Deliberative Mini-publics As a Partial Antidote to Authoritarian Information Strategies

Authoritarian and proto-authoritarian regimes control a growing number of states throughout the world. Among the information strategies that these regimes use to gain and maintain support are the dissemination of false or misleading policy information and the use of manipulative policy frames. Deliberative mini-publics can partially counter those strategies by distributing accurate policy information and employing non-exploitative policy frames that affirm the dignity of members of the polity as free and equal citizens.

Authoritarian and proto-authoritarian 1 regimes are taking power throughout the world (Diamond, Plattner, & Walker, 2016;Fishkin & Mansbridge, 2017). Since 2000, approximately 25 formerly democratic states-including Russia, Turkey, and Venezuela-have become dictatorships or have experienced a substantial degradation of democratic institutions (Diamond, 2015). In addition, several states with populist governments-such as Hungary and Poland-have exhibited "democratic backsliding" (Bermeo, 2016, p. 5) in terms of the erosion of norms concerning free and fair elections, separation of powers, protection of minority rights, or independent media or civil society (Krastev, 2017;Márquez, 2017). Moreover, in 2016, United States voters elected a populist president whose governance practices tend towards authoritarianism (Balkin, forthcoming;Dionne, Ornstein, & Mann, 2017).
Of the many techniques that authoritarian and proto-authoritarian governments employ to wield power-including the "personalization" and executive "concentration of power" (Márquez, 2017, pp. 62, 66;Krastev, 2017), corruption, election rigging, and at times "violent repression of" dissent (Levitsky & Way, 2010, p. 8)-some concern the shaping of policy information disseminated to the public. One set of these tactics involves the creation of false or misleading policy information (Márquez, 2017;Merloe, 2016;Pomerantsev, 2016) that erroneously depicts reality as congruent with the authoritarian ruler's worldview. Another is the use of policy frames that exploit citizens' fears and prejudices, especially divisive frames that characterize politics as a struggle between a virtuous majority and minorities who threaten the social order (e.g., Bolonyai & Campolong, 2017;Ferguson, 2014;Krastev, 2017;Offe, 2017;Van Dijk, 2000;Wodak, 2015).
A key consequence of the use of such tactics is a decline in the quality of publicly available policy information, which hinders citizens' ability to exercise meaningful 1 In this essay, I follow political scientists who conceive of a spectrum of regime-types having poles of full democracy and full authoritarianism (e.g., Cheibub, Gandhi, & Vreeland, 2010;Freedom House, 2017;Levitsky & Way, 2010;Marshall, Gurr, & Jaggers, 2017;Pemstein, Meserve, & Melton, 2010; for an overview, see Clark, Golder, & Nadenichek Golder, 2017). Many such accounts posit a list of attributes of democracy-often derived from Dahl (1971)-and place regimes further toward the authoritarian pole the fewer democratic attributes the regimes exhibit. For example, Levitsky and Way (2010, p. 7) categorize as full democracies regimes that hold "free" and "fair" "elections," protect "civil liberties," and provide a "level playing field" for political competition; as "competitive authoritarian" regimes those that hold competitive elections but violate one or more of the other criteria to some degree; and as "fully authoritarian regimes" those that lack competitive elections and substantially violate the other criteria. I use "protoauthoritarian" to denote regimes-described by scholars as "competitive authoritarian" states (Levitsky & Way, 2010, p. 7) or "illiberal democracies" (e.g., Van De Walle & Smiddy Butler, 1999;Zakaria, 2003)-that fall between full democracy and full authoritarianism on a spectrum of state attributes. control over the state. In addition, these manipulative information strategies convey a symbolic message about the relationship between the state and members of the political community: that those members are subordinate to, and are effectively controlled by, the state (Wodak, 2015).
Nonetheless, practices of deliberative democracy (Fishkin & Mansbridge, 2017;Gastil, 2008;Nabatchi et al., 2012) have the potential to counter the effects of authoritarian information strategies. Interventions involving deliberative minipublics-groups of citizens who are generally representative of the populace, engage in facilitated deliberation about policy, and often distribute the results of their deliberations to the public (Fung, 2007;Grönlund, Bächtiger, & Setälä, 2014;Smith, 2009)-are likely to be particularly effective. The reason is that mini-public processes are designed to provide high-quality policy information to the minipublic members as the basis for their deliberations, and, with respect to minipublics that publicly distribute the results of their deliberations, to furnish highquality policy information to the populace.
Further, a new framework for measuring information quality in deliberative processes (Richards, 2017) enables deliberative practitioners and scholars to evaluate the quality of both types of policy information related to mini-publics. That framework also permits comparison of deliberative information to other kinds of policy information-including information emanating from authoritarian or proto-authoritarian states. In an authoritarian setting, this new framework for assessing deliberative information quality allows scholars, deliberative practitioners, and citizens to compare meaningfully the quality of policy information from state sources with the quality of such information distributed by processes of citizen deliberation.
Moreover, the procedures of deliberative mini-publics symbolically depict a relationship between citizens and the state in which citizens act as free and equal members of the political community who exercise sovereignty over the state. That depiction offers an alternative to the vision of the polity expressed by authoritarian information strategies. What's more, deliberative mini-publics can be implemented alongside activist strategies employed to challenge authoritarianism. This essay begins with a review of scholarship concerning the relationship of authoritarianism to deliberation, with a focus on the role of information in deliberation. Next, I describe information strategies employed by authoritarian and proto-authoritarian states. I then explain informational practices of deliberative mini-publics. Next, I present the new framework for evaluating the quality of deliberative information, and demonstrate its application with evidence from a study of policy evaluations produced by mini-publics. I then discuss how the informational techniques of mini-publics can improve the quality of policy information available to citizens governed by authoritarian or proto-authoritarian states, and how the new framework for evaluating the quality of policy information can demonstrate that quality. The implementation of mini-public deliberation in authoritarian conditions is discussed. Finally, limitations of this approach are set out, before the essay concludes.

