Affinity Groups , Enclave Deliberation , and Equity

There is growing appreciation for the value of holding enclave dialogue and deliberation among marginalized peoples in their own affinity groups, as one stage in a larger conversation with the broader public or with public officials. These enclaves may be disempowered by enduring political inequalities, or in relation to a particular issue under discussion, or by the act of deliberation itself. Recent research and practice has demonstrated that well-structured dialogue and deliberation in enclaves can increase the inclusion, participation, and influence of members of society who have been excluded from public discourse, while avoiding the dangers of coercion, sectarianism, conformism, error, and illegitimacy. We review normative arguments and empirical evidence for the judicious use of affinity group enclaves to advance equity. We show multiple ways in which enclaves can be incorporated into democratic projects and processes that also include discussion among more representative samples of the public and with government. We offer design principles for affinity group discussion, which are illustrated by a recent series of dialogues on Facing Racism in a Diverse Nation, organized in the U.S. by Everyday Democracy. Finally, we discuss conditions in which enclave deliberation is most likely to be needed to create equity and suggest an agenda for future research.

Organizers of dialogue and deliberation employ several common strategies aimed at achieving equal inclusion, participation, and influence in civic forums.In forums that are open to all who want to join, each participant typically has an equal opportunity to attend, speak, and, if applicable, an equal vote.Forums that restrict participation to a sample of the public take further steps to practice equality.To achieve proportional representation of members of marginalized groups, organizers often recruit random samples or quasi-representative microcosms of the public, or recruit participants in part through networks of social service or civil society organizations (Leighninger, 2012).Some forums subsidize the costs of participationincluding information acquisition, time, and moneyby providing background materials about the issues, translation services, paying stipends to participants, and the like (Lee, 2011).To create conditions for equal participation and influence, facilitators set ground rules that encourage sharing of speaking time, respect for participants regardless of status or identity, and openness to a broad range of communication styles (Gastil & Levine, 2005).Each of these strategies seeks inclusion of the disempowered on more equal discursive terms than are often found in traditional public meetings, which can be dominated by more privileged citizens, or by officials or policy experts, and which are not designed to engender cooperative talk between community members as equals (Gastil, 2008).
While these are important strategies, they can be insufficient.Even forums that most aim to create representative microcosms of a community are hard pressed to include proportional numbers of community members who are disadvantaged by their education, income, race, gender, age, and political interest (Jacobs, Cook, & Delli Carpini 2009;Ryfe & Stalsburg 2012).Research often finds that despite organizers' best efforts, more privileged participantswhite, male, highly educated, and professionalspeak and influence decisions more than other participants (for summaries, see Black, 2012;Karpowitz & Mendelberg, 2014;Karpowitz & Raphael, 2014).Information, issues, and choices are often framed from the perspective of the powerful, even when presented as neutral or in terms of the "common good" (Young, 2000;Christiano, 2012).
In this article, we argue that incorporating stages of enclave discussion among disempowered people within larger political forums or processes can help move us beyond formal equality to achieve more substantively equitable dialogue and deliberation. 1Democratic theorists have long recognized that members of less privileged groups need to confer among themselves in civil society associations in order to contribute autonomously and effectively to discussion in the wider public sphere (Fraser, 1992;Mansbridge, 1996;Sunstein, 2000).We extend this insight to civic forums, processes, and institutions that aim to engage the whole community, maintaining that it would be better for equity, and ultimately for the quality of deliberation, to integrate opportunities for discussion among the least powerful.We argue that enclaves can counteract background inequalities among participants, the difficult dynamics of small group discussion among people of different statuses, and the dominance of associations and ideas of the privileged in the wider political system.And we believe these benefits of enclaves can be realized not just in advocacy groups or social movements, but in the institutions of democratic deliberation that have been developed over the past few decades, from innovative government-led methods of public consultation and stakeholder engagement to forums such as Deliberative Polls, Consensus Conferences, Citizens Assemblies, and the like.
In this light, enclave discussion is not necessarily an inferior version of crosscutting talk among a microcosm of the public, which is often the dominant ideal in deliberative democratic theory and practice.Indeed, enclaves are a feature of the traditional political institutions from which many contemporary civic forums draw metaphorical legitimacy and some design features.Consider the role of enclaves in the namesake institutions of our "21st Century Town Meetings," "Citizens Assemblies," "Deliberative Polls," and the like.Citizens who want to bring proposals to Town Meetings meet in like-minded groups to develop their arguments beforehand (Mansbridge, 1983).Members of legislative assemblies form caucuses based on common issue priorities and interests.Individual polling responses are shaped in part by our networks of family, friends, and others with whom we discuss politics.Like all forms of political communication, talking in enclaves poses some threats to good dialogue and deliberation, and we discuss ways of overcoming these dangers.But we start from a belief that enclaves are natural and necessary organs of healthy political institutions rather than warts on the body politic.
We begin by defining the kind of enclaves we are advocating, which share marginalized perspectives or social locations rather than essentialized identities, and the ways in which their members may be disempowered in deliberation among heterogeneous groups.Next, we draw on the empirical literature to describe the contributions that enclaves of the disadvantaged can make to creating more equitable and higher quality civic deliberation.We also describe the potential dangers of enclave discussionssuch as extremism, sectarianism, and conformismand why we see these dynamics as pitfalls that can be avoided by good deliberative design rather than as iron laws of political communication.For us, the key is to connect enclave deliberation among the marginalized well to other elements of the political system, and so we review several ways in which enclaves have been integrated productively into larger structures of democratic deliberation in forums and institutional processes that aim to represent a whole polity.To illustrate some more specific design principles for enclave deliberation, we present an extended example drawn from a set of dialogues in the U.S., entitled Facing Racism in a Diverse Nation.Finally, we discuss conditions in which enclave deliberation is most likely to be needed to create equity and sketch out an agenda for future research on the topic.

