INTRODUCTION
The perceived crisis of democracy has produced a plethora of proposals of how democracy can be revitalized and made more legitimate in the eyes of citizens. This ranges from reforms in the current representative system (e.g., Levitsky and Ziblatt Reference Liu and Shiraito2018) to more “attentive” forms of democracy (Hibbing et al. Reference Lafont2023). A very popular proposal for revitalizing democracy is direct citizen involvement and sortition (Guerrero Reference Hainmueller and Hopkins2014; Landemore Reference Cummings2020), with scholars and practitioners pinning their hopes on deliberative minipublics (OECD Reference Parkinson2020). Minipublics comprise various dialogical participatory formats, where usually a randomly selected representative or at least diverse group of citizens discuss pressing policy issues and make recommendations to political decision-making.
Around the globe, minipublics have been used to address constitutional issues (e.g., the Conference on the Future of Europe [CoFoE]), political gridlock (e.g., Citizen Assembly on abortion in Ireland), long-term issues (e.g., global Climate Assemblies), and political collusion (e.g., Citizen Assembly on electoral reform in Canada) or are considered as a tool to fight political corruption. Empirical research shows that minipublics have several positive effects on citizens’ democratic attitudes. They increase political trust among participating citizens (Grönlund, Setälä, and Herne Reference Guerrero2010), reduce polarization (Fishkin et al. Reference Gastil and Knobloch2021), strengthen faith in democracy (Boulianne Reference Boulianne2019), promote intergenerational justice by focusing on the long-term (Smith Reference Smith2021), and provide non-participating citizens with information and recommendations to make informed political choices (Már and Gastil Reference Moellering2020). Against staunch critics, the internal workings of minipublics also largely conform to democratic and deliberative aspirations (Neblo, Esterling, and Lazer 2018).
These positive experiences—in combination with the claim that the “spectacle of electoral politics is so dispiriting at the moment” (Landemore in Cummings Reference Leeper, Hobolt and Tilley2022)—have led some democratic theorists and activists to call for institutional reforms that would increase the (direct) decision-making power of minipublics (Buchstein Reference Buchstein2019; Landemore Reference Cummings2020). Chwalisz (Gardels Reference Gardels2022) puts it as follows: “While citizens’ assemblies today are largely advisory and complementary to our existing electoral institutions, it is not impossible to imagine a future where binding powers shift to these institutions—or where they perhaps even replace established governing bodies in the longer term.” Indeed, Landemore (Reference Cummings2020) has proposed a vision of “open democracy” which (at least partly) replaces the legacy institutions of the representative system (especially parliaments) with minipublics.
Yet calls for minipublic empowerment have been vigorously challenged. The philosopher Cristina Lafont (Reference Landemore2019) argues that “lottocratic” forms of representation are undemocratic since non-participants never know whether the deliberating counterparts share their policy goals, interests, and values. If deliberative minipublics were authorized to making binding decision, non-participating citizens would blindly defer to a majority of (deliberating) citizens who are not like them, especially when the latter have changed their minds (a frequent finding in deliberative events). Consequently, Lafont (Reference Landemore2019) proposes that minipublics should be limited to an advisory and “deliberation-promoting” function rather than an empowered decision-making function.
To date, this debate is largely normative and is dominated by a positive view of the legitimacy-enhancing role of minipublics, with advocates often claiming that deliberative minipublics will revitalize democracy and help cure the current democratic malaise. We do know from the sparse existing research that citizens seem to have attitudinal sympathy for minipublics (Talukder and Pilet Reference Talukder, Pilet, Reuchamps, Vrydagh and Welp2023), while at the same time being reluctant to grant them strong empowerment (Bedock and Pilet Reference Bedock and Pilet2023; Christensen Reference Christensen2020; Goldberg and Bächtiger Reference Grönlund and Setälä2022; Pilet et al. Reference Pow2023).
Our starting point consists of embedding previous research in institutional approaches and especially innovation theories which may help better understand complex reactions of citizens toward institutional innovations such as minipublics. Innovation theories predict that adoption and resistance trends may be simultaneously present in individuals’ minds: individuals may find innovations attractive, but due to their novelty, they are uncertain how innovations will work in practice. Focusing on three different country contexts, the United States, Ireland, and Finland (see below), we first analyze whether the pattern predicted by innovation theory—generic support for deliberative innovations with simultaneous reluctance for empowerment (binding decisions versus recommendations only), autonomy (such as coupling versus decoupling from existing institutions), and asking for “additional provisions” (such as descriptively representative composition, large versus small size or clear versus narrow majorities for recommendations or decisions)—holds across different institutional and cultural contexts.
But this type of analysis is inherently conservative in the sense that it primarily focuses on those who know nothing about minipublics and have no experience with them. So what would people think about minipublics, not as they exist today, but in a system of the future where they are established players in the political game, and have involved and reached out to many citizens? Our article takes a crystal ball perspective for the first time: it analyzes whether and how support for minipublics and their design features changes when citizens have some familiarity with minipublics and have made concrete experiences with them (i.e., have participated in a minipublic).
Drawing on innovation theories, we predict that the more citizens know about minipublics and the more positive experiences they have made with real minipublics, the more supportive they are of an empowered and autonomous role for minipublics without additional provisions. Advocates, furthermore, argue that citizens are more likely to trust deliberative minipublics, relative to other political institutions. So we also look at whether there is evidence for these claims, focusing on citizens’ trust of those other citizens who comprise deliberative minipublics. We predict that low political trust and high trust in other citizens as political decision-makers are associated with calls for greater empowerment and autonomy of minipublics without additional provisions.
We test these predictions in the context of three identical and preregistered conjoint experimentsFootnote 1 in Finland, Ireland, and the United States, conducted in 2022. Conjoint experiments present participants with hypothetical scenarios and identify the attributes of hypothetical scenarios (in our case: deliberative minipublics) and then ask for evaluations of these scenarios (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hangartner and Yamamoto2014). The three country contexts (United States, Finland, and Ireland), in turn, were selected to maximize variation in levels of trust and experience with minipublics at the macro level (Gerring Reference Gherghina and Geissel2006) while simultaneously focusing on comparable cases, namely wealthy and established liberal democracies. Whereas the United States represents a political system with relatively low level of political trust, Finland represents a typical high trust society; Ireland, in turn, represents a country with a high level of minipublic experience.
Our results, first of all, carry a contentious but nuanced message to minipublic advocates and critics: non-participants view minipublics quite positively but struggle with empowered and autonomous roles without additional provisions. Despite large country-level differences between the United States, Ireland, and Finland regarding political trust and experience with minipublics, findings across the three countries are almost identical. By the same token, support for minipublics and their design features is indeed conditional on experience and familiarity with minipublics as well as individual-level political trust and trust in other citizens as political decision-makers: actual participation in minipublics, familiarity, low individual political trust, and high individual trust in other citizens as political decision-makers (especially when the two combine) produce more support for empowered and/or autonomous minipublics. But this does not imply that such citizens are actually in favor of empowered and autonomous minipublics; rather, our results indicate that they are just more open to this. At the end of the day, most citizens want minipublics to advise or (at best) complement representative institutions, not replacing them. Our results, secondly, also carry a bigger lesson for institutional reformers, especially in current times of democratic disorder. They show that while democratic innovations may engender general support from many citizens, their implementation is fraught with reservations (even among those familiar with it), which requires new design thinking (Saward Reference Saward2021).