Deliberative Information Under Authoritarian Rule
The relationship between deliberation and authoritarianism has been addressed by numerous scholars (e.g., He, 2014;He & Warren, 2011Leib & He, 2006;Pirsoul, 2017;Saito, 2008;Zhou, 2012), several of whom have discussed the role of information. Some of this literature emphasizes the capacity of deliberative information to foster learning. For Dong and Shi (2006), information is a focus of the Chinese citizen-deliberation forums called "Democratic Talk in All Sincerity" (p. 218), which enable citizens to gather policy information from officials and allow officials to collect accurate information about public opinion on issues and policies. Pirsoul (2017) argues that, if practices of public deliberation were permitted in Middle-Eastern authoritarian states, "[t]his epistemic value" of deliberation could improve public understanding of stigmatized minority groups, such as Shias, and undermine those regimes' divisive policy frames (p. 17). For Saito (2008), access to policy information during Ugandan local-council deliberations can foster the development of citizens' "civic values" and encourage the fashioning of "collective solutions" to policy problems (p. 177). He's (2006a) survey research showed that Chinese citizens associated forums for citizen deliberation with an entitlement to acquire policy information, the ability "to learn about the perspectives of" other citizens, and the "right to monitor" state action (pp. 189-190). Thus, in some authoritarian settings, citizens link deliberative information not only to the learning that is central to epistemic accounts of democracy (e.g., Bohman, 2006;Estlund, 2008;Landemore, 2013), but also to the accountability at the core of monitory theories of democracy (e.g., Keane, 2009).
The quality of deliberative information under authoritarianism has also been addressed in earlier scholarship. For example, He (2006b) found that where the state controlled deliberative processes, nearly half of Chinese citizens surveyed reported "speak[ing] only partial truths" during deliberation for fear of "the consequences of what they say" (p. 139). Consequently, He (2006b), colleagues (2010), andRosenberg (2006) recommended that processes of citizendeliberation in China use non-state organizers, moderated group discussion, private completion of participant surveys, and similar techniques-all of which tend to boost information quality-in order to encourage honest communication during deliberation. Discussing results of a 2005 Deliberative Poll in China that employed such techniques, Fishkin et al. (2010) reported that public officials had remarked on the high quality of the information yielded by that deliberative process.
The quality of policy information in authoritarian states such as China is degraded, not only by the distribution of "selective and incomplete" information, but also by the high cost of technology such as "internet subscription[s] and cable fees" that are beyond the reach of many citizens (Tan, 2006, p. 199). Tan (2006) proposed changing the design of village-level citizen-deliberation processes in order to improve information quality, such as by adding sessions in which policy information is provided to citizens by officials or well-informed village elites, and in which citizens share with officials the citizens' local knowledge of social problems. Zhou (2012) notes that such design changes have been implemented in processes such as "citizen evaluation meetings" in China (p. 8).
Another line of inquiry has explored cultural factors that are consistent with deliberation in some authoritarian states. Thus, Dryzek (2009) argued that deliberative practices have the potential to flourish under authoritarian rule informed by ideologies of Islam, with its tradition of "consultation" in governance, and Confucianism, due to the latter's principle "of reasoned consensus" (p. 1396; see He, 2014). Yet Dryzek (2009) acknowledged features of societies under such rule-such as the roles of "enclave deliberation," "religious fundamentalism," and "ideological conformity" (pp. 1396-1397)-that can restrict the public dissemination of high-quality information generated in deliberative processes.
Further, some scholars have identified civil-society organizations as integral to nurturing deliberation under authoritarianism. According to Leib (2006), "nongovernmental institutions" in authoritarian states such as China contribute to policy deliberation in part by "filtering information" about issues and policies and then basing their policy-debate interventions on such information (pp. 123-124). Yet this role may go unfilled in authoritarian states that suppress civil-society organizations (Márquez, 2017;Merloe, 2016).
Another function of deliberative information emphasized in earlier scholarship is that of enhancing the quality of policy. Xu (2006) contended that the high-quality information arising from public deliberation in authoritarian states such as China can enable the state to correct policy errors and yield improved policy outcomes. This governance-enhancing function of deliberative information also informs a key tension in this literature concerning the democratizing potential of deliberation.
Although citizen deliberation in authoritarian regimes may foster democratization-in part by enhancing the democratic "capacity" of individual citizens and governance institutions (Dryzek, 2009(Dryzek, , p. 1379)-the high-quality information generated by deliberative processes is at least as likely to entrench authoritarianism by improving the policies and bolstering the legitimacy of autocratic states (He & Warren, 2011. Having surveyed relevant prior literature, I now describe information strategies of authoritarian and protoauthoritarian states.