Enclaves
Cass Sunstein introduced the term "enclave deliberation" to describe communication among "like-minded people who talk or even live, much of the time, in isolated enclaves" (2002, p. 177).While Sunstein warned that the dangers of this kind of talk can lead people to adopt extreme views, he also noted that enclaves that are connected to the broader public sphere can contribute to a more equitable and pluralistic democracy, as we will address in more detail below.In the meantime, we can say that there is growing scholarly interest in theorizing the role of enclave deliberation in public life (Karpowitz & Raphael, 2014;Raisio & Carson, 2014;Setälä, 2014;Sunstein, 2000;Vasilev, 2013) and researching its contributions to diverse policy discussions, including about Internet access (Karpowitz, Raphael, & Hammond, 2009), elderly care (Baur & Abma, 2012;Jansen et al., 2015), health care research (Nierse & Abma, 2011), disabilities services (Raisio, Valkama & Peltola, 2014), and youth issues (Bulling et al., 2013).
However, because enclaves have been defined in multiple ways in the literature, we need to clarify what kinds of enclaves we want to integrate into deliberative forums and processes.First, our focus is on the value of enclave discussion in organized community forums and political processes, not among informal conversational partners (as analyzed, for example, by Mutz, 2006), or within advocacy organizations and social movements (e.g., Polletta, 2002), or political parties.Second, we are not suggesting that public forum organizers should form enclaves based on participants' common pre-discussion preferences about the outcome of the decision at hand (as in much of the experimental jury research summarized by Sunstein, 2000), given the concerns about group polarization discussed below.Depending on the goals of the forum, it may be useful to integrate enclaves of people who share policy discourses (as in the climate governance meetings analyzed by Stevenson & Dryzek, 2012), or shared values or beliefs (as in the church-based discussion groups studied by Neiheisel, Djupe, & Sokhey, 2009), or common cultural or linguistic identity (as in dialogues held among Belgian Flemish and Walloons researched by Steiner and Caluwaerts and reported in Steiner, 2012).However, among the most important goals of deliberation is to help participants re-examine their interests, beliefs, and values, and then to align them with specific policy preferences.Therefore, to avoid essentializing participants, organizers need to be cautious about attributing these characteristics to participants before discussion.
We see shared social perspectives, or structural locations in society, as the most legitimate basis for enclave deliberation of the disempowered in the widest range of contexts.Iris Marion Young defines perspectives as including the "experience, history, and social knowledge" derived from individuals' locations in social groups (2000, p. 136).Though perspectives may overlap with social identities to some extent, Young cautions against a crude essentialism that would reduce people's perspectives to a function of their race or gender, for example, and assume that this attribute consistently confers a common identity on the group's members, much less common interests, beliefs, or values.Considerable diversity of views and identities can be found among people of any social category.Yet Young also rejects a simplistic view of persons as unencumbered free agents.In her view, individuals are "positioned in social group structures rather than having their identity determined by them" (p.136).For Young, "social perspective consists in a set of questions, kinds of experience, and assumptions with which reasoning begins, rather than the conclusions drawn" (p.137).Because a "perspective is a way of looking at social processes without determining what one sees" (p.137), people who share a common vantage point often have different perceptions of their own interests, plural beliefs, and diverse policy preferences.Yet, "especially in so far as people are situated on different sides of relations of structural inequality, they understand those relations and their consequences differently" (p.136).Because of their different social positions, participants bring different experiential knowledge to deliberation, and that knowledge will be more difficult (but not impossible) for people in different positions to understand and appreciate.
To exemplify this notion of a social perspective as a shared way of looking that does not determine exactly what one sees, Young discusses the Pittsburgh Courier, a longstanding newspaper for African-Americans.The newspaper embodies an African-American perspective on the world by focusing on events and institutions in which blacks are the main actors.The Courier approaches local and national stories, including ones that are not only associated with African-Americans, from angles informed by issues and experiences of special importance to black Americans.Yet, within this shared perspective, the Courier also reports on controversies that dramatize blacks' conflicting views of what is in their best interests and what values they should embrace.Likewise, the opinion pages of the Courier include articles from a wide range of ideological positions, from libertarian to socialist, and from advocacy of economic separatism to integration.It is this kind of enclave dialogue and deliberation that we think can be a useful adjunct to civic forums of the broader public.
While scholars tend to talk about "enclaves," practitioners of democratic deliberation may be more likely to refer to "affinity groups," and we will use these terms interchangeably because we think scholars and organizers are speaking about the same things, and wish to speak to both groups.Of the many dictionary definitions of "affinity," we are not referring to an attraction based on natural likenesses, or a relationship based on common biology or origins, or intermarriage.Rather, we mean something like Young's "perspectives," and closer to the Latin root of affinis: adjacent, bordering, close in location but not identical, and sharing a sympathy for others who are susceptible to similar influences (www.merriam-webster.com/dictionary/affinity).

