Introduction
For at least two decades, traditional representative democracy has been in ‘crisis’ (Dalton Reference Dalton2004; Flinders, Wood, and Corbett Reference Flinders, Wood, Corbett, Elstub and Escobar2019). Stemming from perceived and real deficits in core democratic functions, political disaffection, and public discontent in liberal democracies is on the rise. Deliberative democrats argue that the decline of public trust and legitimacy of representative democracy can be mitigated by integrating deliberative innovations into the broader political system (Curato and Böker Reference Curato and Böker2016). Specifically, involving citizens in deliberative forums such as minipublics, where they engage in informed and respectful discussion on policy issues with fellow citizens and political actors, is expected to enhance the legitimacy of democratic governance (Fishkin Reference Fishkin2009; Dryzek et al. Reference Dryzek, Bächtiger, Chambers, Cohen, Druckman, Felicetti, Fishkin, Farrell, Fung, Gutmann, Landemore, Mansbridge, Marien, Neblo, Niemeyer, Setälä, Slothuus, Suiter, Thompson and Warren2019).
Claims about the legitimising potential of deliberative democracy are commonly based on a normative model of procedural legitimacy that stresses the importance of direct citizen involvement – that is, going beyond elite-driven decision-making to involve ordinary citizens and their expression of opinions and needs in the development of political outcomes (Dryzek Reference Dryzek2010; Caluwaerts and Reuchamps Reference Caluwaerts and Reuchamps2015). Yet, the question of what generates public perceptions of legitimacy in democratic governance is an empirical one. Do populations believe that deliberative democracy processes such as deliberative minipublics (DMPs) would strengthen the legitimacy of existing representative institutions and enhance their democratic functions by enabling a more genuine fulfilment of popular sovereignty? Relatedly, existing literature on the effects of deliberative innovations predominantly focuses on direct benefits for participants,Footnote 1 rather than demonstrating the diffuse effects of deliberative processes on the wider citizenry (Werner and Marien Reference Werner and Marien2022; Germann et al. Reference Germann, Marien and Muradova2024). Amid growing calls to institutionalise deliberative minipublics (DMPs) not as replacements for legislative empowerment but as ‘auxiliary supplements to representative institutions’ (Lafont and Urbinati Reference Lafont and Urbinati2024: 233), there remains an outstanding need to empirically evaluate their effects on public legitimacy perceptions and their macro-level implications for democratic health (Warren Reference Warren2025).
In this article, we examine whether adding citizen deliberative assemblies to the decision-making processes of elected legislatures can increase democratic legitimacy among the wider public, and if so, which specific features amplify that effect. The study draws on a scenario in which a citizens’ panel (a form of deliberative minipublic) is convened to inform a policy inquiry conducted by a multi-party committee of the Australian federal parliament. To our knowledge, no previous empirical research has examined the effects of deliberation in this specific parliamentary context.
An emerging body of experimental research suggests that deliberative processes can produce diffuse effects that have the potential to influence the attitudes and public opinion of non-participants (Ingham and Levin Reference Ingham and Levin2018). Within this literature, some scholars emphasise the role of institutional design, suggesting that design features such as an accurate descriptive representation of broader society (Paulis et al. Reference Paulis, Pilet, Rojon and Vittori2024; Germann Reference Germann2025) are important for the procedural legitimacy of deliberative assemblies, even for those not directly involved. Acceptance of decision outcomes and support for institutions are boosted by the fact that the voices of ‘ordinary people’ are being heard (Van Dijk and Lefevere Reference Van Dijk and Lefevere2023:138). Some scholars demonstrate the importance of the degree of ‘authorisation’ granted to deliberative bodies (Goldberg et al. Reference Goldberg, Lindell and Bächtiger2024). Whereas others contend that citizens are more likely to view a decision-making process as legitimate if the outcome is one with which they agree or which benefits their individual interests (Hibbing and Theiss-Morse Reference Hibbing, Theiss-Morse, Sullivan, Snyder and Sullivan2008; Esaiasson et al. Reference Esaiasson, Persson, Gilljam and Lindholm2019).
We contribute to and extend this existing literature by explicitly testing the effects of information cues about how participants feel treated by decision-makers. The deliberative norms of consequentiality (Dryzek Reference Dryzek2010; Felicetti Reference Felicetti2016) and listening (Scudder Reference Scudder2021) are both considered central to the democratic quality of deliberative processes (Mansbridge et al. Reference Mansbridge, Bohman, Chambers, Christiano, Fung, Parkinson, Thompson, Warren, Parkinson and Mansbridge2012: 12). Drawing on procedural justice theory (Lind and Tyler Reference Lind and Tyler1988), we assume that these cues operate as signals of procedural fairness, thereby shaping legitimacy perceptions. That is, citizens appreciate decision-makers who both listen to and fairly consider the input of ‘ordinary people’. We therefore empirically test whether legitimacy gains depend on these relational norms – whether diffuse legitimacy benefits are enhanced when people feel the input of their fellow citizens is (a) given fair consideration in the making of policies which affect them, and (b) genuinely listened to by political representatives. We further expect these effects to be most pronounced among individuals who strongly prefer a deliberative model of democracy, as it aligns with their normative orientations towards political decision-making.
We investigate this through a vignette experiment embedded in an online-access survey of 4,200 adult Australian citizens. Participants were randomly assigned to read or not read a news article about a hypothetical parliamentary committee tasked with making recommendations on national nutrition policy, thereby placing the survey respondents in the role of hypothetical non-participants.Footnote 2 Within this scenario, respondents were further randomly exposed to several conditions that varied key features of the process: whether the parliamentary committee receives input from a deliberative citizens’ panel; whether the committee takes on board the citizen panel’s advice; whether this uptake of the citizen panel’s recommendations is obligatory; and whether members of the citizens’ panel report feeling sincerely listened to by parliamentarians during the process. This experimental design enables us to assess whether and under what conditions information about the integration of a deliberative event into parliamentary decision-making has a diffuse effect on the wider public’s legitimacy perceptions.
The results of the study lend support to normative expectations that legislative decision-making processes, supplemented by deliberative assemblies, generate higher legitimacy perceptions compared to traditional representative practices (Setälä Reference Setälä2017; Warren Reference Warren2025). Even when exposure is limited to a low-intensity and indirect information cue (a short newspaper article), our findings confirm that learning that a parliamentary committee incorporated input from a process of citizen deliberation increases perceptions of procedural legitimacy. Although the baseline effect is substantively small, legitimacy beliefs rise more substantially when the informational cue indicates that participating citizens received fair consideration and felt genuinely listened to by the parliamentary committee. These findings confirm that consequentiality and listening are central to enhancing the democratic functions of existing electoral institutions in the eyes of the public. Moreover, the legitimating force of consequentiality and listening underscores the importance of adhering to deliberative communication norms in democratic governance, rather than relying solely on institutional design features. That is, the benefits of deliberative innovations occur not simply when citizens are given space to speak, but when their input is genuinely heard and taken seriously by decision-makers.