Authoritarian and Proto-authoritarian Information Strategies
The public-information component of authoritarian and proto-authoritarian regimes' governance strategies tends to include both the dissemination of false or misleading policy information and the use of divisive policy frames that play upon citizens' worst instincts. With respect to false or misleading information, examples include the prodigious lying about policy matters by U.S. President Donald Trump and his spokespersons (Dionne et al., 2017;Leonhardt & Thompson, 2017), and the dissemination of conspiracy theories by authoritarian regimes in Hungary, Venezuela, and Russia to explain adverse economic circumstances or foreignpolicy developments (Bolonyai & Campolong, 2017;Márquez, 2017;Pomerantsev, 2016). Also contributing to the misleading character of publicly available policy information in authoritarian states is government suppression of independent sources of such information, through state control of mass media and the squelching of civil-society organizations (Levitsky & Way, 2010;Márquez, 2017;Merloe, 2016;Pomerantsev, 2016). 2 Beyond distributing false or misleading policy information, authoritarian and protoauthoritarian regimes regularly frame policy issues in divisive terms (Bolonyai & Campolong, 2017;Ferguson, 2014;Wodak, 2015). These "polarising frames" (Calvert & Warren, 2014, p. 208) are manipulative (Bolonyai & Campolong, 2017) in that they purposefully play upon citizens' fundamental anxieties, "inflam[ing] the passions of" the majority of citizens (Dryzek, 2009(Dryzek, , p. 1381) in a "politics of fear" (Wodak, 2015, p. 2).
Many authoritarian and proto-authoritarian leaders apply these divisive frames to the issue of immigration. For example, Hungary's prime minister combines "discourses of fear" stigmatizing immigrants as a threat to the nation, with defensive language invoking national and cultural identity, in a use of the "topo[s] of threat and savior" (Bolonyai & Campolong, 2017, pp. 255, 257). Authoritarian leaders heighten the provocativeness of these frames with inflammatory metaphors linking immigrants to "invasion," "flood," "parasites," and "disease" (Wodak, 2015, pp. 56-57, 60).
In addition to immigrants, Jews are frequently targeted by authoritarian leaders' divisive policy frames. In Europe, right-wing populist officials regularly employ overt and veiled anti-Semitic rhetoric (Wodak, 2015), and leaders of Poland's ruling Law and Justice party have publicly praised demonstrators advocating anti-Semitic policies (Charnysh, 2017). In the U.S., Donald Trump's presidential campaign disseminated anti-Semitic messages such as a Star of David superimposed on a field of dollars, as well as images of George Soros, Janet Yellen, and Lloyd Blankfein accompanied by statements concerning "those who control the levers of power" and "global special interests" (Burns, 2016;Conway, 2016;Haberman, Barbaro, & Mahler, 2016;Marshall, 2016).
Stoking fear and polarizing political opinion are not the only functions of manipulative authoritarian policy frames. As the savior and father/children topoi suggest, those frames also have a symbolic function: that of depicting the political community in hierarchical terms, as a relationship between a dominant leader and a subordinate, dependent populace.

Informational Practices of Deliberative Mini-Publics
By contrast with authoritarian regimes, processes of citizen deliberation, and especially mini-publics, feature characteristics that tend to increase the quality of publicly available policy information. In a mini-public, a (generally) representative group of citizens gathers to deliberate on and evaluate social problems, policy proposals, or political candidates, and, frequently, publicly distributes their evaluation in order to encourage the populace to deliberate (Fung, 2007;Ryan & Smith, 2014;Setälä, 2017).
The first set of mini-public features that tend to heighten information quality are focused on the validity of information. In many mini-public processes, participants' key sources of information consist of "balanced briefing materials" (Fishkin, 2009, p. 53) based on sources recommended by experts, as well as testimony of experts and stakeholders subject to examination by participants (Hendriks, 2005;Knobloch et al., 2014). Participants during deliberation then further vet information from multiple sources for consistency Pennington & Hastie, 1992). Moreover, design features of mini-publics, such as facilitated small-group discussion and procedures that foster empathy among participants, tend to minimize bias during deliberation (Morrell, 2014). In addition, in minipublic processes that distribute policy information to the populace, such information is often further fact-checked before distribution (e.g., Gastil et al., 2016). Thus information processed by mini-publics tends to be high in accuracy (Fishkin, 2009;Fuji Johnson, Black, & Knobloch, 2017).
Second, participants in many types of mini-publics are randomly selected from the populace (Bächtiger, Setälä, & Grönlund, 2014), and are thus likely to be perceived as representative of the public. Many mini-public processes also feature procedural safeguards to keep interested parties from influencing deliberations (Vidmar, 2000;Vidmar & Hans, 2007), while mini-publics' information-vetting signals that minipublic participants are highly informed and competent to make policy decisions (Richards, 2017). Accordingly, the policy information vetted or created by minipublics is likely to be deemed credible by the public (Warren & Gastil, 2015).
Moreover, in mini-public processes that randomly select participants from the population, participants are likely, on average, to think, speak, and act like the majority of citizens, and to be intimately familiar with the communicative practices of, and typical methods of processing policy information employed by, their fellow citizens (Richards, 2017). Thus mini-public participants, when creating policy information to be distributed to the populace, are likely to select information that is useful to ordinary citizens (Richards, 2017(Richards, , 2018. Those participants are also likely to use techniques of phrasing, formatting, and organization that render such information accessible and relevant to most citizens and readily applicable by most citizens in their political decision making (Richards, 2017). Consequently, minipublics are likely to produce policy information that most citizens consider readable, relevant, and useful to their voting decisions.