In the following, we first rethink support for minipublics from the perspective of institutional and innovation analysis. We understand support of minipublics and their design features in connection with familiarity and experiences as well as political trust and trust in other citizens as political decision-makers. We then present the research design and methods. The article ends with a discussion of the results and their implications for the role of minipublics in democratic systems and for democratic renewal.
An Institutional and Innovation Framework to Understand Support or Resistance Toward Minipublics
We use institutional and especially innovation theories to address when and under what conditions citizens are supportive or resistant to the adoption of (empowered and autonomous) minipublics. In a nutshell, we assume that support for (empowered and autonomous) minipublics (without additional provisions) is conditional, dependent on one’s own experiences with minipublics, as well as perceptions of political trust and trust in other citizens as political decision-makers.
While institutional theories have flourished in political science since the late 1980s, institutional research has hardly been applied to the study of deliberative innovations such as minipublics. On this account, minipublics are novel institutions, implemented in an already existing institutional architecture. Historical institutionalist theories predict that established institutions will always yield higher support and legitimacy feelings (Hall and Taylor Reference Hardin1996). The reason is straightforward: beliefs and actions are strongly shaped by existing institutions, which in our case are the legacy institutions of the representative system (e.g., parliaments).
Innovation theories are more nuanced regarding resistance and adoption trends. They describe the process of adopting an innovation (as a novel institution in our case) as “the process which an individual […] passes from first knowledge of an information, to forming an attitude toward the innovation, to a decision to adopt or reject, to implementation of the new idea” (Rogers Reference Rojon, Rijken and Klandermans2003, 163). In general, innovation theories suggest that the adoption of innovations is more likely the higher their perceived usefulness of the innovation and the more people believe that the innovation will bring relative advantages compared to existing alternatives (Claudy, Garcia, and O’Driscoll Reference Claudy, Garcia and O’Driscoll2015; Straub Reference Straub2009). Innovation adoption, however, hinges on “compatibility” (i.e., the degree to which an innovation is perceived as consistent with existing values and past experiences), “complexity” (i.e., how easy the innovation is to understand), “triability” (i.e., the amount of experimentation in practice), and “observability” (i.e., how visible the innovation is). Despite a focus on such attributes of innovations, traditional innovation approaches still imply a “pro change bias” (Talke and Heidenreich Reference Talke and Heidenreich2014), assuming that individuals are generally open to change and new products. By contrast, a more recent and less established stream of innovation research argues that individuals might also have good reasons to resist innovations. This research points to functional barriers, such as the need to change routines or uncertainty about the exact consequences of the innovation, or psychological barriers, such as innovations clashing with entrenched traditions or norm that produce innovation resistance (Antioco and Kleijnen Reference Antioco and Kleijnen2010; Claudy, Garcia, and O’Driscoll Reference Claudy, Garcia and O’Driscoll2015).
What does this mean for the support of minipublics and their design features? Our theoretical framework predicts that citizens will assess novel institutions within an already existing and path-dependent political and institutional context. If citizens think that the legacy institutions of the representative system (e.g., the parliament and government) still work in a satisfactory way, they may see little reason to change these institutions, especially not in radical ways. At the same time, when going through the list of adoption factors, minipublics are fairly attractive innovations. On the one hand, minipublics contain attractive democratic features, such as directly involving citizen voices in an inclusive and discursive way. Besides, some citizens may also have instrumental reasons to try something new since they are unhappy with how legacy institutions work (a point to which we return below). Depending on their institutional setup, minipublics may also be fully compatible with the existing institutional architecture: if they are advisory-only and are strongly tied to legacy institutions (e.g., because they include both citizens and political actors), they may not jeopardize the working of established institutions. On the other hand, minipublics are also fairly complex, un-tried, and not very visible institutions (Rummens Reference Rummens and Reuchamps2016). They may raise fears and uncertainty on the part of outsiders who are not familiar or have no experience with them, including allegations of citizen incompetence and biased decision-making. Hence, citizens might not only be reluctant to grant minipublics strong empowerment and autonomy, they might also ask for additional provisions—such as large size or clear majorities for recommendations or decisions which make the signal of minipublic recommendations or decisions stronger (if the majority opinion is used to find the “correct” view; Arnesen et al. Reference Arnesen, Broderstad, Johannesson and Linde2019, 181); for familiar institutions, such additional provisions may be less relevant. In sum, citizens may have both reasons for adoption and resistance toward minipublics. In this regard, recent trends in innovation research claim that adoption and resistance trends may not constitute full opposites in citizens’ minds (Claudy, Garcia, and O’Driscoll Reference Claudy, Garcia and O’Driscoll2015, 528): Citizens may find minipublics generally attractive for democratic or instrumental reasons; but due to their complexity and their limited visibility, citizens may simultaneously be uncertain how minipublics will operate in practice.
Previous research lends some support to these predictions, showing simultaneous trends of attitudinal sympathy for minipublics and strong reservations toward minipublic empowerment (Bedock and Pilet Reference Bedock and Pilet2023; Christensen Reference Christensen2020; Goldberg and Bächtiger Reference Grönlund and Setälä2022; Pilet et al. Reference Pow2023; Rojon, Rijken, and Klandermans Reference Rothstein and Stolle2019; Talukder and Pilet Reference Talukder, Pilet, Reuchamps, Vrydagh and Welp2023). However, empirical findings do not always converge. Pow (Reference Pow, van Dijk and Marien2023) found in Northern Ireland that citizens view bodies with randomly selected citizens as legitimate as bodies in which citizen representatives are elected (or which combine citizens and elected politicians). van Dijk and Levefere (Reference Dinesen, Sønderskov, Sohlberg and Esaiasson2022) found in a survey experiment in Belgium that citizens have higher legitimacy feelings when minipublic recommendations are taken up. The problem of existing studies is that they concentrate on single countries or pose the question of empowerment in a different way, namely as political authorities willing (or unwilling) to take up minipublic recommendations (which may produce a “knee-jerk” reaction on part of citizens to favor uptake rather than dismissal).
Our first goal is to provide more robustness for previous findings and replicate them in three different country contexts, namely the United States, Ireland, and Finland (for the rationale of the country selection, see below). Drawing on previous research and innovation theory, our first basic hypothesis is as follows:
H1: Citizens in general have attitudinal sympathy for minipublics but will be simultaneously reluctant to grant them strong empowerment (i.e., minipublics making binding decisions) and autonomy (decoupled from existing institutions and organized by NGOs) and will ask for additional provisions (descriptively representative composition, large size, clear majorities for recommendations or decisions as well as consideration of all interests).