Power
Socially and politically privileged groups have more opportunities, preparation, and resources to engage in political discussion and action; indeed, many settings for public political discussion (such as public hearings, meetings, and media discourse) already serve as enclaves of the advantaged in society.In the interest of equitable rather than merely equal treatment of participants in dialogue and deliberation, we recommend forming supplemental enclaves of people who come from less powerful social locations.While political power can be multifaceted and fluid, we can still identify three kinds of participants that may be disempowered in dialogue and deliberation if compensatory measures are not taken.
The politically disempowered include people who have been disadvantaged over the long term in the larger political system outside the forum.These groups may be formally excluded from aspects of the political system, for example because they are denied voting rights or legal standing in judicial or administrative arenas.Such groups may be hampered by the lingering effects of past exclusion, or may lack the same access to resources for effective organization and action enjoyed by more privileged groups (often defined by income, education, and race or ethnicity).Members of disempowered groups typically have a weaker political voice, participating at lower rates than others in efforts to influence institutional and public policy making through organized lobbying, campaign contributions, legal advocacy, voting, direct action campaigns, and other methods of affecting policy (Verba, Schlozman, & Brady, 1995).As a result, government serves these groups' interests or preferences less well when they conflict with those of better organized groups who have greater resources to press their case (Bartels, 2008;Gilens, 2012).In sum, a group is disempowered if political or economic conditions are tilted against its ability to participate equally in political life, the group has a weaker political voice, or government is consistently less responsive to the group's policy preferences or interests.
Another group, the situationally disempowered, may be in a weaker position in regard to the issues on the agenda in a forum, regardless of the group's status in the political system, or of their socio-demographic attributes or privileges.These groups may find themselves in disadvantaged positions in the social, economic, or cultural realms, not only in politics oriented toward the state.In any of these spheres, the situationally disadvantaged encounter conditions that limit their ability to participate, speak, and be heard on equal terms.For example, in relation to freedom of speech, even the most affluent American youth are relatively disempowered by laws and customs that give schools great leeway to limit and sanction student expression, both on and off campus, in the interest of maintaining discipline, curbing bullying or hate speech, and the like (Papandrea, 2008).
A third kind of disadvantage may accompany and compound political or situational disempowerment: dialogue or deliberation itself may be a daunting arena for some people, especially if it is restricted to a style of reasoning that is combative, abstract, dispassionate, impartial, or that presumes specialist knowledge, or if it is conducted in a unfamiliar language.While research on public forums finds conflicting evidence of deliberative disempowerment, it is troubling to see multiple recent studies showing that even in fairly well-designed forums, people with less education or income, of lower social status, immigrants conversing in their second tongue, and women can be less likely to speak or influence others in mixed groups (Black, 2012;Gerber, 2015;Han, Schenck-Hamlin, & Schenck-Hamlin, 2015;Hansen, 2010;Himmelroos, 2011;Karpowitz & Mendelberg 2014).
While recognizing these inequalities of political resources, voice, and influence, we are mindful of the danger of reinforcing the marginalization of "the marginalized" by labeling them as such.The multiplicity of ways in which people can be disempowered that we have described above indicates that we see people as disempowered historically rather than permanently, situationally rather than universally, and in relation to more privileged groups rather than absolutely.Indeed, if we did not think that less powerful people retain considerable agency to strengthen their position, it would make little sense for us to propose enclave deliberation as one step toward achieving more equitable participation and influence in public life.In addition, the internal diversity of views and experiences found among disempowered groups, which should be shared in affinity group discussions, is a testament to the fact that people's identities are never defined only by the ways in which they are disempowered.People always bring much more to the table than the sum of their disadvantages.A crucial challenge for theorists and designers of enclave deliberation is to help participants identify these disadvantages, while simultaneously addressing people in ways that help them overcome marginalization from public discussion, rather than furthering their disempowerment.

The Virtues of Enclaves
With these definitions of enclaves and power, we turn to our main arguments for integrating enclaves of marginalized groups into broader public deliberation.We think that this strategy can help achieve more equitable recruitment, participation, and influence in civic forums, in ways that increase their deliberative quality and legitimacy.

Inclusive Recruiting
Many of the forums that have most successfully recruited members of marginalized groups offer them opportunities to confer in their own enclaves, not just in mixed groups.2Neighborhood forums held as part of Brazilian Participatory Budgets (Baiocchi, Heller, & Silva, 2011), Chicago community policing beat meetings (Fung, 2004), and urban redevelopment in New Orleans after Hurricane Katrina (Fagotto & Fung, 2006) suggest that the least powerful are most likely to attend forums when they can confer with others who are psychologically and physically proximate about how to address their most pressing needs, and have a direct channel to decision makers.Many forum organizers turn to voluntary associations to help recruit the disempowered because associations often provide the only trusted connections that disadvantaged people have to public life.When the marginalized are geographically clustered, as in low-income neighborhoods, forming enclaves can often be accomplished by holding meetings in these neighborhoods, rather than at a central site that is less convenient or more intimidating (City Hall, a university, and so on).

Equitable Participation and Influence
Within the forum, enclaves may also compensate for inequities of voice among participants.One need not make the mistaken assumption that all members of politically or deliberatively disadvantaged groups are ineffective deliberators to recognize some relevant aggregate group differences that may be reproduced in forums.Educational and income inequities appear to be the most powerful influences on how much experience adult Americans have with many kinds of political discussion.A nationally-representative survey finds that less educated people are less likely to engage in many kinds of "discursive participation," including discussions about politics in face-to-face public meetings, informal conversation, Internet discussions, and talk aimed at influencing others' political opinions or votes (Jacobs, Cook, & Delli Carpini, 2009).Low-income people, women, African-Americans, Latinos, and the young also take part in some of these kinds of political talk less frequently.In particular, each of these groups except youth is less likely to try to persuade others about a political issue, which is the stuff of many civic forums. 3Building deliberative experience in enclaves could help to strengthen marginalized people's ability to participate fully and persuasively in mixed-group discussions, enhancing consideration of views on their merits by decreasing communicative inequalities and status differences.
In most forums, deliberation occurs in small groups, the dynamics of which might be improved by enclaves for several reasons.First, enclaves can mitigate the pressures of tokenism.The voices of the least advantaged who attend forums may be muted by being dispersed across many small discussion groups.It can be a great burden to be the only member of a disadvantaged group in deliberation.To be a token is often to feel hyper-visible and constantly scrutinized, uncomfortably different from others, the subject of others' generalizations about one's group, and the need to contend with others' expectations that one speaks for or represents one's group (Niemann, 2003;Saenz, 1994).Tokens often feel they must devote much attention to managing impressions of themselves, adjusting their speech and behavior to guard against criticism from their in-group for betraying their "own kind" and to preclude attacks from out-group members for confirming negative stereotypes about their group.In reaction to these stresses, tokens may isolate themselves or disassociate from social interactions in which they are physically present.Deliberation is a challenging form of political communication under any circumstances, but especially for participants from non-dominant groups laboring under the psychological conditions of tokenism.For example, with regard to gender tokens, experimental research on deliberation finds that women typically speak less, are less likely to mention issues of distinctive concern to women, are less likely to receive positive feedback from men, and are seen as less authoritative when they are the lone female member of their group, while there are no comparable effects of tokenism on men (Karpowitz & Mendelberg, 2014).This same research underscores the power of enclave deliberation to amplify women's voices, especially as the gender composition of the group interacts with the decision rule used to resolve issues (such as majority rule or unanimity).Karpowitz and Mendelberg's (2014) experiments show that women benefit in several ways when they deliberate amongst themselves.In women-only groups, regardless of decision rule, group talk times increase, women raise issues of distinctive concern to them often (though they also do so in groups where women predominate but men are also present), they are less sensitive to critical or negative comments from others, they generate a group norm that is especially supportive and inviting, they articulate a preference for more generous support to the poor, and the group makes decisions that are in line with those more generous preferences.Put differently, women are able to generate outcomes that better match their preferences when they gather in enclave groups.In contrast, women's voices and influence are most diminished when they are in the minority and majority rule is used to make decisions.Women who benefit from enclaves the mostthat is, those with the greatest gains in efficacy and participationare the women who enter enclave discussion with the lowest levels of confidence in their ability to participate in group discussion (Karpowitz & Mendelberg 2016).
These findings hold several implications for civic forums, in which deliberation is typically designed to occur in mixed-gender groups.One is that if forums take no steps to correct for the deliberative disempowerment of women in mixed groups, women are likely to continue to be less frequent and influential contributors.Another is that groups that do not use unanimity or that lack female majorities are likely to arrive at conclusions that are less favorable to equalizing distribution outside the forum.However, there are guidelines that could promote gender equity, which can be tailored to the forum's contribution to the larger political context: To avoid the maximum inequality, avoid groups with few women and majority rule.To minimize male advantage, assemble groups with a supermajority of women and use majority rule.To maximize women's individual participation, gender homogeneous groups are best.(Karpowitz & Mendelberg, 2014, p. 141) While research on real-world forums sometimes finds more equitable participation or influence by women, the less educated, and people of color, the more optimistic studies often do not measure the potential effects of group composition and dynamics; most of the hopeful research is focused on a single forum design, the Deliberative Poll, which does not require participants to come to a group decision; and the studies generally do not show whether more and less privileged participants agree that their participation and influence were equitable (Karpowitz & Raphael, 2014, pp. 134-136).