Mini-publics not only aim to improve the quality of the content of policy information; they also strive to contextualize policy information in ways that discourage divisiveness and encourage deliberation (Kadlec & Friedman, 2007). Mini-publics often frame policy information in terms of three or more policy choices, each related to a vision of the future-a long-term policy goal-and related values (Fishkin, 1997;Kadlec & Friedman, 2007;Melville, Willingham, & Dedrick, 2005). Proposed policies "are not mutually exclusive," and the frames and values contextualizing those proposals concern "issues" rather than "personalities[,] … partisan divisions" (Melville et al., 2005, p. 43), or group identities (Calvert & Warren, 2014). Thus those policy proposals, frames, and values are aimed at de-polarizing policy discussions. Moreover, descriptions of issues in mini-public information resources also address likely actual outcomes of each policy choice, including adverse consequences (Melville et al., 2005). Minipublics that do not frame policy information by means of multiple future scenarios nonetheless often contextualize such information with descriptions of the goals and likely actual consequences of proposed policies Richards, 2017Richards, , 2018. Thus, the distinctive framing techniques employed in deliberative minipublics encourage citizens to transcend the polarizing terms of authoritarian policy discourse, reflect on the merits of social problems and policy proposals, and wrestle with policy trade-offs (Kadlec & Friedman, 2007;Mathews, 1999).
Further, these information-related procedures of deliberative mini-publics have a symbolic function . In contrast to the hierarchical vision of the political community expressed by authoritarian information practices, the information strategies employed in deliberative mini-publics depict the polity as a community of "free and equal citizens," who are themselves the source of political authority (Rawls, 2005, p. 369;Cohen, 1989). Mini-public procedures cast citizens as those who control the state (Pettit, 2012), rather than as those controlled by it.

Framework for Evaluating the Quality of Deliberative Information
What's more, a new framework allows scholars and practitioners to evaluate the quality of policy information more effectively than before (Richards, 2017). This framework enables the assessment of the quality of information used in or distributed by deliberative mini-publics, and the comparison of deliberative information to policy information disseminated by other sources (Richards, 2017), including authoritarian regimes.

The "CRAAV" Framework
Currently, scholars and practitioners of deliberation struggle to evaluate the quality of policy information involved in deliberative processes. The reason is that criteria for conducting such evaluations have never been consolidated into a single framework, but remain dispersed among many publications (e.g., Fishkin, 2009;Knobloch et al., 2013;Sprain, Carcasson, & Merolla, 2014). Thus, practitioners have trouble identifying both the standards for assessing the quality of deliberative information, and measures for applying those standards to real-world information resources.
A new approach addresses those problems by synthesizing the various standards for assessing information quality in deliberation into a coherent framework, and specifying qualitative and quantitative measures for applying those standards to information resources associated with actual deliberative processes. Informed by a review of the literature on the role of information in democratic deliberation, this framework posits five standards for evaluating deliberative information quality: credibility, relevance, accessibility, applicability, and validity ("CRAAV") (Richards, 2017).
One information-quality standard often found in the deliberative literature goes variously by the name of "recommending force" (Fishkin, 2009, p. 84), trustworthiness (Crosby, 2003), and legitimacy (Gastil, 2000). These terms all seem to designate the idea of credibility, or the extent to which citizens can believe and rely on information (Ryfe, 2006).
Scholars of deliberation have also discussed as a criterion of information quality the extent to which policy information relates to citizens themselves or their circumstances (Matthews, 1999) or to social problems or policy solutions (Melville et al., 2005). Those terms seem to refer respectively to the subjective and objective aspects of relevance, meaning the relationship between information and some entity (Harter, 1992;Richards, 2018).
Several scholars have also advanced as standards for assessing deliberative information notions such as readability  and clarity (Crosby, 2003). Those terms seem to designate accessibility, or citizens' ability to comprehend policy information.
Another information-quality criterion discussed by deliberative scholars is referred to in terms of necessity (Gastil, 2000), sufficiency (Knobloch et al., 2013), and usefulness (e.g., Kropczynski et al., 2015). These expressions all point to the idea of applicability, or the extent to which information can be employed to carry out tasks or meet people's needs.
With respect to the utility of this framework, scholars have developed quantitative and qualitative measures with which the CRAAV criteria can be applied to information resources involved in real-world deliberative processes (Gastil et al., 2016;Knobloch et al., 2014;Richards, 2017). Those measures can be employed by means of surveys, focus groups, interviews, content analysis, and readability scores (Richards, 2017).
For example, in a recent study, the CRAAV framework was applied to policy evaluations that were written by mini-public panels and distributed to wider publics to encourage deliberation among the latter (Richards, 2017). 3 Results showed that the mini-publics' policy evaluations garnered high marks on relevance, applicability, and validity, but yielded mixed scores on accessibility and credibility (Richards, 2017). These results enable practitioners to focus on the latter two characteristics when reforming that mini-public process in order to enhance the quality of policy information distributed by mini-public panels.
This study also demonstrated the use of the CRAAV framework for comparing the quality of information from different sources. In comparison with policy summaries written by government officials, policy evaluations produced by mini-public panels performed as well or better on all CRAAV measures except for credibility, with citizens generally expressing more trust in the official summaries than in the minipublics' evaluations (Richards, 2017). By contrast, when compared to policy arguments written by citizens who were not involved in formal deliberative processes, the mini-public policy evaluations fared well on all CRAAV standards other than accessibility: the language of the arguments produced by citizens in the absence of formal deliberation proved about two grade-levels easier to read on average than the mini-public policy evaluations (Richards, 2017). Thus the CRAAV framework can also be used to benchmark a deliberative information resource against information from other sources-such as authoritarian or protoauthoritarian regimes.