Now, this general trend may be moderated by familiarity and experience as well as political trust and trust in other citizens as political decision-makers. Indeed, minipublic advocates might rebut that previous research entails an overly conservative view, failing to assess support for empowered and autonomous minipublics in a new era of governance where minipublics are established players in the political game. On this view, scholars need to know what citizens would think if they are familiar with minipublics and have made concrete experiences with them. Familiarity and experiences provide a crystal ball look into the future of minipublics: while it may be well true that unknown and untried innovations produce resistance, familiarity and especially positive experiences with institutional innovations might significantly enhance adoption trends. Political trust and trust in other citizens as political decision-makers, in turn, provide a further window into the circumstances under which adoption or resistance trends prevail: while the amount of political trust is critical whether legacy institutions are seen as legitimate (or not), trust in other citizens as political decision-makers matters whether citizens are willing to delegate policy-making to citizen-led institutions such as minipublics.
The Role of Familiarity and Experience
The willingness to adopt innovations usually requires some familiarity with the innovation. When individuals have no experiences with an innovation, this may cause uncertainty about its exact consequences (Claudy, Garcia, and O’Driscoll Reference Claudy, Garcia and O’Driscoll2015). In this regard, minipublics usually have very limited visibility to the broader public (Rummens Reference Rummens and Reuchamps2016) and are abstract institutions outside citizens’ experiences with democratic practices. Our samples provide evidence for this claim: while 13 percent (United States) to 20 percent (Ireland, with the highest frequency of minipublic uses) of our respondents claim they know about minipublics, less than 5 percent have already made direct experiences, that is, participated themselves in a minipublic. By contrast, a large share of citizens, namely 40 percent (Ireland), 45 percent (Finland), and about 50 percent (United States), state that they have never heard about minipublics at all (see Supplementary Appendix B, Table B.5). However, social cognitive theory assumes that individuals can learn from their own experiences as well as from experiences of others (Bandura Reference Bandura1977). Hence, direct experiences (e.g., being part of a minipublic exercise) and indirect experiences (e.g., media reports on minipublics and their proceedings) can serve as availability heuristics on how a minipublic works in practice.
We argue that some previous familiarity with minipublics, positive experiences with and especially direct participation within them will make citizens more supportive of minipublics and more open to minipublic empowerment and autonomy. More experienced citizens might also have learned from either direct participation or information about minipublics that ordinary citizens are generally competent and capable of good-quality deliberation (as empirical research has found; Gerber et al. Reference Gerber, André, Susumu, Simon and Samuel2018). We also know from empirical research that participants of minipublics are usually very satisfied with the experience (Delli Carpini, Cook, and Jacobs Reference Delli, Michael, Cook and Jacobs2004). Thus, when citizens have observed that minipublics perform quite well in practice, this may not only contribute to a more positive view of minipublics but also reduce fears and uncertainty regarding minipublic empowerment, autonomy, and additional provisions. This leads to the second hypothesis:
H2: Citizens who are previously familiar with minipublics, have made (positive) experiences and have participated in a minipublic are more open to strong empowerment and autonomy of minipublics and ask for fewer additional provisions compared to citizens with no (or negative) experiences.Footnote 2
The Role of Trust
The viability and vitality of institutions are closely related to institutional and political trust (Warren Reference Warren and Warren1999). When institutions continuously fail to deliver basic goods or are largely unresponsive to citizen demands, satisfaction with democracy and trust in politics and institutions decreases and demands for institutional change becomes more likely. We focus here on trust rather than satisfaction, even though the correlation between satisfaction with democracy and institutional and political trust is high.Footnote 3 Trust is the more fundamental variable pertaining to normative expectations toward political institutions and actors, whereas satisfaction may also be regarded as an indicator of attitudes to policy outputs (Grönlund and Setälä Reference Grönlund, Setälä and Herne2007). However, introducing novel institutions also raises the question how much citizens would actually trust these new institutions to perform better. In case of minipublics where agents are citizens rather than elected representatives, the main source of trust then is no longer political trust, but trust in other citizens as political decision-makers. Our basic claim is that the “usefulness” and “relative advantage” (Claudy, Garcia, and O’Driscoll Reference Claudy, Garcia and O’Driscoll2015) of minipublics and their design features are conditional on the amount of political and trust in other citizens, or differently put, on the extent to which citizens trust existing political institutions and politicians as well as alternative or novel actors such as citizens.
Both political trust and trust in citizens as political decision-makers rely, on the one hand, on interpersonal trust (i.e., trusting politicians and citizens), either because someone trusts another person for personal reasons or because someone has confidence in a group of people without knowing them personally, but shares their social backgrounds or viewpoints (Hardin Reference Hibbing, Theiss-Morse, Hibbing and Fortunato2002). On the other hand, trust can rely on trust in institutions, that is, trusting parliaments (composed of politicians) and minipublics (composed of citizens).Footnote 4 Institutional trust can be conceptualized as a relationship between an individual and an institution and its achievements, without individuals having to know or trust specific persons (Lepsius Reference Levitsky and Ziblatt2016). Interpersonal and institutional trust is related, however, since institutions necessarily always include individuals (Moellering Reference Neblo, Esterling and Lazer2014).
We claim that trust can affect support for minipublics and their design features for two reasons. We identify the following two mechanisms that produce more support for empowered and autonomous minipublics.
Mechanism 1: Because citizens do not trust politics.
Many citizens today mistrust representative institutions (Citrin and Stoker Reference Citrin and Stoker2018). A considerable number of citizens believe that political institutions are not responsive to their needs and do no longer fulfill their expectations and act in the public interest (Craig, Niemi, and Silver Reference Craig, Niemi and Silver1990).Footnote 5 In our sample, only about half of the people are satisfied with the way democracy works in their countries. In Finland, the share of satisfied people is a bit higher (60 percent of the respondents rated the question above midpoint). On average, people in the United States are the most dissatisfied in our sample, with more than 20 percent stating to be deeply dissatisfied with the way democracy is working.Footnote 6 In Finland and Ireland, the share of deeply dissatisfied people hovers around 13 percent (see Supplementary Appendix B, Table B.11). Same patterns emerge for political trust, although the share of people who distrust politics is even higher. Trust in politics is highest in Finland, while it is lowest in the United States, where, for instance, more than a third deeply distrust political institutions and about half of the people do not trust politicians (see Supplementary Appendix B, Tables B.6 and B.7).