The legitimacy gains hold across the range of democracy conceptions respondents endorse – direct, deliberative, technocratic, and trustee. Although this finding was unexpected, it suggests that embedding citizen deliberation within representative decision-making can generate legitimacy benefits that extend beyond those already predisposed to deliberative models. As such, rather than positioning these findings on one side or the other of the familiar dichotomy between electoral versusa deliberative models of democracy, we argue that deliberative processes complement and reinforce the perceived legitimacy of existing legislative institutions (Warren Reference Warren2017). Institutionalising minipublics as auxiliaries to electoral-representative bodies may therefore help bridge the gap between citizens and political decision-making, mitigate disaffection, and contribute – albeit modestly – to addressing contemporary democratic malaise (Suiter and Deligiaouri Reference Suiter and Deligiaouri2023; Lafont and Urbinati Reference Lafont and Urbinati2024).
The article proceeds in the following four parts. We begin by reviewing existing scholarship on the formation of legitimacy beliefs, situating this literature in relation to empirical research on democratic innovations and setting out our theoretical expectations of diffuse effects. This is followed by an account of our research design and data. Next, we present and analyse the empirical results. We conclude by reflecting on the broader implications of our results and identifying directions for future research.
Democratic legitimacy and deliberative innovations
Legitimacy, defined as ‘the belief that authorities, institutions, and social arrangements are appropriate, proper and just’ (Tyler Reference Tyler2006: 376), is fundamental to democracy. It ensures that people ‘accept and abide by laws even if such laws do not benefit their individual interests, because they were passed according to legitimate procedures’ (Van Ham et al. Reference Van Ham, Thomassen, Aarts and Andeweg2017: 3). As a critical component of political support, legitimacy has long-term and diffuse consequences for the stability of democratic systems (Claassen Reference Claassen2020). There is a well-established expectation that the institutional arrangements within a democratic system significantly shape public perceptions of democratic legitimacy. Procedures deemed to be fair cultivate what Clayton et al. (Reference Clayton, O’Brien and Piscopo2019: 116) describe as ‘a reservoir of goodwill’, enabling institutions to retain public trust even in the face of unpopular or occasionally flawed decisions.
Recent scholarship contends that, alongside the structural drivers of declining democratic satisfaction,Footnote 3 perceived deficiencies in the institutions and practices of representative politics are fuelling public disillusionment and eroding democratic legitimacy. Election-based representative democracy ‘is being criticised for not being inclusive, responsive or accountable enough’, observes Nicolaïdis (Reference Nicolaïdis2024: 6), ‘and, at the same time, for being too responsive to short-term popular preferences’. In response to these tensions, scholars and practitioners have advanced institutional reforms designed to expand direct citizen involvement in political decision-making, including proposals to integrate deliberative forums in representative systems.
Theory holds that the introduction of deliberative innovations can positively shape the public’s outlook on politics and society, including people’s perceptions of democratic legitimacy (Gastil Reference Gastil2021). By enhancing meaningful public participation in decision-making and constructing ‘new kinds of representative bridges to decision-makers’ (Warren Reference Warren2025: 9), deliberative innovations have the potential to address the underlying drivers of discontent with electoral democracy. When citizens perceive a greater connection to, and responsiveness of, political processes, their level of political support is expected to increase.
The most well-known deliberative democratic innovation is the deliberative minipublic (DMP). DMPs are defined as ‘carefully designed forums where a representative subset of the wider population come together to engage in open, inclusive, informed, and consequential discussion on one or more issues’ (Curato and Farrell Reference Curato and Farrell2021: 3). Though not a panacea for the democratic challenges of our time, DMPs are widely regarded as a promising intervention to address legitimacy deficits within representative democracies (Kuyper and Wolkenstein Reference Kuyper and Wolkenstein2019). Yet, proposals for DMPs vary, with Lafont and Urbinati (Reference Lafont and Urbinati2024: 3) identifying a spectrum from those advocating for the substitution of electoral democracy with empowered and autonomous deliberative assemblies, to those imagining these innovations as supplementary to current institutions. Whereas the former involves conferring authority from legislative assemblies to assemblies of randomly selected citizens, the latter group of proposals aims to empower the public to have greater participation and influence in the political process without undermining the power of elected assemblies or of citizen voting. Empirical research to date indicates that although some express a degree of openness toward empowered and autonomous minipublics, most citizens prefer minipublics to operate in ways that complement, rather than replace, existing representative institutions (Goldberg et al. Reference Goldberg, Lindell and Bächtiger2024).
One such minimally radical institutional proposal concerns the functioning of parliamentary or legislative committees. Although committees are formally designed to operate as deliberative forums that incorporate public input into policy development, in practice they rarely realise this aspiration. Various mechanisms have been proposed to strengthen the integration of public voices into committee processes, such as community forums or stakeholder roundtables, but a more structured approach would involve formally linking a DMP to a parliamentary committee, enabling the minipublic’s deliberations to inform and enrich the committee’s work (Hendriks and Kay Reference Hendriks and Kay2019). Despite their participatory capacity, democratic innovations and deliberative democracy research have thus far paid limited attention to these spaces.
Regardless of the specific institutional design or level of authority envisioned, DMPs can only meaningfully respond to public dissatisfaction with electoral-based democratic institutions if their influence extends beyond the individuals who directly participate. That is, for DMPs to fulfil the legitimacy-enhancing role attributed to them in the normative literature, they must generate broader, diffuse effects that influence the attitudes and perceptions of the public at large (Knobloch et al. Reference Knobloch, Barthel and Gastil2020).Footnote 4
Theorising diffuse effects of deliberative mini publics
Deliberative engagement processes typically have limited public visibility (Curato and Böker Reference Curato and Böker2016) and occur infrequently; thus, they remain largely outside citizens’ existing experiences and understandings of democratic participation. This raises serious questions about the mechanisms through which their effects could be ‘scaled up’ or transmitted to non-participating publics and the wider democratic system. It also explains why some scholars remain sceptical about the potential for deliberative innovations to produce macro-level legitimacy gains.
Insights from established research on information cues, heuristics, and endorsement effects (see Mutz Reference Mutz2011; Dewan et al. Reference Dewan, Humphreys and Rubenson2014), however, suggest that deliberative processes can act as a signalling mechanism of public input into government policy decisions. There are at least three reasons to expect such a signal to bolster public perceptions of legitimacy. First, deliberative processes may signal instrumental control by allowing individuals to express their views or challenge evidence. Research on procedural justice suggests legitimacy is strengthened when representative processes offer opportunities for influence and are publicly justified (Grimes Reference Grimes2006; Arnesen and Peters Reference Arnesen and Peters2018). Because deliberative innovations are explicitly designed to provide these forms of control and justification, they are likely to increase acceptance of outcomes, even when the substantive decision is contested. Second, deliberative mechanisms can serve a relational signalling function by emphasising respectful and inclusive procedures. Such processes communicate that individuals are recognised and valued members of the political community, which can strengthen identification with democratic institutions and the decisions they make (Lind and Tyler Reference Lind and Tyler1988). Third, in settings marked by complexity or low transparency, deliberation may reduce uncertainty by activating a fairness heuristic. Citizens may use perceptions of fair procedures as cues to infer trustworthiness and impartiality, thereby enhancing legitimacy even in the absence of substantive agreement (Van den Bos and Lind Reference Van den Bos, Lind and Zanna2002).