Moreover, because the measures used to apply the CRAAV criteria to deliberative information resources tap diverse aspects of each criterion, framework results allow practitioners to pinpoint particular facets of information quality for improvement. For example, in the aforementioned study applying the CRAAV framework to policy evaluations written by mini-public panels, qualitative interviews revealed that citizens' concerns about the credibility of the mini-public policy evaluations mainly involved uncertainty about the information on which mini-public panelists' based their deliberations, and whether procedural safeguards insulated mini-public deliberations from influence by particular interests (Gastil et al., 2016). Those concerns could be addressed through quite specific reforms, such as increased transparency (Karpowitz & Raphael, 2014) about the information resources furnished to mini-public participants and the procedures that protect those participants from outside influence.
The information strategies of deliberative mini-publics and the features and output of the CRAAV framework for evaluating deliberative information having been described, I now explore the potential for mini-publics and the CRAAV standards to improve policy information in authoritarian and proto-authoritarian states.

How Deliberative Information Can Improve the Quality of Policy Information
In what respects might policy information produced by mini-publics improve the quality of publicly available policy information in societies ruled by authoritarian or proto-authoritarian regimes? The first consequence concerns the overall accuracy of policy information available to the public. As noted above, several design features of mini-publics promote the production of information of high validity. Were mini-publics to be held regularly in an authoritarian state and to regularly distribute policy information to the wider public, the accuracy of the policy information available to that public would rise, all else equal.
History provides examples of public receptivity to more accurate information distributed by deliberative institutions under authoritarian rule or during transitions away from such rule. For example, publics proved receptive to democracy campaigners' and civil-society organizations' promotion of "truth" as a core value in struggles against the Communist regime in Czechoslovakia (Havel et al., 1985;Linz & Stepan, 1996). In South Africa, Chile, and other former authoritarian states, truth commissions provided accountability, more accurate historical records, and a measure of justice during those nations' transitions from authoritarian rule (Gutmann & Thompson, 2000).
Particularly regarding mini-publics whose participants are representative of their populations, citizens are likely to view policy information produced by minipublics as high in credibility (Warren & Gastil, 2015). Since policy information from mini-publics is likely to be expressed in language similar to that of ordinary citizens and to be designed to be useful to ordinary citizens in learning and making choices about policy issues, members of the public are also likely to judge such information as high in relevance, accessibility, and applicability (Richards, 2017).
Moreover, policy information distributed by mini-publics to the populace generally includes non-polarizing frames featuring descriptions of goals and probable, actual consequences of policy choices (Richards, 2017), as well as multiple outcome scenarios linked to values concerning issues and principles rather than divisive group identities (Calvert & Warren, 2014;Kadlec & Friedman, 2007). Citizens are likely to perceive contrasts in quality and framing between policy information from mini-publics and such information provided by the authoritarian state. Policy information disseminated by mini-publics may also encourage citizens to engage in reflection-including critical reflection (Curato, 2015;Dryzek, 2009;Setälä, 2017)-on public problems and policy proposals (Gastil et al., 2016;Niemeyer, 2014;Richards, 2018).
Such reflection by citizens may foster learning about issues and policies (Delli Carpini, Lomax Cook, & Jacobs, 2004;Fishkin, 2009) as well as about the range of their fellow citizens' views, arguments, and values concerning social problems and policy proposals (Cappella, Price, & Nir, 2002;Gutmann & Thompson, 1996;Mathews, 1999). Such insights could lead citizens to change their policy preferences (Cohen, 1989;Yankelovich, 1991). These insights could also encourage citizens to develop greater tolerance for, and grant legitimacy to, others' beliefs, attitudes, and values, and so promote social harmony (Dryzek & Niemeyer, 2006;Gutmann & Thompson, 1996). Thus, deliberative policy information featuring de-polarizing frames could thwart to some extent the divisive effects of authoritarian policy frames (Dryzek, 2009;Pirsoul, 2017).
Beyond these changes in knowledge, beliefs, and attitudes, citizens exposed to improved policy information from mini-publics may undergo various kinds of personal transformation. First, citizens may feel more empowered as civic actors (Pincock, 2012), since policy information and rich policy frames provided by minipublics should foster citizens' sense-making about policy (Baden & De Vreese, 2008;Gastil et al., 2016;Richards, 2018), which in turn should increase citizens' confidence in making political choices (Richards, 2018). Second, citizens' exposure to better policy information from mini-publics and citizens' ensuing experiences of political deliberation may spur the development of citizens' civic identity and values (Black & Lubensky, 2013;Saito, 2008). Third, since some citizens may employ policy information from mini-publics to monitor state action (He, 2006a;Keane, 2009), citizens' heightened sense of political efficacy may manifest in calls for greater state accountability (Dryzek, 2009).
Fourth, citizens' access to high-quality policy information from mini-publics and resultant experiences of political deliberation may kindle citizens' desire for a more deliberative politics (Burkhalter, Gastil, & Kelshaw, 2002;. That desire may in turn boost citizens' demands for better policy information from the state and mass media. Again, history furnishes an example. Members of Poland's Solidarity trade union, after many months of exposure to higher-quality policy information generated by their own deliberative processes, came to demand "credible and truthful information in the media" still controlled by the Communist state (Ekiert & Kubik, 1999, p. 36).