Several studies find a positive relationship between political dissatisfaction and support for the direct involvement of citizens, especially with regard to direct-democratic instruments (Gherghina and Geissel Reference Goldberg and Bächtiger2019) In a study from Belgium, Bedock and Pilet (Reference Bedock and Pilet2023) found that support of binding uses of sortition is stronger when distrust in parties and politicians combines with higher levels of political efficacy. However, Christensen, Karjalainen, and Lundell (Reference Christensen, Karjalainen and Lundell2016) found that distrustful citizens are less supportive of democratic innovations. These citizens do not seem convinced that democratic innovations—which are frequently organized top-down—arise from good intentions of authorities. Yet, there is a dearth of studies exploring the support for minipublics in connection with individual-level trust and specific design features.
We expect that low political trust might conduce citizens to be more supportive of institutional innovations such as minipublics, hoping that the latter will perform better than existing ones and better serve the public as well as their personal interest. By contrast, citizens with high political trust are more likely to stick with the legacy institutions of the representative system (Bornstein and Tomkins Reference Bornstein and Tomkins2015, 180). However, citizens with high political trust might still support minipublics as complements to legacy institutions, whereby minipublics may even increase trust of existing institutions. The third hypothesis is thus the following:
H3: Citizens with low political trust are more open to strong empowerment and autonomy of minipublics and ask for fewer additional provisions compared to citizens with high political trust.Footnote 7
Mechanism 2: Because citizens trust other citizens.
“Trust-based” approaches to minipublics (MacKenzie and Warren Reference MacKenzie and Warren2012; Warren and Gastil Reference Warren and Gastil2015) argue that deliberating citizens in minipublics represent especially trustworthy democratic agents, since they are capable of making informed assessments of the merits and downsides of policy proposals without congealed partisan or interest-based constraints. Moreover, citizen representatives in minipublics may evoke “like-me” feelings (Pow, van Dijk, and Marien Reference Progrebinschi2020) among non-participating citizens since the former have similar backgrounds or share similar experiences. However, trusting minipublics requires that citizens trust their fellow citizens for being competent policymakers. Consequently, citizens with high trust in other citizens as political decision-makers and minipublics may be more supportive of minipublics as well as more open to minipublic empowerment and autonomy. In our sample, trust in citizens as political decision-makers and minipublics is remarkably higher than trust in politics, with high trust scores between about 50 percent in Finland and the United States and about 60 percent in Ireland (see Supplementary Appendix B, Tables B.8 and B.9).
This trust-based account of minipublics has been contested, however. According to Lafont (Reference Landemore2019), non-participating citizens—including citizens with high trust in other citizens as political decision-makers—have no reason to trust minipublics since they will never know whether their deliberating fellow citizens actually share their interests, values, or policy goals. As Lafont (Reference Landemore2019) puts it: “many [non-deliberating citizens] will find out that the majority of the sample is not like them, since they actually oppose their view, values and policy objectives on the issue in question.” The “mirror” component in minipublics also falters at “the moment a group’s opinions start to depart from the distribution found in the general population” (Parkinson Reference Pilet, Bol, Vittori and Paulis2006, 81). For Lafont (Reference Landemore2019), citizens can only trust democratic agents whose political views they broadly share or whose judgments they have reasons to trust (126). Yet, Lafont’s argument is purely normative, and it is an open empirical question whether citizens with high trust in other citizens as political decision-makers would buy this argument.
To date, few studies have examined the link between perceptions of other citizens, trust in other citizens as political decision-makers, and support for minipublics. They show that anticipated capacities of fellow citizens (Bedock and Pilet Reference Bedock and Pilet2021) and “like-me” perceptions (Pow, van Dijk, and Marien Reference Progrebinschi2020) positively affect support for minipublics. Yet, it is less clear whether citizens support minipublics because they trust them for their institutional setup (consisting of citizens), or because they trust citizens. We will test for trust in minipublics and trust in citizens separately. The fourth hypothesis is as follows:
H4: Citizens with high trust in other citizens as political decision-makers are more open to strong empowerment and autonomy of minipublics and ask for fewer additional provisions compared to citizens with low trust in other citizens as political decision-makers.Footnote 8
Finally, we draw on research focusing on the link between political trust and “social trust” (e.g., Dinesen et al. Reference Esaiasson, Persson, Gilljam and Lindholm2022) and explore whether political trust and trust in other citizens as political decision-makers interact when low trust in politics coincides with high trust in citizens. Our (tentative) expectation is that citizens with low political trust and high trust in other citizens as political decision-makers are particularly open to strong empowerment and autonomy of minipublics and ask for less additional provisions compared to the rest.
While our main focus is on individual attitudes toward minipublics, we nonetheless take context into account. We adopt a so-called diverse case strategy (Gerring Reference Gherghina and Geissel2006, 97ff.), trying to maximize variation in levels of trust and experience with minipublics at the macro level by selecting countries with extreme values on the variables of interest while simultaneously focusing on comparable countries, namely wealthy, established, and liberal democracies. Whereas Finland represents a typical high trust society with relatively few experiences with publicized and visible minipublics—which is a precondition that familiarity with minipublics can emerge—the United States is characterized by lower levels of political trust and little experience with minipublics as well.Footnote 9 Ireland, by contrast, had sustained experiences with publicized, consequential minipublics and has a medium trust level (in-between Finland and the United States).Footnote 10 The three countries enable us to compare potential effects of higher levels of experience with minipublics versus little experience (Ireland versus United States and Finland) as well as high versus low political trust (United States versus Finland).
DATA AND METHODS
We use data from three identical conjoint experiments in Finland (n = 2,005), Ireland (n = 2,007), and the United States (n = 2,045) which are representative for the population aged 18 and older. For each country, quotas were set for gender, age, education, and region. As for Ireland and the United States, data collection was administered by the survey-sampling company Psyma between January and March 2022. For Finland, respondents were recruited through Qualtrics between December 2021 and February 2022. Almost all the defined quotas were in the targeted corridor, except for Finnish and Irish respondents aged 60 years and older. More information on the samples and populations can be found in Supplementary Appendix B, Tables B.1–B.3.
The three countries were selected in order to ensure macro variation in levels of trust and experience with minipublics. We use a conjoint design to separate design features and authorization mechanisms of minipublics and to build knowledge on support for minipublics in various institutional and cultural contexts. Conjoint experiments are particularly attractive because they allow for estimating several effects simultaneously without having to observe all possible combinations of attributes (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hangartner and Yamamoto2014). The conjoint design included nine attributes, with two to three attribute levels (Table 1). In order to avoid primacy and recency effects, we randomly assigned the attribute order for each respondent in all three surveys. Full randomization allows us to estimate the causal effects of each attribute on the probability of preferring specific minipublic design features.
Table 1. Institutional Design of Minipublics
Since we expected that most citizens are not familiar with minipublics and their design features, we presented respondents with two information packages (video and argument sheets with arguments on various design features) prior to the conjoint experiment. Additionally, respondents also had the opportunity to access a glossary with information on the design features when answering the conjoint questions (see Supplementary Appendix A).