Deliberative Quality and Legitimacy
In addition, it may be easier for people to reconsider their views in peer groups than in more diverse settings.Discussions of differences among perceived peers may help to counteract the "false consensus effect," which leads individuals to overestimate their peers' agreement on a particular preference (Quattrone & Jones, 1980).This misperception can lead individuals to swallow their reservations and continue espousing positions mistakenly attributed to their group, creating a cycle of group adherence to these positions.Deliberation within enclaves can offer safe spaces for individuals to rethink their positions as they discover that "people like me" can and do think differently about politics (Kahan, Slovic, Braman, & Gastil, 2006).Experimental research on the power of social group cuing finds that even participants who employ systematic reasoning are still more likely to accept arguments presented to them by a source with whom they share a perceived group allegiance.For example, experimental participants have been found to be more favorably disposed to a policy position if they think it is proposed by their own political party than by an opposing party, even when the proposed policy was the same (Cohen, 2003).In well-facilitated enclaves, people should be able to hear multiple arguments from people with whom they share a perspective, interfering with any uniform group cues about how "people like us" agree on an issue, so that each participant can exercise her own agency.
Enclaves can also help non-dominant participants exercise their collective agency by helping to ensure that views that are often marginalized in the public sphere get a full and fair hearing in the forum.These views may be omitted from the initial issue framing, arguments, and menu of policy options provided to participants in briefing materials, hearings, or agendas, leaving less powerful participants unaware of these ideas, or swimming against the tide when trying to raise them.This risk is highest in forum designs that pre-define a fixed range of policy choices, such as some Citizens Juries.Even in more open designs, such as Consensus Conferences, participants enter having been exposed to dominant ideas more than marginalized ones in public discourse.It is extremely difficult for even the most experienced and even-handed organizers to anticipate all of the views and policy choices that could be aired in a forum, including ones that might be raised by large numbers of participants.This is the case even for expert organizers and especially for novices.While it is much easier to critique a forum than to organize a good one, it is critical to consider how limitations on participants and ideas can undercut the forum's ability to take up policy proposals made outside the forum by less powerful voices.
Incorporating enclaves at some point in the deliberations can be a valuable way to identify and compensate for these limitations of perspective and knowledge.For example, two of us saw the value of enclave deliberation through our own experience in organizing and evaluating a forum on municipal broadband.This was an enclave forum that convened only citizens from social groups with least access to high speed Internet service.Participants were asked to deliberate over whether a proposed municipal broadband system in Silicon Valley was a desirable option for addressing the digital divide, and, if so, how the project should be designed to meet the needs of the underserved.The project aimed to inform officials, who were in an early stage of planning such a system, and who were interested in considering how it could extend service to the unconnected.In drawing up the agenda and briefing materials, we failed to anticipate and address an issue that ended up being among the most important to participants: protecting their privacy and security online.The issue was introduced by a blind participant who told the story of having his identity stolen.Others chimed in with similar concerns, extending the conversation to fears of lost privacy on the Internet.It is unlikely that the issue would have risen to the top of the forum's agenda had the participants been dispersed in many small groups with more privileged deliberators who did not share the experience of people without Internet access, anti-virus software, and knowledge of the major security threats to avoid online.In an enclave, participants' conclusions on this issue ended up forming a major, and unexpected, part of the group's recommendations.The personal experiences of all members of the group help them to link issues that the organizersincluding a university professor who regularly taught about the digital divide, online privacy, and securityfailed to connect beforehand.This example also illustrates the power of enclaves to supplement expert knowledge with the experiential, local, or situated knowledge of the least advantaged.Too often, the production of expert knowledge is least responsive to the economically and politically disempowered because they are considered weak markets for sales, votes, or grant money (Christiano, 2012).Enclave forums of the disempowered may be among the best possible places to correct this imbalance.Those who share a weaker position can identify common concerns and questions that they want to answer more effectively than in mixed forums.These citizens can subject political and scientific experts to public questioning from people whom specialists have had little professional incentive to consult, given the institutional conditions in which most experts work.At its best, the deliberative turn in public consultation by governments is motivated by genuine respect for the situated understanding of education by students and parents, of health care by patients and family caregivers, of social services by aid recipients, and of planning and land use by community residents (Nabatchi & Leighninger, 2015).These constituencies are sometimes convened in enclave forums.
As these examples suggest, incorporating deliberation among the disempowered within broader civic forums or policy-making processes can broaden the range of voices and views, improving the quality of deliberation and decision making.As Cass Sunstein observes of enclaves in associations and movements: A certain measure of isolation will, in some cases, be crucial to the development of ideas and approaches that would not otherwise emerge and that deserve a social hearing.Members of low-status groups are often quiet within heterogeneous bodies, and deliberation in such bodies tends to be dominated by high-status members.Any shift . . .that increases the number of deliberating enclaves will likewise increase the diversity of society's aggregate "argument pool."(2000, p. 105) In this vein, enclaves can help enlarge our definitions of the common good, which is often defined ideologically in ways that serve the powerful (Young, 2000).Civil society enclaves have helped dissenters to challenge hegemonic common sense that once defined what is best for allthat slaves are children who need the protection of their masters; that voting is best handled by responsible, white, propertied males; and the like.Within the forum, enclaves might also help challenge and supplement partial definitions of the common good.
To summarize, making space for enclave discussion can enhance inclusion of non-dominant groups by strengthening their motivation to attend forums.Enclaves can support fuller participation by the marginalized by developing their deliberative capacities, overcoming tokenism, and providing safe spaces to reconsider their views.Groups of non-dominant deliberators may be more likely to consider unconventional arguments and policies that would increase equity outside the forum, question hegemonic definitions of the common good, and fill in the gaps in forum organizers' and experts' perspectives.If connected to crosscutting discussion among the broader public or among officials, enclaves can foster consideration of additional knowledge and options in these forums.