Collectively, these mechanisms offer strong theoretical grounds to expect that embedding deliberative elements within representative systems can enhance perceived legitimacy by improving both the procedural quality of decision-making and the interpretative frameworks citizens use to evaluate political processes. Thus, it is reasonable to expect that, when citizens are aware of deliberative processes, they will perceive them as more legitimate than non-deliberative alternatives.
A growing body of empirical evidence supports this optimism that deliberative forums can, in fact, mitigate dissatisfaction with politics in the wider public. Boulianne (Reference Boulianne2018) finds that even one-off mini publics can strengthen perceptions of political legitimacy by ‘increasing political efficacy and possibly building trust between citizens and government’ (Reference Boulianne2018: 134), while recent experimental studies suggest that integrating DMPs into representative institutions can improve public perceptions of government trustworthiness, responsiveness, and willingness to listen, especially when contrasted with the less participatory legacy institutions of representative democracy (Jacobs and Kaufmann Reference Jacobs and Kaufmann2021; Germann Reference Germann2025).
Still, the broader signalling effects of deliberative innovations crucially hinge on whether information about these processes reaches audiences beyond the participants themselves. That is, public perceptions of deliberative processes are shaped not only by the institutional design and deliberative quality of a process itself, but also by how it is communicated and framed. Information may diffuse from participants to non-participants through interpersonal communication or through mediated channels. Lazer et al. (Reference Lazer, Sokhey, Neblo, Esterling and Kennedy2015) find that structured deliberative events stimulate broader political discussions through participants’ social networks. However, for most people, mass media remains the primary source of political information (Strömbäck and Shehata Reference Strömbäck and Shehata2010), and public awareness is strongly influenced by the agenda-setting power of the news media (Luo et al. Reference Luo, Burley, Moe and Sui2019). It follows that non-participating publics are far more likely to encounter DMPs (if at all) through media coverage rather than direct interpersonal contact. Accordingly, it is crucial to examine how media-based information cues about citizen deliberation shape legitimacy perceptions among the wider public (see Fishkin Reference Fishkin2009).
Combining these theories of the social psychological foundations of procedural fairness with agenda-setting research and existing studies on the diffuse effects of DMPs embedded into representative institutions, we hypothesise that, all else being equal, receiving information about a political process involving deliberative citizen engagement increases perceptions of procedural legitimacy, compared to the status quo. [H1]
Furthermore, we expect that individuals’ evaluations of the legitimacy of a deliberative process will be closely tied to information about its procedural characteristics. For example, existing research suggests that the perceived legitimacy of DMPs depends on the extent to which the profile of participants is considered to ‘fairly represent’ the wider public (Paulis et al. Reference Paulis, Pilet, Rojon and Vittori2024). When DMPs lack descriptive representation, they fail to confer legitimacy benefits (Paulis et al. Reference Paulis, Pilet, Rojon and Vittori2024; Germann Reference Germann2025). Clayton et al. (Reference Clayton, O’Brien and Piscopo2019) find that gender-balanced decision-making bodies significantly enhance perceptions of procedural legitimacy, reinforcing the view that descriptive representation signals inclusiveness and strengthens democratic legitimacy.
Rather than concentrating on structural design features such as representativeness, random selection, deliberation length, or use of expertise, we examine whether legitimacy perceptions are shaped by consequentiality (the extent to which the deliberative process influences political decision-making) and listening (whether participants perceive themselves as being heard by decision-makers). These two relational dimensions are widely seen as being ‘at the heart of meaningful democratic deliberation’ (Scudder Reference Scudder2021: 119). By foregrounding these deliberative norms, we move beyond institutional design to consider how such relational qualities of deliberative processes are perceived and interpreted by the broader public.
First, consequentiality, understood as when a process has a tangible impact (whether direct or indirect) on decisions or actions external to the deliberation itself (Dryzek Reference Dryzek2010; Felicetti Reference Felicetti2016), is an important criterion for evaluating deliberative processes. Deliberative theory assumes that people attach importance to the way their input, or the input of their fellow citizens, influences the political process and that they want their opinions and concerns to be taken seriously (Font and Blanco Reference Font and Blanco2007). Research suggests that failing to act on deliberative outcomes can diminish the positive effects on direct participants, yet it remains unclear to what extent failure to act influences legitimacy perceptions among the broader public. Extrapolating previous findings on the effects of consequentiality on participants of deliberative processes, we expect that receiving information about a deliberative process involving consequentiality increases perceptions of procedural legitimacy, compared to receiving information about a deliberative process that does not involve consequentiality [H2a]. To explore this, we consider the deliberative norm of consequentiality as being met when decision-makers either voluntarily give ‘fair consideration’ (Scudder Reference Scudder2020) to the input citizens share in deliberation (‘uptake’) or when they are obliged to adopt the outcomes of citizen deliberation (‘obligation’).
We are also interested in the extent to which legitimacy beliefs are influenced by the quality of listening of those in power. As Bickford (Reference Bickford1996) argues, ‘being listened to’ is fundamentally an interpersonal and relational experience. However, in the context of diffuse effects, it remains unclear whether the intrinsic value of this experience can be transmitted second-hand. Do perceptions of legitimacy increase when citizens are exposed to accounts in which their fellow citizens report feeling genuinely heard? Or put another way, can the effect of ‘sincere listening’ be conveyed through another’s experience?Footnote 5 Similar to the second-hand effects of consequentiality, we expect that receiving information about a deliberative process involving sincere listening increases perceptions of procedural legitimacy, compared to receiving information about a deliberative process that does not involve sincere listening [H2b].Footnote 6
Beyond the institutional features or relational qualities of a deliberative process, individual-level attributes such as prior knowledge, democratic preferences, and baseline political trust are also likely to shape how citizens respond to information about citizen deliberation and whether such information enhances their legitimacy beliefs. Existing evidence indicates that support for DMPs is conditional on experience and familiarity (Goldberg, Lindell, and Bächtiger Reference Goldberg, Lindell and Bächtiger2024: 1394). Even without direct participation, informational exposure to deliberative processes may increase familiarity, enhance conceptual fluency, and signal institutional legitimacy through concrete examples, consistent with the mere exposure effect (Zajonc Reference Zajonc1968). We extend this reasoning by positing that individuals’ assessments of the fairness and effectiveness of democratic processes are also filtered through their underlying conception of how democracy should function. Individuals’ understanding of democracy, including their procedural preferences and preexisting trust in deliberative or other participatory remedies, may condition their evaluations of public deliberation (Lu and Chu Reference Lu and Chu2021; Geissel Reference Geissel2023). An essentially contested concept, democracy carries different meanings for different actors (Whitehead Reference Whitehead2002: 14). Taking this contestability seriously requires recognising that conceptions of democracy and understandings of political representation are not limited to the liberal representative model but encompass other models such as participatory, radical, deliberative, and direct democracy (Kurki Reference Kurki2010: 372). Each model or conception of democracy interprets society and the power dynamics within it differently, leading to variations in the institutional forms considered central to ‘democracy’ and the values deemed most essential for legitimate democratic governance. For example, electoral-representative democracy prioritises voting, elections, and political parties, whereas deliberative democracy emphasises mechanisms of ‘voice’ beyond elections. Ignoring these different normative expectations, alongside citizens’ beliefs about what democracy should ideally accomplish (Ulbricht Reference Ulbricht2018: 1399), risks misdiagnosing the sources of democratic dissatisfaction and misjudging the efficacy of potential remedies.