The three conjoint experiments focus on various (real and hypothetical) institutional design features of minipublics that are critical in the theoretical debate on minipublics as well as relevant in the practical application (Table 1). This ranges from authorization to autonomy (initiative and composition) and to additional provisions such as recruitment, size, consensus, and aim. Besides institutional design features, we also included different issue types—distinguishing between salient issues that are more technical (climate change) and salient issues that are less technical (refugee crisis)Footnote 11—as well as outcome favorability (i.e., whether the substantive outcomes of deliberative forums conform to the substantive policy preferences of non-participants).Footnote 12 The wording was identical for all attributes in the three country contexts, except the immigration issue. This issue (unavoidably) taps into different aspects of immigration in the different countries; our goal was to focus on the most salient ones in the three countries at the time when the survey was conducted. Yet, the issue dimension and the related question of outcome favorability only functions as additional variables in our analysis of minipublic empowerment.
Support for Minipublics
To evaluate which minipublic scenario produces higher support, we employed a choice-based conjoint design where respondents were randomly presented with five comparisons of minipublics. We used a paired conjoint design since this design performed best when it came to replicate “natural benchmarks” (Hainmueller, Hangartner, and Yamamoto Reference Hall and Taylor2015). For each comparison, respondents were asked to choose their preferred scenario. We use the forced choice-based outcome variable as our main quantity of interest, which equals 1 if a specific minipublic scenario was chosen and 0 otherwise.
Additionally, we reran analyses using a rating outcome variable,Footnote 13 allowing participants to disapprove both scenarios.Footnote 14 However, we find no substantive differences between forced choice and rating outcomes (see Supplementary Appendix E in the Dataverse, Figure E.1).
In addition, we employed a question on the retrospective assessment of minipublics to assess overall support of minipublics. We asked participants to what extent they think that minipublicsFootnote 15 are an appropriate way to involve citizens in political decision-making. This was measured on a 7-point scale whereby 1 indicates that “they are very inappropriate “and 7 indicates that “they are very appropriate.”
Familiarity and Experience
We included two variables to measure familiarity and experience with minipublics. First, we asked all respondents whether they are familiar with minipublics, using the following categories: (1) “I am not familiar with citizens’ forums”; (2) “I have heard about citizens’ forums before”; (3) “I know much about citizens’ forums, but have not participated”; and (4) “I have already participated in one or more citizens’ forum myself.” The first and second categories indicate that citizens are not familiar with minipublic; the third and fourth categories indicate that citizens had some indirect or direct experiences with minipublics. In a second step, we asked participants who indicated that they have at least heard about minipublics (category two or higher) to state on a 7-point scale how positive or negative they judge their experience with minipublics. We dichotomized the 7-point scale at the median (MD: 4) to create two groups, whereby values above the median indicate positive experience and values below the median indicate negative ones. Of course, this is a relatively crude measure, not capturing what types of positive and negative experiences citizens have made with minipublics; given the time constraints of our conjoint experiments, we thought that this information would be too demanding to obtain from survey respondents. Finally, we also focus separately on those citizens stating that “[they] have already participated in one or more citizens’ forums.” Notice that these are very few citizens (hovering around 4 percent in the three country contexts), requiring that we use a combined analysis of the three country contexts to have enough statistical power in conjoint analysis.
Political Trust and Trust in Other Citizens as Political Decision-Makers
We use standard items to measure political trust and use novel measures to capture trust in other citizens as political decision-makers (Pow, van Dijk, and Marien Reference Progrebinschi2020). Regarding political trust we focus on both institutional and interpersonal dimensions of political trust. The institutional dimension was measured by asking respondents how much they personally trust the government and the parliament, which we combined into a single trust variable. The interpersonal dimension was measured by asking how much they personally trust politicians. All items use a 7-point scales where 1 means “I do not trust at all” and 7 “I do extremely trust.” The correlation between the two is 0.82. Regarding trust in other citizens as decision-makers we again focus on institutional and interpersonal dimensions. The institutional dimension focuses on trust in minipublics. Respondents were asked whether they think that minipublics in general can be trusted, with a scale ranging from 1 “you cannot trust citizens’ forums at all” to 7 “you can fully trust citizens’ forums.” The interpersonal dimension was measured by asking how much respondents trust ordinary citizens to make political decisions that are in the public interest and how much they trust ordinary citizens to make good political decisions. Both were combined into a single variable. All items use 7-point scales where 1 means “strongly disagree” and 7 “strongly agree.” The correlation between the institutional and interpersonal dimension of trust in citizens as political decision-makers is 0.62. Finally, we check correlations between political trust and trust in citizens as political decision-makers we combined all political trust variables and all trust in citizens variables to two separate indices. The correlation between political trust and trust in citizens as political decision-makers is 0.22.Footnote 16 For subgroup analyses, all four trust variables were dichotomized at the median to form two distinctive groups (low trust and high trust);Footnote 17 for robustness checks, we have used different cut values (see Supplementary Appendix E in the Dataverse, Figure E.4, E.5).
Statistical Analysis
We estimate the average marginal component effect (AMCE) each attribute has on choosing a certain scenario (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hangartner and Yamamoto2014), which in our case can be interpreted as the average differences in the probability of preferring a minipublic when comparing two different attribute levels (e.g., binding decision versus recommendation). While AMCE analyses have been widely used in political science and become state of the art to causal interpretations (Hainmueller and Hopkins Reference Hainmueller, Hopkins and Yamamoto2015), Leeper, Hobolt, and Tilley (Reference Lepsius2020) argue that AMCEs may be difficult to interpret and misleading for subgroup analysis. Hence, we additionally estimated marginal means for familiarity and trust to provide a descriptive representation of the findings (see Supplementary Appendix D in the Dataverse, Figures D.2, D.4, D.6, D.7, D.10, D.11, D.14, D.15, D.18, D.19); results, however, are consistent with the ones reported from AMCEs. Notice that the moderating variables such as familiarity and trust are not randomized, precluding causal interpretation.
FINDINGS
We present findings in three steps. First, we focus on general support for minipublics in connection with their design features. Second, we test how familiarity and experience affects support for minipublics and their design features. Finally, we analyze support for minipublics and their design features in relation with political trust and trust in other citizens as political decision-makers.
First, we checked for average support of minipublics in general. On average, respondents in our sample view minipublics as fairly desirable tools in political decision-making (M = 4.85; SD = 1.36 on a 7-point scale), partly corroborating our expectation that minipublics are attractive institutional innovations. In this regard, Finnish respondents are a bit more reluctant on average (M = 4.47; SD = 1.29) compared to respondents in Ireland (M = 5.05; SD = 1.31) and in the US (M = 5.04; SD = 1.39). As mentioned before, the overall satisfaction with democracy in Finland as well as trust in political institutions is comparably high, slightly reducing the need to turn to alternative institutions such as minipublics.