Avoiding the Dangers of Enclaves
While we are mindful of the risks posed by enclave deliberationincluding group polarization, sectarianism, conformism, error, and illegitimacywe think they are neither inevitable nor unavoidable.

Coercion and Group Polarization
Experimental research in social psychology sometimes finds that homogeneous groups are prone to group polarization, in which "members of a deliberating group predictably move toward a more extreme point in the direction indicated by the members' predeliberation tendencies" (Sunstein, 2000, p. 74).Groups seem to be most likely to go to extremes if they meet all three of the following conditions: they see themselves as sharing an identity that is made salient to them during deliberation (such as party affiliation, race, gender, or profession), they meet regularly over time, and they insulate themselves from competing views (Abrams et al. 1990;Mercier & Landemore 2012).Yet the critical factor is that members of the group enter deliberation tilting toward a particular decisional preference, such as a guilty or innocent verdict in a mock jury, or a pro-life or pro-choice stance on abortion (Baron, 2005;Sunstein, 2000).
These conditions differ markedly from the kind of affinity group deliberation we are proposing.We have argued that this kind of deliberation is most legitimate when it assembles people with shared perspectivescommon structural locations in society that are less powerfuland least legitimate when it sorts people into groups with similar pre-deliberation preferences on the issue at hand.These enclave discussions should be one step in a larger public process that includes cross-cutting talk among participants from different social groups, experts, or elected officials.We are not suggesting that organizers of enclaves should encourage participants to hold similar policy views, or sequester them for long stretches of time, or even that organizers should exclusively appeal to a common identity among the group (more on this below).Indeed, to the extent that perspectives are correlated with preferences, well-designed enclave deliberation has the potential, as we highlight below, to reduce such correlation by helping participants come to recognize and appreciate the full diversity of views among them.
However, as Sunstein notes, even if group polarization occurs, it is not inherently undesirable.At any given place or time, a laudable position that is perceived as "extreme" by the majoritysuch as calling for the abolition of slavery in the United States in the early 1800s -may be more just than a "moderate" stance on the issue.Thus, to reject all instances of polarization within disempowered groups would be to adopt a conservative bias against innovative views or a centrist bias against "extreme" positions.
The quality of deliberation depends not on how many people change their opinions or in what direction, but on why they hold their positions after deliberation and whether those positions are reasonable.If participants seriously consider a broad range of reasons and conclusions, if they are well-informed by evidence, and if participants deliberate as equals, then it can be as legitimate for their discussion to reinforce or strengthen their pre-existing positions as to change those positions.The important question is not whether groups move further in a direction to which they were previously inclined, but whether they do so for bad reasons, such as fear of the social judgments of the group majority or a failure to consider minority viewpoints within the group, or whether members shift for good reasons, such as clarifying their views.We should only be concerned about coercive polarization, which arises from social pressure and consideration of a narrow range of arguments.The real danger is not group polarization but conformity to domination by the majority.
Fortunately, most studies of real-world forums find no strong evidence of group polarization, for any reasons.This is a repeated result of research on Deliberative Polls (Farrar et al., 2009;Fishkin 2009, 131-33;Luskin, Fishkin, & Jowell. 2002;Siu, 2008).But perhaps these forums avoid polarization because they engage microcosms of the public as a whole in cross-cutting discussions rather than in enclaves, and end by surveying participants individually rather than requiring them to make group decisions?
Yet studies of forums with enclaves also show little or no evidence that groups shift further in the direction to which they were initially inclined because of social pressure or limited argument pools.For example, in the consensus conference on high-speed Internet access mentioned above, which convened an enclave of people from groups with least access to broadband, researchers found that participants perceived greater diversity of views among the group as the conference progressed, yet were able to agree on some common policy recommendations (Karpowitz, Raphael, & Hammond, 2009).Another study found "considerable variation and internal critique" (Nierse & Abma, 2011, p. 418) among enclaves of people with intellectual disabilities asked to make recommendations for future medical research on their conditions.In another civic forum on nanotechnology, enclave groups of skeptics and enthusiasts about technology both arrived at similar policy recommendations (Kleinman, Delborne, & Anderson, 2011).In discussions among Dutch-speaking and French-speaking Belgians about how the state could accommodate each language group, researchers found no differences in how often participants acknowledged the merit of each other's views in monolingual groups (of Dutch and of French speakers) and multilingual groups (of Dutch and French speakers together) (Steiner, 2012).Research on a Finnish mini-public about immigration policy compared enclaves of pro-immigration, anti-immigration, and mixed-position groups, and found that participants as a whole shifted in favor of loosening restrictions on immigration during the forum, including members of the mixed groups, and both sets of enclave groups (Grönlund, Herne, & Setälä, 2015).While the pro-immigration groups polarized slightly, they do not seem to have done so for troubling reasons.All groups shifted in a pro-immigration direction, not just the pro-immigrant groups.And participants who shifted their positions most dramaticallytoward either extreme or toward the centerreported low levels of perceived social pressure to change their minds (Lindell et al., 2016).Taken together, these studies offer mounting evidence against fears of group polarization in well-structured and facilitated enclave deliberation, whether these affinity groups share common perspectives (such as living with disabilities or without Internet access), values and beliefs (about technology), identities (including language and culture), and even policy preferences (about immigration).

