced by a publication-focused model of official statistics with no interactivity. A deliberative public would require annotation features, discussion forums, and collective sense-making tools. An agonistic public would need methodology visibility, contestation mechanisms, and the ability to propose alternative metrics. An active public - one capable of collective intervention - would require action pathway support, coalition-building tools, and feedback loops to providers.
The public sector transparency agenda assumes that providing access to data will automatically constitute an active public capable of holding providers accountable. The theoretical literature suggests this assumption is unfounded. Data access constitutes a consumer or spectator public unless additional infrastructure supports the transition to deliberative, agonistic, or active modes. The claim that current dashboards rest on untested assumptions about how publics form and act is one I have developed empirically in 135 Hypotheses, which documents the full set of assumptions embedded in public-sector executive dashboard design and assesses the evidential confidence behind each one; roughly a third of the design assumptions underpinning current dashboards have weak or no evidential support.
Design Requirements for Public-Constituting Dashboards
Drawing together the theoretical frameworks, dashboard design that supports public constitution would need to address several interconnected requirements. Le Dantec's framework implies connecting aggregate data to specific communities and impacts, making visible what is not measured as well as what is, and supporting users in tracing systemic causes of experienced effects; it would also need to enable identification of similarly affected parties and support collective interpretation and sense-making rather than positioning users as isolated consumers, building ongoing capacity for collective action rather than treating data access as the terminus.
Marres's account of material participation adds that the material conditions of data consumption - the five-minute window, the printed board pack, the mobile screen - are not merely constraints to design around but constitutive of what kinds of engagement are possible; designing for effective rather than merely nominal transparency means attending to these material conditions and addressing the problem of relevance by connecting affected parties to platforms for action. DiSalvo's agonistic framework pushes furthest: the choices embedded in data infrastructure - which metrics are included, how scores are calculated, what level of aggregation is used - should themselves be visible and contestable, with mechanisms for productive disagreement rather than the foreclosure of debate through the rhetoric of objective data.
Implications for Public-Sector Dashboard Strategy
The Fundamental Reframe
This theoretical analysis suggests a fundamental reframing of public-sector dashboard strategy. The question shifts from "how do we provide performance data to the public?" to "what kind of public do we want to constitute, and what infrastructure would support that constitution?" The first question treats the public as pre-existing and data as neutral. The second recognises that dashboard design choices actively shape who is assembled as a public and what political capacities they possess.
I develop this reframe in two directions in later posts. Competing Theories of Change maps the distinct theories of change a dashboard might embody and examines the behaviour barriers that prevent intended use, while The Sunlight Hypothesis traces the historical evidence for each mechanism. The series introduction on dashboards as policy instruments draws these threads together into the structural argument that internal operational dashboards and external transparency publications are different policy instruments serving different theories of change, and that designing them as one thing serves neither purpose well. The Politics of Performance Transparency then applies Kimbell and Tonkinwise's (2025) framework on government service design politics to the question, developing the argument that public-sector dashboards simultaneously perform three government functions - rules enforcement, service provision, and policy implementation - while the service framing dominates and the other two remain implicit.
Strategic Options
Three strategic options emerge from the analysis. The first is to continue positioning dashboards as information services for individual users - the current consumer public model - accepting that this constitutes passive, individualised publics and that transparency in this mode is largely symbolic. The second is to redesign dashboards to support collective interpretation and discussion, adding features for annotation, commentary, and dialogue, while accepting that this opens data to contestation and may generate uncomfortable discussions. The third is to design dashboards as infrastructure for constituting politically capable publics: making methodology visible and contestable, supporting coalition-building and collective action pathways, and accepting that this transforms dashboards from information provision to democratic infrastructure.
The choice between these options is itself political. It cannot be resolved by technical analysis or user research alone. It requires explicit deliberation about what role performance data should play in public sector governance, what kind of accountability we want to enable, who has legitimate authority to interpret and act on performance information, and what modes of public engagement with public service performance are desirable.