Moving beyond these descriptive findings, we examine the effect of design choices on support for minipublics. Figure 1 reports the AMCEs of all attribute levels with 95 percent confidence intervals. Points without bars indicate the reference level for each attribute. For group composition, for example, mere citizen groups is the reference category. The estimated effect for mixed groups indicates that respondents were 5.7 percentage points more likely to choose a minipublic that was composed of politicians and citizens than a minipublic that was composed of citizens only.
Figure 1. Effects of Minipublic Attributes on Support across Countries
Note: This figure reports the AMCEs. The choice outcome (whether respondents have chosen a scenario) is the dependent variable. Standard errors clustered at the individual level to consider that each respondent made several comparisons. N = 60,570 (6,057 respondents x 10 scenarios). Complete results and information on observations and variables can be found in Supplementary Appendix C, Table C.1.
Overall, Figure 1 shows that support for minipublics strongly hinges on their institutional design features.Footnote 18 Corroborating H1, we find that citizens in general and across countries (Figure 2) are more likely to reject empowered minipublics (making binding decisions) compared to advisory minipublics. More precisely, the effects for authorization show that a minipublic with binding decision-making capacity is about 4.1 percentage points less likely to be chosen than one that advises elected officials. There is no discernible effect of the minipublics making recommendations followed by public referenda, indicating that a minipublic with this level of authorization was not more or less likely to be chosen than one that advises elected officials.
Figure 2. Differences between Countries
Note: This figure reports the AMCE. The choice outcome (whether respondents have chosen a scenario) is the dependent variable. Standard errors clustered at the individual level. Finland n = 20,050; Ireland n = 20,070; United States n = 20,450. Complete results and information on observations and variables can be found in Supplementary Appendix C, Table C.2.
The marginal means, however, indicate descriptively that while a minipublic making recommendations to public officials is the most preferred and minipublics making binding decisions is the least preferred scenario with minipublics making recommendations followed by a public referendum located in-between. The latter category is not significant in the United States though. Note, however, that these findings do not indicate a general rejection of empowered uses of minipublics. It only shows that citizens tend to be more open to advisory roles compared to empowered uses. Furthermore, we speculate that support for minipublic recommendations with a direct-democratic referendum might either reflect a general preference for referendums or indicate that citizens want minipublic recommendations to have some effect on policymakers, albeit not in a direct way (van Dijk and Lefevere Reference Dinesen, Sønderskov, Sohlberg and Esaiasson2022). Moreover, our results indicate that citizens tend to be more open to non-autonomous minipublics compared to autonomous ones, which are tightly coupled with the representative system (with a 5.7 percentage points increased likelihood to be chosen when the scenario included clear preference for mixed group memberships compared to citizens only).
Our findings also indicate that respondents are more likely to ask for additional provisions, namely random selection (compared to self-selection), larger groups (compared to smaller groups), and recommendations backed by a clear majority (compared to narrow majority). No difference can be found whether minipublics should carefully ponder all interests or be efficient in decision-making. Regarding issue type, citizens are more likely to choose minipublics for pressing issues such as fighting climate change but find them less appropriate to tackle salient and less technical issues such as immigration (with the exception of Ireland). We can only speculate why this is the case: one explanation is that citizens may prefer minipublics for issues where they think that some common ground can be found (such as climate change) compared to deeply polarized issues (such as immigration) where this is not the case (Pilet et al. Reference Pow2023). Finally, and in line with previous research, minipublics are also more likely to garner support when the output corresponds to the respondent’s own policy preference.
In sum, the findings reveal that most respondents are more open toward non-empowered and non-autonomous minipublics with additional provisions. A further intriguing finding is that there are very few differences regarding countries (see Figure 2 and Supplementary Appendix D in the Dataverse, Table D. 1, Figures D.1–D.3, including a formal test). While binding decisions and minipublic autonomy do not find support in all three country contexts (compared to recommendations and recommendations followed by a referendum), US citizens are less likely to choose minipublics making recommendations followed by a referendum compared to Ireland and Finland (they are indifferent between recommendations to public officials and recommendations followed by a referendum).
Next, we ask how familiarity and experiences with minipublics affect support. First, Table 2 shows the average support for minipublics among citizens with various levels of familiarity and direction of experiences. The findings indicate that respondents who are not familiar with minipublics or only have heard about them find them a little bit less appropriate for involving citizens in political decision-making. However, respondents who have already made positive experiences seem to be more optimistic toward minipublics compared to citizens with negative experiences. The direction of these effects is fairly consistent across the three countries (see Supplementary Appendix D in the Dataverse, Figures D.8, D.12, D.16).
Table 2. Average Support for Minipublics Conditional on Levels of Familiarity, Participation, and Direction of Experiences
Note: Dependent variable is retrospective assessment of minipublics, 7-point scale. All reported mean differences are statistically significant (p < 0.001).
Second, we examine how familiarity and experience affect support for design features. In this regard, Figure 3 shows the differences in AMCEs by levels of familiarity, participation, and direction of experiences. Partly consistent with H2, the findings for familiarity show that indirect or direct experience with minipublics increases the chance that citizens are more open to autonomous minipublics (indicated by the negative difference effect for mixed groups, i.e., the effect for citizens who are familiar with minipublics minus the effect for citizens who are not familiar with minipublics). Familiarity, however, does not seem to affect citizens’ openness to empowered uses of minipublics. Moving to the actual participation in a minipublic and the type of experiences citizens have made with them, Figure 3 shows that citizens having made positive experiences with minipublics or having participated in a minipublic put more weight on strong empowerment compared to citizens who have made negative experiences and have not participated in a minipublic. Three comments are in order. First, we have probed for differential effects of those who have rated their participation experience as positive (about 75 percent) or negative (about 25 percent). We find no differences for evaluations of empowerment and autonomy (see Supplementary Appendix D in the Dataverse, Figure D.4). Second, the result for positive experiences, however, is not fully robust when different cut values for positive and negative experiences are applied (see Supplementary Appendix E in the Dataverse, Figure E.6). Third, the analyses in Figure 3 only indicate that people who have some familiarity with minipublics, have participated, or have made positive experiences are more open to strong empowerment and/or autonomy); but it does not mean that they are really in favor of this (see Supplementary Appendix D in the Dataverse, Figures D.4, D.5, and E. 17). Fourth additional provisions, issue characteristics, and outcome favorability do not produce any differences between citizens with no (or negative) experiences (see Figure 3 and Supplementary Appendix D in the Dataverse, Figure D.4).
Figure 3. Effects of Minipublic Attributes on Support Conditional on Familiarity and Experiences
Note: Effects show the difference (increase/decrease) in the probability of choosing a scenario for a particular attribute level relative to its baseline level for citizens who (a) are familiar with minipublics (left), (b) have participated (middle), and (c) have positive experiences (right) minus the probability of choosing a scenario for citizens who (a) are not familiar with minipublics (left), (b) have not participated (middle), and (c) with negative experiences (right) for the same attribute level relative to its baseline category. Baseline categories are not shown. Complete results and information on observations and variables can be found in Supplementary Appendix C, Table C.3.