Sectarianism
There is also a danger that creating enclaves within the forum will breed sectarian thinking and selfish bargaining with those outside the group.But this critique runs the risk of assuming that arguments about various groups' interests are not a healthy part of democratic deliberation.We should not ask disadvantaged participants to put aside their own interests.The arguments of the marginalized about how their interests are affected by policies are important contributions to public understanding of those policies' fairness (Mansbridge et al., 2010).Indeed, if society did not regularly slight the interests of the disempowered, there would be less reason to convene them in their own affinity groups.As we argued above, it may be easier for marginalized people not only to identify the common challenges they face but also to reconsider their views in peer groups than in more diverse settings.At the same time, all of us should be encouraged to justify our interests in terms that link our sense of what is best for us to ideas of what is true, fair, or good for others who are also affected by the issue.Disempowered people need to take others' basic interests into account, both for the sake of persuading them expediently and as a matter of justice.This kind of public-minded reasoning is not at odds with groups expressing their interests as long as such expressions do not ignore the interests of other groups and the interests of the public as a whole.
Why not leave these kinds of discussions to advocacy organizations of the marginalized, which may appear to be their more natural home?While dialogue can be pursued productively in social movements, there are good reasons to supplement it with enclave civic forums.One is that most members of disadvantaged groups do not join most social movements, so forums can involve some people who would not take part in policy discussions otherwise.Second, competent civil society associations may not be ready to address some issues that are the topic of government consultations, planning groups, and other civic forums.Third, discovering self-interest and group interest happens in ways that are less prone to sectarian thinking in affinity group forums than in advocacy organizations.In forums, participants are more likely to be unorganized community members who are not initially united by commitment to a particular cause, less familiar with one another than they are in organizations, and who interact for briefer periods and are expected to abide by discussion rules established by forum organizers.The goal of many forums is to convene or make decisions on behalf of a social whole, rather than a part.These aspects of forums are likely to temper sectarianism, even during phases that involve affinity group discussion.Additionally, it may be that enclaves are more likely to be publicminded the more that the political system is responsive to them (Setälä, 2014).While there is always a danger that any kind of civic engagement may be coopted by leaders, affinity group forums that are connected well to other institutions can give the disempowered a reasonable belief that their voices will be considered by the rest of society rather than falling on deaf ears.

Error
Should we be concerned that citizens in enclaves will learn less than they would in microcosmic forums, where participants' views might be enriched by more diverse citizen perspectives?The Finnish immigration study mentioned earlier provides the only direct test of this claim of which we are aware.The researchers found no significant differences in information gains about immigration or general political knowledge between the enclave discussion groups (for or against immigration) and the mixed groups (Grönlund, Herne, & Setälä, 2015).Of course, disempowered people have valuable knowledge of their own to contribute, which may circulate best when the least advantaged speak among themselves.Not only are they likely to assemble their own experiential knowledge more effectively than in mixed settings, but they are also likely to formulate different questions for experts and officials, which can elicit information about the effects of policy options on less privileged sectors of society; both dynamics would contribute to the store of knowledge that can inform a larger deliberative process.

Illegitimacy
One of the main reasons why organizers gravitate toward microcosmic forums is that they can offer a symbolic representation of a polity.In contrast, enclave deliberation can appear partial in both senses of the wordpartisan and incomplete.It is difficult enough for most microcosmic forums to get their conclusions adopted by officials and institutions.How could enclaves possibly achieve the necessary legitimacy to affect the world beyond the forum?
As with all forums, the challenge is connecting affinity group discussions to power centers.As we noted at the outset, enclaves already operate effectively within some existing democratic institutions, including ones that involve the public.Consider the well-known Porto Alegre Participatory Budget, in which people deliberate in their neighborhoods before sending representatives to district and city level discussions.Budget priorities defined in the relative enclaves of each neighborhood must survive further rounds of deliberation at the district and citywide levels.While disentangling the policy impacts of civic forums from other factors in the political process is always difficult, Participatory Budgeting in Brazil has been credited with increasing the transparency of public spending, reducing corruption and dependence on government patronage, redistributing public investment to the most neglected neighborhoods, and increasing civic engagement, solidarity, and the perceived legitimacy of local government (Baiocchi, Heller, & Silva, 2011).