In this way, we theorise that democratic orientations operate as a demand-side moderator of informational effects. Accordingly, we test whether antecedent democratic conceptions condition the legitimacy-enhancing impact of deliberation. We expect an interaction effect: holding all else constant, receiving information about a political process involving deliberative citizen engagement increases perceptions of procedural legitimacy more strongly for those who hold deliberative conceptions of democracy [H3].
Methods
We assessed the merit of these expectations with an experimental research design that allowed us to test whether a representative decision-making process embedded with citizen deliberation performs better in generating democratic legitimacy relative to the status quo process without deliberation. Using a representative sample of Australian adult citizens, we estimated the causal impacts of deliberation on public perceptions of legitimacy, which we operationalised as government trustworthiness, responsiveness, and willingness to listen. This section describes our hypothetical scenario, the experimental design, and our measures.
The case
Australia has relatively extensive use of deliberative processes at the subnational level, including the commissioning of deliberative processes by public authorities and statutory agencies as part of their strategies for citizen engagement (Webb et al. Reference Webb, Parry, Ercan, Asenbaum and McCaul2026). A landmark development in 2020 saw the Victorian state government introduce legislation mandating the use of ‘deliberative engagement practices’ by local government authorities (Government of Victoria 2020). At the national level, however, Australia has no permanent or ad hoc processes established to date, meaning it trails behind countries such as France and the United Kingdom, which have utilised DMPs at the national level (OECD 2023). Recent scholarship suggests that this pattern may reflect institutional inertia or resistance at the federal level, rather than a lack of public appetite for a stronger citizen voice in national politics (Webb et al. Reference Webb, Grömping and Foley2025, Reference Webb, Parry, Ercan, Asenbaum and McCaul2026).
The scenario
The scenario of the study is the convening of a citizens’ panel (a type of DMP) to inform an inquiry held by a multi-party parliamentary committee situated within the Australian Federal Parliament. In our hypothetical case scenario, the parliamentary committee is tasked with making recommendations on national nutrition policy. Specifically, we focus on the policy proposal to introduce a tax on sugar-sweetened beverages (‘sugar tax’). The issue represents an ongoing policy debate in Australia, characterised by active lobbying from business and professional interest groups on either side and an absence of strong partisan alignment (Allen and Allen Reference Allen and Allen2020; Anaf et al. Reference Anaf, Fisher, Handsley, Baum and Friel2021; Cullerton et al. Reference Cullerton, Demeshko and Waller2024). Public opinion on the sugar tax remains divided, with support ranging between one-third and two-thirds of Australians (Cullerton et al. Reference Cullerton, Baker, Adsett and Lee2021; Pettigrew et al. Reference Pettigrew, Booth, Dunford, Scapin, Webster, Wu, Tian, Praveen and Sacks2023). This makes the issue an ideal test case for our research question, as deliberation is generally recommended for issues that are contested but not deeply polarised (Bächtiger et al. Reference Bächtiger, Spörndli, Steenbergen, Steiner and Rosenberg2007).
Experimental design
The fieldwork for this survey was conducted between 17 September and 9 October 2024, with a sample of N = 4,225 Australian citizens aged 18 and older recruited via an online-access panel. We adopted a pragmatic approach to sample planning, with our budget-constrained sample size achieving power to detect ‘very small’ effects of approximately f = 0.057 (d = 0.114), which aligned with our a priori expectations.Footnote 7 Although our pool of respondents is drawn from a non-probability sample, it is suitable for the purposes of our study, which is focused on the causal effects of treatments, rather than population-level estimates (Baker et al. Reference Baker, Brick, Bates, Battaglia, Couper, Dever, Gile and Tourangeau2013; Coppock et al. Reference Coppock, Leeper and Mullinix2018). The survey used introductory screening questions to fill quotas based on age, gender, state of residence, and educational attainment.Footnote 8 Respondents were then asked sociodemographic questions and asked about political attitudes, including interest in politics, left-right ideological orientations, self-reported partisan identification, political trust, and internal efficacy, among others. Importantly, respondents were also asked a pre-treatment question about their support for a sugar tax.Footnote 9
After these questions, respondents were randomly assigned in equal numbers to one of six experimental conditions: one control group (‘status quo’) and five treatment groups, as shown in Figure 1.Footnote 10 Each group was shown a vignette of a fictitious news article describing the scenario of a parliamentary committee process about the sugar tax. The headline and parts of the article body were experimentally manipulated to vary aspects of the scenario across four factors, each with two levels: (1) Mode: Did the committee receive input from a deliberative citizens’ panel [yes/no]? (2) Uptake: Did the committee take on board the citizen panel’s policy recommendation [yes/no]? (3) Obligation: Was the committee obligated to follow the citizen panel’s recommendation [yes/no]? And (4) Listening: Did the members of the citizens’ panel feel sincerely listened to [yes/no]? All six vignettes are included in online Appendix C of the supplementary materials.Footnote 11 Some treatment conditions are logically nested (e.g., listening and obligation within uptake and uptake within mode). Our design is therefore an incomplete factorial experiment, akin to a multi-arm parallel-group experiment in medical research to test key contrasts (Juszczak et al. Reference Juszczak, Altman, Hopewell and Schulz2019).
Experimental conditions and flow.
Figure 1. Long description
A flowchart illustrating the experimental conditions and flow for studying the effects of deliberative minipublics on public legitimacy perceptions. The process begins with an introduction and consent, followed by the collection of socio-demographics, attitudes, and policy preferences. Participants are then randomized into different modes. If they receive information about the Parliamentary Committee, they proceed to the deliberation uptake step. Depending on their uptake, they are categorized into different groups based on obligation and listening. The flowchart includes decision points where participants either proceed to listening or not, leading to various outcomes such as deliberation uptake with obligation and listening, deliberation uptake with obligation, deliberation ignored, or status quo. The final outcomes are analyzed for procedural legitimacy and a manipulation check, followed by a debrief.