Finally, we ask how political trust and trust in other citizens as political decision-makers affect support for minipublics. First, Table 3 again shows the average support for minipublics among citizens with low versus high political trust (left column) and high versus low trust in other citizens as political decision-makers (right column). As expected, support for minipublics is higher among citizens who trust other citizens. Support, however, is slightly higher among citizens with high political trust, corroborating Christensen, Karjalainen, and Lundell (Reference Christensen, Karjalainen and Lundell2016) that distrusting citizens tend to be generally less supportive of democratic innovations (these differences are small, however). These differences are consistent across the three countries (see Supplementary Appendix B, Tables B.14–B.16).
Table 3. Average Support for Minipublics Conditional on Low versus High Levels of Trust
Note: Dependent variable is retrospective assessment of minipublics, 7-point scale. All reported mean differences are statistically significant (p < 0.001).
Consistent with H3 and H4, Figures 4 and 5 show the differences in effects for all attribute levels for low institutional and interpersonal political trust compared to high political trust (Figure 4) and for high institutional and interpersonal trust in other citizens as political decision-makers compared to low trust in other citizens as political decision-makers (Figure 5). First, partly corroborating H3, citizens with low political trust are more open to minipublic autonomy, indicated by the negative effect regarding differences for government (top-down organization) compared to mixed group composition. Although, however, low institutional trust conduces to a higher probability to choose minipublics followed by referenda compared to mere advise to public officials, binding decision-making appears not to be a desirable mechanism, even in the context of low institutional trust. Moreover, our findings indicate that additional provisions tend to be equally important to citizens with high and low political trust. The only significant difference is for large groups. Issue type and substantive considerations seem not to matter differently either.
Figure 4. Effects of Minipublic Attributes on Support Conditional on Low Political Trust
Note: Effects show the difference (increase/decrease) in the probability of choosing a scenario for a particular attribute level relative to its baseline level for low political trust minus the probability of choosing a scenario for high political trust for the same attribute level relative to its baseline category. Baseline categories are not shown. Complete results and information on observations and variables can be found in Supplementary Appendix C, Table C.4.
Figure 5. Effects of Minipublic Attributes on Support Conditional on High Trust in Other Citizens as Political Decision-Makers
Note: Effects show the difference (increase/decrease) in the probability of choosing a scenario for a particular attribute level relative to its baseline level for high trust in other citizens as political decision-makers minus the probability of choosing a scenario for low trust in other citizens as political decision-makers for the same attribute level relative to its baseline category. Baseline categories are not shown. Complete results and information on observations and variables can be found in Supplementary Appendix C, Table C.4.
Second, corroborating H4, citizens with high trust in other citizens as political decision-makers put more weight on minipublics making binding decisions compared to citizens with low trust in other citizens as political decision-makers. This is indicated by the positive effect regarding differences for binding decision-making compared to advising elected officials. Contrary to our expectations, however, high trust in other citizens as political decision-makers is not related to giving minipublics an autonomous role. Additional provisions, issue type and substantive considerations are equally important to citizens with high and low trust in other citizens as political decision-makers. In general, effects both for low political and for high trust in other citizens as political decision-makers seem to be consistent across countries, except for the United States, where we did not find any differences in effects for high compared to low trust in citizens as political decision-makers (Supplementary Appendix D in the Dataverse, Figure D.18). Taken together, strong empowerment is mainly associated with high trust in other citizens as political decision-makers and, albeit to a lesser extent, with high trust in minipublics. Decoupling, by contrast, appears to be mainly related to low trust in political institutions and politicians.
Additionally, we explore whether both mechanisms (low political trust and high trust in other citizens as political decision-makers) reinforce each other.Footnote 19 To do so, we split respondents into two groups. The first group consists of respondents who scored low on political trust, but high on social trust (in our sample 27.5 percent are simultaneously not trusting politics but trusting citizens); the second group consists of all other respondents.Footnote 20 Figure 6 shows that there is a reinforcement effect: citizens who simultaneously distrust politics but trust other citizens are not only more open to binding decision-making and public referenda compared to advisory roles only but also more clearly against top-down procedures and more reluctant toward minipublics composed of citizens and public officials compared to bottom-up procedures and mere citizens groups. Additional provisions and substantive considerations, in turn, appear to be equally important to both groups. Note, however, these effects are most evident for Finland (high trust society) and somewhat less so for Ireland. Conversely, in the United States, there do not seem to be any significant differences. Table 4 presents a summary of hypotheses and findings.
Figure 6. Combined Effect for Low Political Trust but High Trust in Citizens as Political Decision-Makers
Note: Effects show the difference (increase/decrease) in the probability of choosing a scenario for a particular attribute level relative to its baseline level for low political trust combined with high trust in citizens as political decision-makers minus the probability of choosing a scenario for respondents who do not fall into this category for the same attribute level relative to its baseline category. Baseline categories are not shown. Complete results and information on observations and variables can be found in Supplementary Appendix C, Table C.5.
Table 4. Summary of the Hypotheses and Findings
Finally, we did robustness and diagnostic checks (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hangartner and Yamamoto2014), as well as attention checks (see Supplementary Appendix E in the Dataverse). With the exception of the direction of experiences (see above), the various tests indicate that the assumptions by Hainmueller, Hopkins, and Yamamoto (Reference Hainmueller, Hangartner and Yamamoto2014) are fulfilled and the presented results are highly robust given different model specifications and different operationalizations of variables. To take into account multiple hypothesis testing, we have implemented a test for conjoint analysis developed by Liu and Shirato (Reference Már and Gastil2023). This entails the Bonferroni correction, the Benjamini–Hochberg procedure, and the adaptive shrinkage (Ash). We find that the confidence intervals reported in the text are highly robust when applying these corrections (see Supplementary Appendix E in the Dataverse, section on “multiple hypothesis testing”).
DISCUSSION
Based on our institutional innovation framework, we have three messages to advocates and critics of minipublics. First, minipublics are attractive institutional innovations—displayed by the moderately positive assessment of minipublics being an appropriate way to involve citizens in political decision-making—but since most citizens are neither familiar with them nor have made concrete experiences, citizens in general are very reluctant to grant them empowerment and autonomy as well as ask for additional provisions. This corroborates expectations from innovation and institutional theories, emphasizing that citizens will assess novel institutions within an already existing and path-dependent institutional and political context and that adoption and resistance trends can occur simultaneously. It also brings together divergent findings from existing research, finding that citizens like minipublics but still have many reservations. An intriguing result here is that these findings are almost identical across dissimilar countries—United States, Finland, and Ireland—with different levels of institutional trust and experience with minipublics.