Integrating Enclaves
The advantages of deliberation among affinity groups of the marginalized may be realized, and the risks minimized, if it is part of a larger process of democratic deliberation.Theorists increasingly conceive of democratic deliberation as ideally infused throughout a systemic political process rather than as something that happens best at a single event that convenes a representation of the whole polity (Parkinson & Mansbridge, 2012).This shift helps us to imagine enclave deliberation of the least powerful as one moment in a larger political conversation.This should enable us to advance equity through deliberation that can stretch across forums and other institutions for representing public opinion or making decisions, rather than assuming that each forum should be designed as if it were solely responsible for these tasks.In this light, forums may not always need to strive for convening a proportional microcosm or representation of the whole polity, or of all who are affected by an issue under consideration, as long as the forum is connected well to other elements of the political system.Democracies rarely expect a single institution or event to represent public opinion or the decision of the whole polity, and so it should be for civic forums.
Enclaves could be integrated into existing deliberative events and processes in a number of ways, a few of which are illustrated in Figure 1.This is by no means an exhaustive list, which is beyond our scope here (and more detail on approaches to inclusion is provided in Karpowitz & Raphael, this issue.)However, we think these examples begin to show how affinity groups of the less powerful can be blended with deliberation among more heterogeneous and representative participants, and how this might take advantage of the promises of enclaves we have discussed while avoiding some of their dangers.
First, enclaves might be convened within large cross-cutting civic forums, such as 21 st Century Town Meetings, which break participants out into many small group discussions.Enclaves of marginalized groups might be asked to review whether the central issues of the forum have been named and framed in ways that resonate with the group's experience, to draw on participants' situated knowledge in discussing the impacts of social problems and policy proposals on the least advantaged, and to add unique policy proposals of their own.
Exchanges also can occur between enclave forums and cross-cutting forums.For example, in a unique Deliberative Poll about how Australia could engage in national reconciliation with its indigenous communities, initial meetings among the indigenous were held in multiple regions to inform policy proposals that were considered later in a single meeting that was nationally representative of all Australians (Issues Deliberation Australia et al., 2001).If the national meeting had been the only one, it might have included only a handful of indigenous participants because of their small slice of the contemporary population.This would have defeated the organizers' goals of convening indigenous and other Australians in small group deliberation because there would not have been enough indigenous to distribute into all of the groups, and those who participated would have been tokens.

Enclave Officials
Enclave in deliberation with officials as representatives of the public Enclaves can also interact with officials who represent the larger public, either in a single forum (as in the broadband Internet forum mentioned earlier) or as part of an ongoing political process (like the Brazilian Participatory Budgets and Chicago community policing meetings discussed above).Each of these examples succeeded at recruiting members of disempowered groups in part because they saw an opportunity to engage in consequential discussion with similarly situated community members that might influence officials.And each example helped limit group polarization by exposing participants not only to their affinity group's perspectives but also to officials' views and information.

An Example: Affinity Groups in the Facing Racism Dialogue
How might these enclave deliberations be organized most productively?We suggest a number of design principles, which were exemplified by a series of dialogues in the U.S. using a process described in Facing Racism in a Diverse Nation and Dialogue for Affinity Groups organized by Everyday Democracy (Abdullah & McCormack, 2008a;Abdullah & McCormack, 2008b). 4The use of these dialogue materials permitted participants to convene in multiple facilitated small groups over several two-hour sessions to build understanding and relationships, explore different sides of racism and inequities, locate common concerns, identify practical steps to address these problems, and make action plans for their communities.Some communities chose to include affinity group dialogues to give people from the same racial or ethnic group a way to talk about issues that affect their group before and after talking with participants from different backgrounds (see Figure 2).Organizers identified several purposes of the affinity groups.The affinity sessions aimed to develop participants' voices by gaining support and practice in talking about difficult issues; build relationships and trust; explore multiple ways to work with others; help participants obtain new insights into their own and others' beliefs; and "unpack our own 'baggage' before joining dialogues with mixed groups" (Abdullah & McCormack, 2008b, p. 4).
Facing Racism presented the affinity groups as one step in a larger discussion, not an isolated event.Briefing materials explained that "[t]hese are not intended to be used as stand-alone sessions," but "are designed to add to a community-wide dialogue on racism" (p.4).The sessions moved participants from making personal and group connections to the issue at the initial affinity session to preparing for mixed-group dialogue at the second session.After six sessions of discussion in heterogeneous groups, including identifying actions that participants might take with all members of the community, the affinity groups reconvened for a closing session to identify any additional actions they might want to take as a group.In this way, the dialogues allowed participants to consider their relationship to racism as individuals, as members of particular racial or ethnic groups, and as members of a wider community, and to consider actions that could be taken both as group members and as members of the general public.
The affinity sessions also encouraged participants to make their own sense of their social location, rather than imposing a single, stifling identity on participants.Individuals who considered themselves multiracial were invited to attend any group in which they felt they belonged, or to attend a group for mixedrace participants, which included unique discussion prompts (such as "how do we define ourselves?";"where do I belong?"; and questions about the experience of "passing" for a different race).The initial discussion prompts asked participants to describe their racial or ethnic background, the first time they realized they belonged to a racial or ethnic group, the first time they noticed people who were different from themselves, and then to discuss what racism meant to them by sharing a brief story of how it had affected their lives.
These broad questions also invited multiple issue framings and reflection on them.
To this end, briefing materials advised facilitators that [s]ome people may want to talk about the meaning of 'racism' and 'discrimination.'Let them talk about it the way they see it.There is no need for the group to agree about definitions, but you can point out common themes" (p. 7).