To the extent that officials of authoritarian regimes also access policy information produced by mini-publics, such information could encourage the crafting of policies based on more informed analysis and selected from a wider set of policy choices (Bohman, 2006;Landemore, 2013). Such policies are more likely to correspond to reality and therefore to be effective. Such policies are also more likely to accord with the range of citizens' actual preferences (Innes & Booher, 2003) and to be accepted by the public (He & Warren, 2011). Thus the improvement in publicly available policy information due to interventions by minipublics may lead to improved governance and better social and economic outcomes for citizens (Xu, 2006).
Considered together, all of these consequences may augment the "deliberative capacity" of an authoritarian or proto-authoritarian political system (Dryzek, 2009(Dryzek, , p. 1382. European history furnishes hopeful examples of such capacity's encouraging "a democratic replacement" for authoritarian rule (Dryzek, 2009(Dryzek, , p. 1388Habermas, 1989). Yet the recent history of China suggests that an authoritarian or proto-authoritarian regime may instead harness that capacity to improve governance, bolster its legitimacy, propagandize more effectively, and consequently reinforce its autocratic hold on the state (He & Warren, 2011. To what purposes can the CRAAV framework be put in an authoritarian environment? First, the CRAAV criteria can be applied to policy information disseminated by authoritarian or proto-authoritarian regimes to reveal the flaws in that information. Results of those evaluations can be publicly disseminated by a range of actors, including political activists engaged in resistance against authoritarian rule. Public access to such evaluations could weaken popular support for an authoritarian regime, encourage citizens to engage in critical reflection about policy information and public problems (Curato, 2015;Dryzek, 2009), and spur citizens' demands for alternative sources of policy information.
Second, the CRAAV standards can be used to assess the quality of information furnished to citizen-participants in deliberative mini-publics. Results of those evaluations have several functions. First, mini-public organizers can employ those results to identify the strengths and weaknesses of the information on which their deliberative processes are based, and to inform efforts to improve those processes. Second, those results can aid citizen-participants in mini-publics in weighing policy information during their deliberations. Third, comparing those results to findings from CRAAV evaluations of policy information distributed by the authoritarian regime can help mini-public participants deepen their understanding of the differences between information from authoritarian-state sources and information introduced to and vetted by a deliberative process. Fourth, organizers of minipublics can disclose the results in order to enhance the credibility of policy information that mini-publics distribute to the populace.
What's more, CRAAV criteria can be used to evaluate the quality of policy information that mini-publics disseminate to the broad public. Mini-public organizers can use those results to improve the design and conduct of mini-public processes. Moreover, members of the public can employ those results to judge the credibility of policy information distributed by mini-publics. Those citizens can also compare results of CRAAV evaluations of policy-information resources from mini-publics and from the authoritarian regime, to engage in critical reflection (Curato, 2015;Dryzek, 2009) in which citizens judge which sources of policy information warrant citizens' trust. To the extent that citizens deem mini-public information more credible than information from the authoritarian state, results of CRAAV evaluations could increase citizens' demands for policy information from mini-publics, which could lead in turn to increases in the quantity of high-quality policy information distributed by mini-publics to the populace.
Of course, access to results of CRAAV evaluations can also be used by public officials to compare the quality of state-provided information to that of information produced by mini-publics. Such evaluations could promote deliberative and democratic values among state officials (Dryzek, 2009), or encourage loyalists of the authoritarian regime to appropriate high-quality information from mini-publics in order to improve the quality of regime policies and thus heighten the legitimacy of the regime (He & Warren, 2017;Xu, 2006).

Implementing Mini-Publics in Authoritarian Conditions
Of the many challenges to implementing deliberative mini-publics in authoritarian or proto-authoritarian settings, two seem particularly pressing. The first concerns the potential of state actors to intervene in mini-public processes in a way that compromises the quality of the information yielded by those processes (He, 2006b). The second challenge involves the likelihood that political activists will play important roles in the organization or conduct of mini-publics; that possibility raises the vexed question of the relationship between activism and democratic deliberation (Levine & Nierras, 2007;Polletta, 2015).
Regarding the challenge of state intervention, as noted above, He (2006b) found that Chinese citizens participating in deliberative-governance processes admitted engaging in self-censorship when state actors were involved in those processes. Fishkin et al. (2010) succeeded in keeping state personnel on the sidelines of at least one citizen-deliberation event organized in China in 2005, and thereby demonstrated the possibility of safeguarding the informational integrity of deliberative processes in authoritarian environments. Whether deliberative practitioners in authoritarian and proto-authoritarian states can consistently keep state actors at a distance from mini-public processes remains to be seen, however.
Another aspect of the state-intervention challenge concerns deliberative practitioners' ability to distribute policy information produced by mini-publics to the populace. Evidence from China suggests that officials in some authoritarian states will allow such public distribution, provided that this distribution is limited to localities and that the content of the distributed information is restricted to particular, local policy questions and is subject to state censorship (He & Warren, 2011. In other authoritarian or proto-authoritarian states that are unwilling to distribute mini-public information through official channels, and that have suppressed independent media and civil-society organizations (Levitsky & Way, 2010;Márquez, 2017;Merloe, 2016;Pomerantsev, 2016), public dissemination of policy information from mini-publics may not be possible. Where an authoritarian state allows the operation of independent civil-society institutions such as labor unions or religious organizations, the example of Poland under Communist rule indicates that policy information from deliberative processes can be widely shared through those institutions' communication systems (Ekiert & Kubik, 1999;Payerhin & Zirakzadeh, 2006).