Through a manipulation check conducted after the measurement of the outcome variables, we confirmed that respondents understood the treatments as intended.Footnote 12 Interestingly, respondents in all treatment arms other than the ‘ignored’ arm perceived that the committee sincerely listened to the citizens’ panel. This suggests that respondents adopt a broader understanding of listening than we anticipated when designing the experiment. In their interpretation, listening appears to be inferred primarily from the committee’s uptake of the recommendations, regardless of whether members of the hypothetical deliberative forum reported feeling ‘sincerely listened to’. We return to this point in the analysis.
Measures
To measure the dependent variable, perceptions of procedural legitimacy, respondents answered a battery of four questions assessing their level of agreement on a 5-point Likert scale from ‘Strongly agree’ to ‘Strongly disagree’. The statements included ‘The committee can be trusted to make the right recommendation for all Australians’; ‘The committee process was fair’; ‘The committee made its recommendation after careful consideration’; and ‘I support the committee’s recommendation.’ The order of the items was randomised. We conducted a confirmatory factor analysis to model the dependent variable. Results indicate that all four items represent a single latent factor; however, the initial model fit is unsatisfactory (Table F1 in the supplementary materials). Based on the modification indices, we therefore allowed the residuals of the items ‘The committee can be trusted to make the right recommendation for all Australians’ and ‘I support the committee’s recommendation’ to correlate. This adjustment is theoretically justifiable, and we expect that responses to both questions may be influenced by substantive agreement with the policy recommendation itself. The modified model demonstrates a satisfactory fit with the data, and we therefore use factor scores based on this model for all subsequent analyses.
To examine treatment effect heterogeneity for hypothesis H3, we used a pre-treatment measure for conceptions of democracy. Our measure builds on existing operationalisations, such as those proposed by Landwehr and Steiner (Reference Landwehr and Steiner2017) and Rapeli and von Schoultz (Reference Rapeli and von Schoultz2023), but requires respondents to make a ranked choice. Specifically, we asked respondents to indicate their preferred approach to political decision-making in a democracy by ranking the following options from most to least desirable: ‘Develop policies in close dialogue with citizens and affected groups’ (deliberative conception); ‘Let experts in different areas make decisions’ (technocratic); ‘Let elected politicians make decisions’ (trustee); and ‘Put important questions to referendums’ (direct). The options were randomised to reduce order effects. The nominal variable ‘conception of democracy’ is coded as each respondent’s first-ranked option (deliberative, technocratic, trustee, or direct).Footnote 13
In addition, we measured pre-treatment sociodemographic and attitudinal covariates, including age (in years), gender (female = yes/no), education (university = yes/no), employment (full-time employment = yes/no), economic situation (able to save = yes/no), and political trust. Political trust was measured pre-treatment as a 5-point Likert scale from ‘Strongly distrust’ to ‘Strongly trust’, based on the prompt: ‘How much do you trust or distrust the Australian Government? When answering this question, please consider the Australian Government in general, rather than the current government or any specific political party.’ We used all covariates to adjust our regression estimates as a means of improving precision (Lin Reference Lin2013; Wager et al. Reference Wager, Du, Taylor and Tibshirani2016).
Finally, an important objection to our theoretical expectations is that procedural justice has limited impact and that citizens’ acceptance of political authority and decisions is primarily driven by outcome favourability (Esaiasson et al. Reference Esaiasson, Persson, Gilljam and Lindholm2019). If this holds true, then procedural innovations, such as introducing deliberative elements into existing representative mechanisms, would have little effect on perceptions of legitimacy when people disagree with the outcome. To account for this possibility, we also adjusted our estimates for respondents’ policy preferences regarding a sugar tax. These were measured pre-treatment on a 5-point Likert scale from ‘Strongly disagree’ to ‘Strongly agree’ in response to the statement: ‘Some people say that the government should introduce a tax on sugar-sweetened beverages that would raise the cost of a 375 ml can of soft drink by 16 cents. Supporters claim the tax will help lower obesity and chronic health conditions in Australia. Other people are opposed to the tax, claiming it will disproportionately affect Australians in low-income households. To what extent do you agree or disagree with the suggestion of introducing a tax on sugar-sweetened beverages?’
Analysis
To systematically assess our theoretical expectations, we structured our analysis in five stages. First, we pooled all treatment conditions and compared this combined group with the ‘status quo’ group to estimate the average treatment effect of information about a citizen’s deliberation on perceptions of legitimacy (H1). Second, we disaggregated the treatment conditions to examine whether the presence of specific deliberative norms (namely consequentiality and sincere listening) produced differential effects on perceived legitimacy (H2a, H2b). Third, we explored whether individual democratic conceptions shape responsiveness to deliberation (H3). Fourth, we conducted additional post-hoc analyses by interacting our treatments with indicators of trust and satisfaction. Finally, we examined the robustness of our results to different model specifications, measurement approaches, and sample exclusions.
Taken together, our findings suggest that exposure to information about deliberative engagement within a parliamentary committee yields a modest positive effect on perceived legitimacy. Importantly, these legitimacy gains are amplified when the deliberative norm of sincere listening is present, regardless of whether the committee is obligated to follow the citizen’s recommendations. This highlights the value placed by the public on being heard, or on perceiving that others are genuinely heard, in their evaluation of deliberative democratic innovations. Notably, the positive effect of deliberative engagement holds across subsets of the population that hold different conceptions of democracy.
The effect of deliberation on perceived legitimacy
To evaluate hypothesis H1, which posits that receiving information about a political process involving deliberative citizen engagement increases perceptions of procedural legitimacy, we examined the effect of the pooled treatments on the dependent variable. Our primary estimand of interest was the average causal effect of information about a parliamentary committee process receiving input from a deliberative citizens’ panel with varying attributes (‘uptake’ = yes/no; ‘obligation’ = yes/no; ‘listening’ = yes/no) relative to a comparison group who received information about the existing parliamentary committee process, that is, a public submissions process without citizens’ deliberation. We employed covariate-adjusted linear regression of our dependent variable on the pooled treatment conditions, with the ‘status quo’ serving as the reference category. To facilitate comparisons, all estimates were standardised using Glass’s Δ, which scales outcomes by the standard deviation in the control condition (Glass Reference Glass1976). Figure 2 displays the estimated effects with 90% and 95% confidence intervals (CIs), alongside a margin of equivalence (MOE) set at ±0.2 standard deviation units. An effect is considered distinguishable from zero if the 95% CI excludes zero. Conversely, an effect is deemed substantively equivalent to zero (i.e., ‘negligible’ in size) when the 90% CI lies entirely within the MOE. Only if a 95% CI excludes zero and the 90% CI overlaps the equivalence bounds is the estimate interpreted as both statistically distinct from zero and non-equivalent to a ‘negligible’ effect (Lakens Reference Lakens2017). All models are fully reported in the supplementary materials, Appendix H.
Effect of information about deliberative citizen engagement on legitimacy perceptions.