The theoretical frameworks reviewed here do not answer these questions. But they do reveal that not asking them - treating dashboard design as a purely technical matter - is itself a political choice that defaults to constituting passive, individualised publics.
Dashboards as Pictures in Our Heads
Lippmann's central insight - that we act on mental pictures rather than direct knowledge of the world - has implications for dashboard design that extend beyond the publics question. Dashboards are the "pictures in our heads" of public service performance. They do not provide direct access to reality; they provide representations that shape how we understand and act upon that reality.
Lippmann identified several mechanisms by which these pictures become distorted. Stereotypes - pre-existing mental categories that filter incoming information - shape how dashboard data is interpreted before it is even examined; "under-performing service", "good performance", and "seasonal pressures" are categories that organise perception before the data is consulted. Decisions about what to include, exclude, aggregate, or disaggregate are acts of editorial selection. People have limited cognitive resources; the executive with a five-minute window between meetings cannot engage with comprehensive performance data, and this material condition of consumption shapes what can be understood. And there is always a lag between events and awareness; monthly reporting cycles mean performance problems may persist for weeks before becoming visible.
The constitutive role of representations is a thread I have developed across several earlier posts. Lakoff and Johnson's argument that metaphors are not decorative but constitutive establishes the foundation: we think through metaphors rather than merely communicating with them. Morgan's work on organisational metaphors extends this to institutional life, where the metaphor through which we see an organisation determines what we think is possible within it. Boland and Collopy's claim that our vocabulary of representations is critical in determining how well or poorly we do applies directly: the representational choices embedded in a dashboard determine what questions can be asked and what questions are structurally invisible. And the argument that representational formalisms conceal as much as they reveal is precisely Lippmann's point about stereotypes and editorial selection, restated in the language of design.
Dashboards as Democratic Infrastructure
The theoretical literature on publics, material participation, and adversarial design converges on a fundamental insight: technologies do not merely serve pre-existing publics but constitute them. Applied to public-sector performance dashboards, this means that design choices determine who is assembled as a public, what issues that public can form around, what capacities for interpretation and action it possesses, and what modes of political engagement are available.
The public sector transparency agenda has proceeded on the assumption that publishing data automatically creates accountability. Publication creates nominal transparency; effective transparency requires infrastructure that supports the constitution of publics capable of collective interpretation and action. Dashboards designed for "executive self-service" or "public transparency" without attention to how they constitute publics will produce the kinds of publics their design implies: individualised consumers of information, disconnected from each other and from the platforms where public sector governance actually occurs.
Whether the public sector transparency agenda should pursue the direction of dashboards as democratic infrastructure is a political question that this analysis cannot resolve. But the analysis does reveal that the question must be asked, and that treating dashboard design as politically neutral is itself a political choice with significant consequences for what kinds of publics public service performance transparency can constitute. The deficit model analysis I develop in a subsequent post examines one specific mechanism through which this foreclosure operates: the assumption that the public's problem is a lack of information rather than a lack of infrastructure for collective sense-making.
References
Dewey, J. (1927). The Public and Its Problems. Henry Holt.
DiSalvo, C. (2012). Adversarial Design. MIT Press.
Ehn, P. (1988). Work-Oriented Design of Computer Artifacts. Arbetslivscentrum.
Kimbell, L., & Tonkinwise, C. (2025). A political dialogue about government service design politics. In L. Penin, A. Prendiville, & D. Sangiorgi (Eds.), The Bloomsbury Handbook of Service Design (Chapter 6.3). Bloomsbury Academic.
Le Dantec, C. A. (2016). Designing Publics. MIT Press.
Le Dantec, C. A., & DiSalvo, C. (2013). Infrastructuring and the formation of publics in participatory design. Social Studies of Science, 43(2), 241-264.
Lippmann, W. (1922). Public Opinion. Harcourt, Brace and Company.
Lippmann, W. (1927). The Phantom Public. Macmillan.
Marres, N. (2012/2015). Material Participation: Technology, the Environment and Everyday Publics. Palgrave Macmillan.
Mouffe, C. (2000). The Democratic Paradox. Verso.