One more salient challenge to the operation of deliberative mini-publics in authoritarian or proto-authoritarian settings lies in the probability that activists will feature prominently among the administrators or participants in mini-public processes. This challenge arises from widely perceived conflicts in goals and tactics between deliberative practitioners and activists (Levine & Nierras, 2007;Polletta, 2015;Young, 2001). For example, activists prioritize prompt action to achieve goals whereas deliberative practitioners prefer extended, talk-based processes involving consideration of both objectives and means of policy. Deliberative practitioners often believe that sound procedures can enable equitable and bias-free deliberation, whereas activists assume that deliberation involves bias and inequality. Many deliberative practitioners also believe that they can create deliberative processes that minimize state involvement, whereas activists worry that state actors will coopt those processes. In addition, activists favor polarizing discourse to motivate citizens to act on established preferences, whereas, as discussed above, deliberative practitioners employ de-polarizing policy frames to encourage reflection, perspective taking, and possible changes in preferences (Fung, 2005;Levine & Nierras, 2007;Polletta, 2015;Young, 2001).
Nonetheless, several scholars have argued that activism is compatible with the practice of deliberation. Levine and Nierras (2007) contend that deliberation should be viewed as "one phase in a cycle of social change that also includes" political activism (p. 11). Fung (2005) endorses a model of "deliberative activism," in which citizens adhere to norms of deliberation while embracing political tactics and objectives appropriate to the levels of inequality and injustice in their society. For Polletta (2015), activists often contribute valuable information to deliberative processes. Moreover, history furnishes examples of the complementarity between deliberative practice and activism in authoritarian settings, notably during the era of Communist rule in Czechoslovakia (Bolton, 2012) and Poland (Payerhin & Zirakzadeh, 2006).

Discussion
This essay has proposed the use of deliberative mini-publics as a means of improving the quality of publicly available policy information in societies under authoritarian rule. Mini-publics have the potential to enhance the quality of policy information available to citizens through design features that result in careful vetting and processing of policy information. In contrast to the divisive policy frames employed by authoritarian regimes, mini-publics employ de-polarizing policy frames that encourage citizens to engage in reflection on issues while taking into account the perspectives and interests of their fellow citizens. I also contended that mini-publics' de-polarizing policy frames symbolically depict an egalitarian view of the political community in which free and equal citizens collectively exercise control over the state. This depiction contrasts with the hierarchical vision of the polity conveyed by authoritarian policy frames. I then addressed practical issues concerning the implementation of mini-publics in authoritarian contexts, including the risk of state intervention and the challenge of managing the relationship between activists and deliberative practitioners.
Notably, this essay foregrounds the role that a new model for evaluating the quality of information in deliberation-the CRAAV framework-could play in bolstering mini-publics' capacity to enhance the quality of publicly available policy information under authoritarian rule (Richards, 2017). The essay explains how results of the CRAAV framework can be employed, not only by deliberative practitioners to maximize the quality of the policy information used within and distributed by mini-publics, but also by citizens to make more informed judgments about the quality of the information on which they base their political choices.
Moreover, in this essay I acknowledge the risks involved in using public deliberation to boost policy-information quality in authoritarian states. The higher quality of policy information disseminated by mini-publics has the potential to nurture the "deliberative capacity" of citizens and state actors (Dryzek, 2009(Dryzek, , p. 1381) and so to encourage democratization (Zhou, 2012). Yet such information may also reinforce authoritarian regimes by increasing the effectiveness of their policies and enhancing their legitimacy (He & Warren, 2011Xu, 2006).
To be sure, the proposals set out above are subject to several limitations. First, information quality may not always be a primary goal of deliberative practitioners, even in authoritarian settings. For example, when deliberative practitioners wish to organize mini-publics on "highly divisive… moral [or] ethical issues" such as abortion, the objective of improving the quality of publicly available policy information may be eclipsed in importance by goals such as the maintenance of social stability (Bächtiger & Beste, 2017, p. 109). In future research, scholars should explore the complexities of conducting processes of citizen-deliberation concerning such issues in authoritarian contexts.
Second, mini-publics' capacity to enhance the quality of publicly available policy information under authoritarian rule depends in part on citizens' ability to process such information, which is in part a function of citizens' levels of education (Dryzek, 2009;Knight & Johnson, 1997;Rawls, 2005;Sanders, 1997). The greater the inequalities in education in a society, the less likely are mini-publics' informational interventions to be effective. The influence of educational inequality on citizens' capacity to process information from mini-publics in authoritarian or proto-authoritarian states could be investigated in future research.
Third, certain factors, other than state suppression of independent media and civilsociety organizations (Levitsky & Way, 2010;Márquez, 2017), complicate the process of distributing high-quality policy information from mini-publics to the populace in authoritarian settings. First, cultural phenomena in societies under authoritarian rule-including rigid ideological or religious views or patterns of "enclave deliberation"-may cut off large shares of the population from information disseminated by mini-publics (Dryzek, 2009(Dryzek, , pp. 1396(Dryzek, -1397. Second, authoritarian rulers' use of automated information-dissemination methods on a massive scale (e.g., Woolley, 2016) means that the policy information distributed by mini-publics may be drowned out in enormous waves of "computational propaganda" (Woolley & Howard, 2016, p. 4886). Third, a similar effect could result from the technological introduction, by foreign actors, of false or misleading policy information or divisive policy frames into public discourse within the authoritarian or proto-authoritarian state (e.g., Brady, 2016;Diamond et al., 2016;Hennessey, 2017). Future research could explore techniques by which deliberative practitioners in authoritarian or proto-authoritarian states could defend against these kinds of technological challenges.