Note: Covariate-adjusted OLS estimates in Glass’s Δ standard units compared to the reference group ‘Status quo’. Thick horizontal lines denote 90% CIs, and thin lines denote 95% CIs. Dotted vertical lines denote a margin of equivalence (MOE) of ±0.20 standard units. Black estimate and CIs = distinct from zero and non-equivalent effect. The mid-grey estimate and CIs = distinct from zero but have equivalent (negligible) effect. Light grey estimate and CIs = indistinct and equivalent (negligible) effect. Adjusted for covariates age, female, university degree, full-time employment, economic situation, regional/remote location, conception of democracy, political trust, policy preference for sugar tax, partisan ID, political interest, internal efficacy, satisfaction with democracy, and affective polarisation.
First, looking at the pooled treatments (topmost estimate in Figure 2), we find that exposure to low-intensity information cues about a parliamentary committee receiving input from a deliberative citizens’ panel causes a small but statistically significant increase in perceived procedural legitimacy relative to the status quo control condition (b = 0.064, SE = 0.029, p = 0.029). Expressed as a standardised mean difference, this corresponds to Glass’s Δ = 0.079, with a 95% confidence interval of [0.027, 0.130]. Based on our equivalence margin of interest of ±0.2 standard deviation units, the 90% confidence interval for Δ lies entirely within these bounds, indicating the effect is substantively negligible despite being statistically significant (Lakens Reference Lakens2017).
The finding supports hypothesis H1, suggesting that information about a process of citizen deliberation embedded into an existing parliamentary process does have positive effects on the broader public’s perceptions of legitimacy. Although the pooled effect falls below the conventional social science threshold for ‘small’ effects (Cohen Reference Cohen1988), such modest magnitudes can nevertheless be substantively meaningful when interventions are light-touch and outcomes are difficult to influence (Abelson Reference Abelson2026: 46). In our case, respondents are exposed only to a brief news vignette, and legitimacy beliefs are typically slow-moving. The very small effect we detect could also reflect respondents’ awareness that citizens’ deliberative input risks being ignored by politicians. This hunch was confirmed when we excluded the ‘ignored’ condition from our pooled estimate, which resulted in a slightly higher effect size of Δ = 0.118 [95% CI: 0.067, 0.170], a result we explore further in the next section.Footnote 14 Moreover, effects within a margin of equivalence may still matter in practice, particularly when interventions are scalable or cumulative over repeated exposure (Lakens Reference Lakens2017). This is applicable here, as the embedding of deliberative procedures in parliamentary committees would result in repeated news items on the process, thus potentially building up legitimacy over time.
The effects of consequentiality and listening on perceived legitimacy
Next, to provide deeper insights into how information about citizen deliberation affects perceptions of procedural legitimacy, we extended our analysis to specific relational features of the deliberative process. Returning to hypotheses H2a and H2b, we examined whether information about the deliberative norms of consequentiality and sincere listening increased legitimacy perceptions. To this end, we analysed the treatment conditions individually, rather than pooled. While Figure 2 above shows regression estimates, here we first estimated a covariate-adjusted omnibus ANOVA using Type III sums of squares, which tested each factor conditional on all other predictors in the model. This analysis revealed a statistically significant effect of experimental condition on perceived procedural legitimacy (F(5, 3572) = 7.989, p < 0.001), warranting further post-hoc comparisons. Conditional on exposure to a deliberative process, we then assessed whether the presence of consequentiality or sincere listening increased legitimacy. Accordingly, we conducted multiple comparisons of Glass’s Δ-standardised group means, evaluating directional hypotheses using one-sided tests and applying Bonferroni correction to account for multiple comparisons. The results are presented in Table 1.
Bonferroni-adjusted post-hoc contrasts for H2a and H2b
Notes: Entries report Glass’s Δ based on covariate-adjusted estimated marginal means from weighted OLS model M2 in Table H1. One-sided p-values, Bonferroni-adjusted to account for multiple comparisons.
Recall that consequentiality is operationalised as the presence of uptake, obligation, or both. To unpack its effects, we examined three contrasts for H2a: (i) uptake relative to ignored deliberation (X3 vs. X2), capturing the distinct effect of uptake; (ii) obligation conditional on uptake (X5 vs. X3), capturing the distinct effect of obligation; and (iii) the combined effect of uptake and obligation relative to ignored deliberation (X5 vs. X2). The results indicate that consequentiality caused increased perceived legitimacy, primarily through uptake. Uptake alone produced a small but statistically significant increase in legitimacy relative to ignored deliberation (Glass’s Δ = 0.117, p = 0.026), while the combined presence of uptake and obligation yielded a larger effect (Δ = 0.166, p < 0.001). By contrast, obligation conditional on uptake did not exert an independent effect (Δ = 0.049, p = 0.702). Substantively, these findings suggest that consequentiality operates mainly via signals that deliberative inputs are taken seriously, with formal obligation adding little additional legitimacy on its own.
We then looked at an analogous set of contrasts to assess the effects of sincere listening (H2b), comparing ‘deliberation with listening’ to otherwise identical conditions without listening, both in the absence of obligation (X4 vs. X3) and in its presence (X6 vs. X5). Information that the participants of the minipublic felt sincerely listened to increased legitimacy perceptions (Δ = 0.130, p = 0.012), indicating a clear positive effect of listening when deliberation is taken up but non-binding. By contrast, adding listening to an already obligatory process did not produce a further increase (Δ = 0.048, p = 0.733). Together, these results show that sincere listening is an important driver of perceived procedural legitimacy, particularly when uptake of the deliberation outcomes is voluntary, while its marginal contribution diminishes once uptake of outcomes is institutionally mandated.
Our interpretation is that, for non-participating publics evaluating deliberative processes within representative institutions, the perceived quality of listening matters more for legitimacy than whether the uptake of the deliberation outcome is institutionally guaranteed. Nevertheless, consequentiality remains a necessary condition, as evidenced by the clear difference between the ‘uptake’ and ‘ignored’ conditions. Importantly, these two dimensions appear partially fused in public perception: our manipulation check shows that all respondents except those in the ‘ignored’ condition felt the citizens’ panel had been listened to,Footnote 15 suggesting a broad understanding of listening as ‘taking on board citizen input’. At the same time, explicit cues of high-quality listening generate an additional positive effect. Indeed, as shown in Figure 2, deliberation characterised by sincere listening produced the largest effect estimate even in the absence of formal obligation.