Second, looking into the crystal ball, familiarity with and actual participation in minipublics does produce more support for empowered and/or autonomous minipublics. Regarding trust, we need to properly distinguish between two types of trust (a topic neglected by previous studies): trust into politics (whereby respondents do not differentiate between politicians and political institutions) and trust into citizens as political decision-makers’ (also forming a compound construct of trust in citizens and trust in minipublics). This, in turn, has complex effects on support for minipublics and their design features: both low individual political trust and high individual trust in citizens as political decision-makers (especially when the two combine) in general lead to more openness toward empowered and autonomous minipublics. But even among citizens who tick all these boxes, that is, have made positive experiences with or even have participated in minipublics, have high trust in other citizens as political decision-makers, and are distrustful of existing institutions, there is a high degree of a damage control calculation regarding novel and untried institutions. None of these citizens is really in favor of strong empowerment and autonomy of minipublics, while additional provisions such as large size or clear majorities for recommendations are continuously required by all types of citizens (see Dataverse, Figures E.17–E.24). For minipublic advocates, this means that even in contexts where the “spectacle of electoral politics is […] dispiriting” (Landemore in Cummings Reference Leeper, Hobolt and Tilley2022) or where minipublics have been used quite frequently (as in Ireland) and citizens have made positive experiences with them, there seems no desire for radical institutional change toward a system of empowered minipublics (Landemore Reference Cummings2020), replacing elections or wider civic engagement. Nonetheless, in the eyes of the citizens, minipublics are attractive institutional additions to the existing institutional architecture. Table 4 summarizes the findings.
These nuanced findings of minipublic support and resistance call for re-thinking of the exact purposes of minipublics in political systems (Bächtiger and Dryzek Reference Bächtiger and Dryzek2024, 194–6). Lafont (Reference Landemore2019) proposes a “participatory track”: minipublics may produce “added value” for political systems when they contest the majority opinion, when they play a vigilant role, or when they anticipate issues. On this “participatory” interpretation, deliberative minipublics may be useful for democratic systems by yielding a “deliberation-promoting” function for the citizenry. There is reason for skepticism, however: minipublics will always need to compete for public attention with actors that are much better equipped to shape public debate.
But there are alternatives. Drawing from the idea of “trust-based” uses of minipublics (Warren and Gastil Reference Warren and Gastil2015), we can think of a “shortcut track” where minipublic recommendations can provide trustworthy input for citizens’ opinion formation rather than being major contributors or shapers of public debate. This trust-based use of minipublics is particularly useful, for instance, when citizens have to make choices in direct-democratic voting under conditions of a distorted public sphere (Gastil and Knobloch Reference Gerring2019). We can also think of a “political track” where input from minipublics is primarily processed by representative institutions and bureaucracies. In this regard, Neblo, Esterling, and Lazer (2018) propose to create more direct exchanges between the represented and representatives. They do so by bringing MPs and randomly selected constituents together, engaging them in a common deliberative forum on pressing issues of public policy.
CONCLUSION
This article has investigated the potential of deliberative minipublics to provide a new set of institutions for democratic renewal, especially by taking a crystal ball view and exploring how citizens would judge empowerment and autonomy of minipublics had they made concrete experiences with them. Replicating previous research, we first show that in the eyes of (non-participating) citizens minipublics are quite attractive institutional innovations, but that citizens in the United States, Ireland, and Finland in general are very reluctant to grant them empowerment and autonomy as well as ask for additional provisions such as large size or clear majorities for recommendations. But even when looking into the crystal ball, this result does not fully change: even though citizens who have made concrete experience with minipublics (i.e., participated in them) as well as citizens who have trust in other citizens as political decision-makers and low political trust are more open to minipublic empowerment and autonomy, they are not really in favor of it. Put differently, even those citizens do not seem to want to replace representative institutions with minipublics (as some minipublic advocates have suggested). Our results thus carry important lessons for institutional reformers more generally. In line with the prediction of innovation theories, they show that while many citizens like democratic innovations, they nonetheless have strong reservations regarding their empowerment. This requires new design thinking (Saward Reference Saward2021) where institutional designers may come up with “blended” designs of democratic governance, mixing democratic innovations with legacy institutions in smart ways.
Future research will need to take a more in-depth view of what the exact rationales are behind adoption and resistance trends, especially the stark reluctance to grant them further empowerment. Is it because citizens buy into Lafont’s “blind deference” argument and are skeptical whether ordinary citizens truly represent their viewpoints and interests; is it because citizens do not have enough experiences with minipublics? Or is it because the current representative system has not completely failed in the eyes of citizens? On the other hand, we also need to zoom in on positive and negative experiences. What does it exactly mean when one has had a positive or negative experience with minipublics? Does positive mean that one just got what one wanted regarding outcomes, or is it because one has made a democratizing experience? And does negative mean that one is dissatisfied with the outcome or the procedure? Qualitative research is needed to uncover these dimensions. Finally, future research on public views on minipublics also needs to look beyond wealthy, established and liberal democracies and include minipublics in the Global South, representing a different setting both when it comes to minipublic designs and to questions of empowerment. In the Global South, we confront, on the one hand, a larger variety of minipublic designs (Progrebinschi Reference Radin2023). On the other hand, given the manifold failures of existing representative institutions including clear as well as symbolic separations between political authorities and citizens, the demand for a (partial) replacement of legacy institutions by “citizens like me” might be stronger. These open questions notwithstanding, our study of minipublic support in three different country contexts shows that all types of citizens (with or without experience and with or without trust in other citizens as decision-makers) do not share the enthusiasm of advocates for a lottocratic revolution of contemporary democratic systems.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit http://doi.org/10.1017/S0003055424001163.
DATA AVAILABILITY STATEMENT
Research documentation and data that support the findings of this study are openly available at the American Political Science Review Dataverse: https://doi.org/10.7910/DVN/ZU4FMC.
ACKNOWLEDGMENTS
We are grateful to valuable comments on earlier drafts of this manuscript. We thank Michael Bechtel, Mark Warren, James Druckman, Dominik Hangartner and our colleagues at KU Leuven (particularly the LEGIT research group), University of Stuttgart, and Åbo Akademi University. Moreover, we want to thank Anja Rieker and Michael A. Neblo for valuable assistance as well as the participants of the Deliberative Democracy Summer School in Turku 2022, the NIG conference in Tilburg 2022 and ECPR general conference in Innsbruck 2022 for their helpful comments.
FUNDING STATEMENT
This project has received funding from the German Research Foundation (project number 432370948), the European Research Council under the European Union’s Horizon 2020 research and innovation program (grant agreement No 350361), and the Academy of Finland (grant 353720) and was supported by the Strategic Research Council of the Academy of Finland under Grant 312676.
CONFLICT OF INTEREST
The authors declare no ethical issues or conflicts of interest in this research.
ETHICAL STANDARDS
The authors declare the human subjects research in this article was reviewed and approved by the Ethics Committee of the University of Stuttgart (May 17, 2021, Az 21–008, see Supplementary Appendix). The authors affirm that this article adheres to the principles concerning research with human participants laid out in APSA’s Principles and Guidance on Human Subject Research (2020).
Open access
Comments
No Comments have been published for this article.