Fourth, effective opposition to authoritarian rule requires a large share of the citizenry to engage consistently in critical reflection about the state and its policies, and yet higher quality information produced by mini-publics, by itself, is unlikely to influence citizens to adopt such a critical posture. Indeed, some deliberative mini-publics may have the effect of discouraging citizens' critical reflection on policy (Böker & Elstub, 2015;Lafont, 2015). According to Böker and colleagues, achieving widespread critical reflection among the citizenry requires changes to mini-public procedures and roles, including granting citizens more control over the design and conduct of mini-publics and ensuring that mini-public outputs have greater influence on policy (Böker & Elstub, 2015). Enabling more critical policy reflection among the public also depends on practitioners' nurturing a more deliberative "political culture" beyond the bounds of institutions such as minipublics (Böker, 2017, p. 32) in societies under authoritarian rule. The potential of these institutional and cultural reforms to influence the capacity of citizens of authoritarian states to engage in critical policy reflection can be investigated in future research. 4 Fifth, the use of the CRAAV framework for evaluating deliberative information, recommended in this article, may divert attention from contextual factors that could reduce the deliberative or critical impact of mini-publics in authoritarian states. These factors include circumstances of the organization of mini-publics-in which state actors or their proxies may exercise undue influence (Gastil, Knobloch, & Kelly, 2012)-other forms of co-optation of mini-public processes by authorities (Levine & Nierras, 2007;Polletta, 2015), and the "cherry-pick[ing]" of favorable mini-public "recommendations" by state actors (Smith, 2009, pp. 93, 176;Setälä, 2017). As noted above, Fishkin et al. (2010) have shown that, in some authoritarian jurisdictions under some circumstances, such challenges can be addressed to some extent through the use of organizational techniques and procedures that keep state actors at arm's length. In future research, the efficacy of such techniques and procedures should be explored in a greater variety of authoritarian settings. 5 Sixth, this essay does not fully address the risk that improved policy information generated by mini-publics will be appropriated by authoritarian states to enhance their legitimacy (He & Warren, 2011. Scholars have considered this risk in discussions of how mini-publics should be linked or "coupled" with formal policymaking institutions (e.g., Hendriks, 2016, p. 46;Setälä, 2017). To minimize the likelihood that the state will appropriate mini-public processes for its own purposes, Hendriks (2016) recommends that such linkages feature loose coupling to permit mini-publics to maintain independence from state institutions, and that these linkages enable "two-directional" or "multi-directional" communication allowing mini-publics and other entities to hold the state to account, as well as facilitating input by mini-publics to state policy decision-making processes (p. 56). Setälä (2017) further urges that mini-publics be implemented on particular issues or by non-governmental entities-such as parliamentary minorities or citizen groupsthat mini-publics enable interaction with "elected representatives" (p. 860) to require accountability (see also Neblo, Esterling, & Lazer, 2018), and that minipublic participants be given more influence over policy decision making through mechanisms such as "suspensive veto powers" (Setälä, 2017, p. 857). The effect of measures such as these on authoritarian states' ability to appropriate mini-public processes in order to enhance state legitimacy should be examined in future studies. 6 Seventh, while this essay has mainly concentrated on conditions in authoritarian states or states growing increasingly authoritarian, aspects of the approach set out in this essay may also be applicable to states transitioning from authoritarian to democratic rule (e.g., Curato, 2015;Dryzek, 2005;Elster, Offe, & Preuss, 1998;Heller, 2009;O'Flynn & Curato, 2015). The role of mini-publics and the CRAAV framework in enhancing the quality of publicly available policy information in such transitional states could be examined in future scholarship.
Further, should deliberative practitioners and their activist colleagues succeed in persuading authoritarian rulers to restore liberal democratic institutions, problems concerning the quality of publicly available policy information may persist, for at least two reasons. First, under liberal-democratic rule, traditional political parties are capable of disseminating substantial quantities of false or misleading policy information or polarizing policy frames, often through alternative media institutions (e.g., Guess, Nyhan, & Reifler, 2017;Mann & Ornstein, 2016;Mooney, 2012). Second, under liberal-democratic regimes, the quality of publicly available policy information may be compromised by attempts to reduce governmental data collection and analysis activities, in an effort to reduce the size and power of the state (Baumgartner & Jones, 2015). Deliberative scholars have begun to address these sources of the degradation of policy-information quality (e.g., Fishkin & Mansbridge, 2017;Miller & Leon, 2017), and the avenues of inquiry that they have opened should be explored more fully in future research.

Conclusion
In a growing number of authoritarian and proto-authoritarian states, governments often degrade the quality of publicly available policy information by distributing false or misleading information about social problems and by employing policy frames that polarize and spark fear among the populace. Practices of democratic deliberation have the potential to counter those authoritarian information strategies to some extent. By disseminating high-quality policy information to the public and contextualizing that information by means of policy frames that focus on issues, goals, and actual consequences of policies, processes of deliberative democracy can encourage reflection, mutual perspective taking, and empathy among citizens under authoritarian rule. In addition, a new framework for evaluating the quality of policy information can help citizens compare the quality of such information produced by the authoritarian state with the quality of policy information distributed by deliberative processes.
Processes of citizen deliberation have the potential to empower individuals and communities, even under authoritarian rule. Careful studies of the influence of deliberative processes on the quality of the information with which citizens make policy choices can suggest new approaches for realizing that potential.