Regarding hypotheses H1 and H2a/b, our design constituted a demanding test of our expectations in several respects. First, we deliberately employed a realistic low-intensity information cue embedded in news reporting. While the resulting effect sizes are small, they are broadly consistent with meta-analyses of political news framing effects, which typically find very small to small attitudinal effects following exposure to competing frames (Amsalem and Zoizner Reference Amsalem and Zoizner2022). At the same time, our effect size estimates of 0.08–0.18 standard deviations (SDs) are substantially smaller (less than a third) than those reported in studies of higher-intensity legitimacy interventions, such as information about the representativeness of minipublics (Germann Reference Germann2025) or the gender balance of decision-making bodies (Clayton et al. Reference Clayton, O’Brien and Piscopo2019). Second, parliamentary committees, although a crucial democratic instrument, are relatively unfamiliar to the public, who are therefore unlikely to have well-formed expectations about a ‘status quo’ committee process without deliberation (Marsh and Halpin Reference Marsh, Halpin, Head and Crowley2015). Third, our preregistered analysis plan for H1 imposed a stringent criterion, requiring all treatment arms, including ignored deliberation, to yield significant positive and non-equivalent effects. Together, these features help explain why H1 was not formally confirmed. Nevertheless, the overall pattern of results suggests that deliberative citizen engagement does confer diffuse legitimacy benefits, provided that recommendations are taken up and participants are perceived as having been sincerely heard.
The influence of respondents’ democracy conceptions on legitimacy effects
We next test hypothesis H3, which posits that information about deliberative processes has stronger legitimacy-enhancing effects among individuals who hold a deliberative conception of democracy. In our sample, respondents’ first-ranked preferences for mode of political decision-making were distributed as follows: deliberative (47.6% of respondents), technocratic (22.8%), trustee (8.8%), and direct (20.9%) conceptions. To evaluate H3, we conducted a moderation analysis in which treatment assignment was interacted with democratic conception within a pooled, covariate-adjusted regression framework. We then derived marginal treatment effects (deliberation minus no deliberation) for each conception group and compared these effects using difference-in-differences contrasts. These comparisons were implemented via covariate-adjusted estimated marginal means with Bonferroni-corrected one-sided tests, with all effects reported as Glass’s Δ. The resulting interaction contrasts are reported in Table 2.Footnote 16
Pairwise comparisons of the pooled deliberation effect across conceptions of democracy
Notes: Glass’s Δ standardised interaction contrasts from covariate-adjusted, weighted OLS model M4 in Table H2. Values represent pairwise differences in the pooled deliberation effect (deliberation−no deliberation) across democratic conceptions. Two-sided 95% confidence intervals. One-sided p-values, Bonferroni-adjusted to account for multiple comparisons.
The moderation analysis provides no evidence that the legitimacy effect of deliberative engagement varies systematically by democratic conception. In the pooled interaction model, none of the interaction terms between deliberation and conception were significant. Consistent with this, Bonferroni-adjusted difference-in-differences comparisons (see Table 2) showed no meaningful contrasts between deliberative democrats and any other group. Relative to deliberative democrats, the pooled deliberation effect was slightly larger among technocrats (Δ = 0.057, 95% CI [−0.185, 0.299]), marginally smaller among direct democrats (Δ = 0.024, 95% CI [−0.225, 0.274]), and marginally larger among trustee democrats (Δ = −0.022, 95% CI [−0.366, 0.323]); however, all differences were small and statistically indistinguishable from zero.
Additional exploratory subgroup analyses showed the largest effect size – and the only one being statistically significant – among trustee democrats (b = 0.189, SE = 0.075, p = 0.012).Footnote 17 However, because such subgroup regressions assess whether the treatment effect differs from zero within groups rather than between them, and because the interaction terms are not significant (see above), these findings should be interpreted cautiously. At most, they suggest that trustee democrats may be particularly receptive to deliberative inputs. Finally, we explored whether individual process features (as in H2a and H2b) were moderated by democratic conception. However, none of the treatment-by-conception interaction terms were statistically significant.Footnote 18
In sum, these results do not support hypothesis H3. The legitimacy-enhancing effects of information about deliberative citizen engagement in representative institutions accrue to non-participants in the wider public across the board, not only to those that already hold a deliberative conception of democracy.
Legitimacy effects among distrusters and dissatisfied
We also examined whether other subgroups exhibit distinct responses to information about deliberation. Disenchanted and populist-minded citizens, for example, may place greater value on empowering citizen voices (Goldberg and Bächtiger Reference Goldberg and Bächtiger2023). Although we did not include a direct measure of populism, we explored this possibility by interacting our treatment conditions with institutional trust and democratic satisfaction. In the pooled model, institutional trust significantly moderated the deliberation effect (Assignment × Trust: b = −0.062, SE = 0.029, p = 0.035), whereas satisfaction did not (b = −0.036, SE = 0.032, p = 0.254).Footnote 19 Post-hoc contrasts confirmed this pattern: among low-trust respondents, deliberation increased perceived legitimacy by Glass’s Δ = 0.156 (95% CI [0.055, 0.256], p = 0.001), whereas the effect was essentially null among high-trust respondents (Δ = 0.003, 95% CI [−0.096, 0.103], p = 0.473). Further disaggregation of the treatments indicates that this responsiveness among distrusters is primarily driven by conditions emphasising the quality of listening (e.g., X4 vs. status quo: Δ = 0.232, 95% CI [0.041, 0.424], p = 0.003).Footnote 20 Substantively, these findings indicate that deliberative signals matter most for citizens who are otherwise sceptical of political institutions. More broadly, they suggest that visible practices of listening can bolster legitimacy among those least inclined to grant it. Deliberative citizen engagement thus emerges as a potential corrective to democratic disaffection, particularly among constituents characterised by institutional distrust.
Robustness
Our findings are robust across a range of alternative specifications.Footnote 21 Re-estimating M1 and M2 with different covariate sets yields consistent signs and comparable magnitudes, with increased precision once attitudinal controls are introduced. The ‘ignored’ condition is negative and statistically significant in models without covariates but attenuates after adjustment, while conditions involving listening and/or obligation remain stable in direction, magnitude and statistical significance across all specifications. Case-resampling bootstrap checks with 2,000 replications further confirm the stability of these coefficients. The results are also unchanged when replacing our latent outcome measure with an additive index or individual items, the latter revealing that the results were mainly driven by respondents’ perceptions of fairness and the perception that the committee made its recommendation after careful consideration. Finally, excluding speeders (fastest and slowest 5%) and straightliners left all substantive conclusions intact, with effect sizes typically increasing. Together, these checks suggest that our core findings are not artefacts of model specification or measurement choice but reflect a consistent pattern in how non-participating respondents evaluate deliberative citizen engagement within representative institutions.
Conclusion
This study sought to deepen our understanding of whether, and under what conditions, democratic innovations such as deliberative minipublics can generate broad, positive impacts on the political attitudes of the wider citizenry. Specifically, we asked whether existing representative institutions (here, parliamentary committees) are perceived as more trustworthy, fair, and effective in the eyes of the public when they incorporate citizen deliberation and which features of such deliberative engagement processes matter most in public evaluations. In moving beyond participant-centred assessments of deliberative forums, the study responds to recent calls to examine the effects of deliberative innovations within the broader context of declining confidence in legacy representative institutions (Warren Reference Warren2025) and complements ongoing efforts to institutionalise minipublics and incentivise elected officials to engage with their outputs (Setälä Reference Setälä2017: 857–859).
Using a nationally representative survey experiment of Australian adults, we embedded deliberative citizen engagement within a realistic parliamentary committee scenario and systematically varied key process features: whether citizen input was included through a citizens’ panel (mode), whether recommendations were taken up (uptake), whether uptake was mandatory (obligation), and whether participants reported feeling sincerely listened to (listening). Procedural legitimacy was measured through perceptions of trustworthiness, fairness, and support for the committee’s decision.
Three central findings emerge. First, exposure to information about citizen deliberation within representative institutions produces a modest but consistent increase in perceived procedural legitimacy relative to the status quo. Although these effects are small in magnitude, they are notable given the low-intensity nature of the intervention and the typically stable character of legitimacy beliefs. Importantly, they indicate that the incorporation of deliberative citizen engagement within existing representative institutions can generate legitimacy gains that spread beyond those who participate directly. Moreover, this finding suggests that the public’s criteria for evaluating legitimate democratic governance align with normative principles of legitimacy that emphasise the inclusion of community voices and interests in decision-making processes.
Second, the strongest legitimacy effects emerge when constituent deliberation is accompanied by explicit cues of sincere listening by the political actors involved. Consequentiality also matters, with legitimacy declining sharply when citizen input is ignored; however, formal empowerment of the deliberative body adds little value. Voluntary uptake of the citizen panel’s recommendations by the committee signals that the process was taken seriously by the politicians involved. For indirectly observing publics, listening and consequentiality appear partially fused: having constituents’ input taken seriously, as expressed in uptake of the recommendations, is equivalent to being ‘listened to’. Yet, high-quality listening provides an additional legitimacy boost, particularly in non-obligatory contexts. In sum, the perceived relational quality of engagement between constituents and decision-makers is pivotal in determining whether deliberative processes can mitigate declining confidence in democratic institutions. In the context of eroding public trust, merely creating opportunities for citizens to voice their views is insufficient. Without demonstrable and meaningful listening by elected officials, such processes are unlikely to enhance democratic legitimacy.
Third, responsiveness to information about deliberative citizen engagement does not vary across citizens’ underlying democratic conceptions. Legitimacy gains are observed among those with deliberative, technocratic, direct, and trustee conceptions alike. For advocates of reform, this is an encouraging finding: deliberative democratic innovations do not resonate solely with citizens already predisposed toward participatory or deliberative ideals. Instead, there is broad receptivity to the claim that structured citizen engagement can strengthen and complement existing representative institutions. Rather than reinforcing a dichotomy between ‘electoral’ versus ‘deliberative’ democrats (Hibbing and Theiss-Morse Reference Hibbing and Theiss-Morse2002), the results point to the cross-cutting appeal of hybrid democratic arrangements that integrate deliberative practices within representative systems. As Jacquet found through qualitative interviews, people desire ‘complex democratic processes where decisions result from the contribution of diverse types of actors with different legitimacies’ (Reference Jacquet2019: 653).
Moreover, the deliberative communication norms of consequentiality and listening matter most to those who trust political institutions least. Among institutional distrusters, information about a process of deliberative citizen engagement within parliament, especially when it emphasises listening, produces substantially larger legitimacy gains, while effects are essentially absent among high-trust respondents. Visible and meaningful practices of listening, this suggests, have the potential to repair legitimacy precisely among those least inclined to grant it.
Although grounded in a specific Australian parliamentary setting, the dynamics identified here are likely to travel to other representative democracies with similar parliamentary institutions and facing similar patterns of elite-citizen disconnection. More specifically, the study carries broader implications for the participatory potential of legislative committees. Our findings point to three ways in which committees may strengthen democratic legitimacy. First, they can expand their deliberative repertoire by moving beyond one-way consultation toward structured processes of dialogue and reflective feedback that incorporate ordinary citizens into formal proceedings. Second, embedding deliberative mechanisms within committee work may counterbalance elite dominance by bringing constituents’ voices more directly into legislative deliberation (Hendriks and Kay Reference Hendriks and Kay2019). Third, because elected officials tend to be more receptive to government-initiated participatory processes than to citizen-led initiatives, situating minipublics within established parliamentary structures may reduce elite resistance and facilitate institutional uptake (Goutry et al. Reference Goutry, Steyvers, Verschuere and Haesevoets2025).
The study is not without limitations. We acknowledge the constraints of vignette-based survey experiments with respect to external validity and behavioural inference. Future research should prioritise field experiments and studies embedded in real-world deliberative processes. That said, Hainmueller, Hangartner, and Yamamoto (Reference Hainmueller, Hangartner and Yamamoto2015) show that effects from vignette and conjoint experiments correspond ‘remarkably well’ with real-world behavioural outcomes. Moreover, recent work by Himmelroos et al. (Reference Himmelroos, Jäske and Setälä2026) reports that findings from an actual minipublic are comparable to diffuse legitimacy effects on wider publics. Both lend confidence that our findings accurately capture the dynamics of public evaluation. Future research should also examine how different media portrayals of deliberative forums (for example, critical, supportive, or partisan) shape these diffuse legitimacy effects to deepen our understanding of how information environments moderate the broader impacts of deliberative innovations. In addition, scholars should explore to what extent similar patterns of legitimacy effects hold in more polarised issue domains (such as immigration or climate change). Finally, a mixed-method design including interviews could explore more deeply how citizens interpret the norms of listening and consequentiality in practice. Understanding these processes is essential if deliberation is to fulfil its promise not merely as a participatory ideal but as a practical tool for rebuilding legitimacy.
Limitations notwithstanding, the findings of this study point to broader implications for democratic governance. They demonstrate that integrating deliberative citizen engagement into representative institutions can enhance public legitimacy and serve as a partial corrective for democratic disaffection, but only when citizens’ contributions are perceived as genuinely heard. In contexts of declining trust and widening distance between citizens and elected officials, deliberation cannot function as a symbolic add-on. Its legitimacy-enhancing potential depends on whether political actors demonstrably and respectfully engage with citizen input.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1475676526101509.
Data availability statement
Replication materials are available at https://doi.org/10.7910/DVN/JLU6MT.
Acknowledgements
The authors are grateful for insightful comments on the research design and preregistration from Kyle Peyton, Tobias Rohrbach, Timothy Gravelle, Ferran Martinez i Coma, Ben Goldsmith, and Francesco Bailo.
Funding statement
This research was supported by a grant from the Institute for Democratic Engagement and Accountability at Ohio State University.
Competing interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Ethical standards
The study was part of a larger project, all aspects of which were pre-registered: https://osf.io/yeu78. Here we focus on research questions 4 and 6 of the preregistration. The study received approval by the Human Research Ethics Committee at the University of Canberra (HRE4677).
Generative AI disclosure
During the preparation of this work, the authors used ChatGPT 4o to improve the language and spelling of some sections of the text. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the manuscript’s content.
Open access