Addressing community change
After asking participants to reflect on some common concerns for their racial group, some of them suggested by the organizers and some by participants, facilitators asked people to prioritize these issues through discussion, identifying points of disagreement as well as agreement.This process helped to expose participants to diverse arguments and information, even before they joined mixed-race discussions later in the process.Additionally, when participants were asked to discuss their experiences "when you talk about racism with people from your same background," follow-up questions asked "where have you found support?" and "where have you found barriers?"(p. 7).In this way, facilitators encouraged exploration of commonalities and differences within the enclave, rather than idealizing the group and generating social pressure to identify shared norms.We have noted that enclaves can be safer places for individuals to reconsider their views when they allow members to see that people like themselves hold a variety of views about their own interests, beliefs, and policy preferences.
By leaving space for participants to add their own concerns to the agenda, the project allowed affinity groups to discuss emerging issues.Coercive group polarization may be less likely in enclaves that tackle issues that are new to public discourse (Karpowitz, Raphael, & Hammond, 2009).These issues are not easily incorporated into group members' existing political schemas, which are shaped primarily by direct experience and oft-repeated mass media messages (Graber, 1989).Because enclave members have not yet experienced these issues or heard much about them from the media, participants do not bring strong opinions about these concerns to the table.For example, some forums have convened "anticipatory publics" (Mackenzie & Warren, 2012) to help steer the development of new technologies and research agendas, such as genetics research, to which enclave forums of people of color have contributed unique policy perspectives based in part on their concerns about discriminatory uses of medical data (Bonham et al., 2009).
While Facing Racism employed ground rules for dialogues that are familiar in many facilitated groups, some of these guidelines are especially significant for enclaves, including "Speak for yourself, not for others" and "You can disagree, but don't personalize it.Stick to the issue.No name-calling or stereotyping" (p. 7).There were some less common rules of engagement as well, such as, "If you are offended or uncomfortable, say so, and say why" (p. 7).These agreements by the group are important to keep dominant speakers from imposing an identity on the group and setting themselves up as arbiters of which views and people are authentic members of the enclave.
community "get a fair chance to succeed" and "have equal access to services" (Abdullah & McCormack, 2008a, p. 30).Participants also discussed the pros and cons of seven approaches to creating change in their communities, from reforming institutions such as schools and banks) to self-reliance by people of color, engaging one another in justifying their preferred strategies in terms that could be persuasive to people of different backgrounds.
To avoid dominant members of the community imposing a single vision of change at the end of the project, Facing Racism organizers invited participants to imagine taking action both within their group and in more heterogeneous organizations.The third optional enclave session, held after the community-wide dialogues, offered participants an opportunity to reconvene in their racial and ethnic groups to compare the list of issues they had initially prioritized together with the top issues that emerged from the community discussions.Facilitators asked the affinity group to account for similarities and differences in these issue priorities, and, "given your experience in both groups, what do you think we need to focus on to improve the community?"(Abdullah & McCormack, 2008b, p. 19).This allowed participants to evaluate whether the full community's action agenda was sufficiently inclusive of historically marginalized voices and likely to be effective to warrant their participation.This approach also invited participants to think about how they might act as a group on behalf of more equitable visions or processes of change by asking "Is there anything that our affinity group wants to work on to address racism?" (p.19).
In both the full community sessions and enclave discussions, Facing Racism had participants prioritize the action steps proposed by the group and then volunteer to work on these actions.In both cases, group decisions were made by majority rule after deliberation but were non-binding: whether or not participants agreed with the final action steps chosen by the group, they could opt out of further efforts to pursue them.This raises the question of what decision rule is most appropriate to enclave deliberation.
Because there are trade-offs between decision rules, it is important for organizers to tailor the decision rule consciously to the goals of the affinity group forums and their relationship to the larger political process.For example, individual polling or majority rule by anonymous balloting may be most appropriate way to express an enclave group's conclusions if the main goal is to maximize individual freedom of thought among its members, avoiding social coercion within the group, whether by a domineering majority or a recalcitrant minority.Deciding by unanimity may serve other ends best, including the goal of rendering a single judgement by the enclave group, and therefore of influencing others (Karpowitz & Raphael, 2014).This could allow enclaves to act as a check on potential domination by more privileged members in mixed group sessions by expressing an endorsement of or dissent from conclusions drawn by the wider forum.
Unanimity is also most likely to yield a better informed and considered group opinion because the need for participants to find common ground gives them strong incentives to share information and opinions, gain knowledge, and give each view a fair hearing (Bächtiger, Grönlund, & Setälä, 2014;Karpowitz & Mendelberg, 2014).This goal of higher quality deliberation might be most desirable in enclave forums commissioned to develop new policy ideas or tackle emerging issues.And unanimity may be more likely than majority rule to mobilize participants to further action on behalf of the group's conclusions because consensus strengthens each member's commitment to the collective verdict (Karpowitz & Raphael, 2014).Future dialogue-to-change projects like Facing Racism may want to use unanimity in the final enclave forums to boost participants' commitment to pursue the actions chosen by the group.
In summary, the Facing Racism dialogues highlight ways in which affinity group discussions can be designed as one step in a more equitable deliberative process that also includes cross-cutting conversations about issues of importance to the public as a whole.To help enclave members discover and draw on their common perspectives, while preserving participants' freedom to arrive at different policy preferences, facilitators can foster exploration of similarities and differences of experience and understandings within the group.Moderators can question those who may try to impose a single identity or set of interests on the enclave, and encourage each participant to articulate their own understanding of their social locations.Organizers can expose participants to diverse issue framings, arguments, and information, especially those in circulation among members of the enclave itself, reminding participants that people who are "in the same boat" can and do see things differently from one another and from others in different situations.This may be easiest for participants to do when discussing aspects of issues that are just emerging into public consciousness and have not yet been framed rigidly in public discourse.Decision rules within the enclave should be chosen carefully to maximize either individual freedom (through majority rule, for example) or to boost the power of the group's voice in the policy process (through unanimity).Facilitators need to prepare enclave members to transition to mixed-group discussion by helping them see the importance of these conversations for enclave members themselves, to recognize the possibility that people of different perspectives may nevertheless share common understandings and values, and to develop skills for taking part in mixed group discussions, especially for appealing both to group and public interests, and working through conflict.Because the enclave's political agency should not be limited by the willingness or ability of public agencies or more privileged community members to pursue the enclave's preferences, it is important to ask enclave members to envision actions they might take within their group as well as in more heterogeneous civic organizations or through government.

Conclusion: Questions, Contexts, and Future Research
We have argued that enclaves need not be lesser forms of participation, or necessarily undermine public-minded thinking, in order to advance equity.By moving beyond a narrow perspective that sees the realization of deliberative ideals as possible only in cross-cutting groups, we can create affinity group processes that enhance marginalized people's inclusion, participation, and influence in public forums.We need a more welcoming, flexible, and intentional range of deliberative structures that create spaces for all community members' full investment and participation.In well-designed enclaves, members of historically and situationally disempowered groups can develop their political understanding as they explore the diversity of their lived experiences, envision themselves as fully participating members of the community, articulate ideas and arguments that can be difficult to raise initially in mixed groups, and widen the pool of reasons considered in public discourseto the benefit of both privileged and marginalized groups alike.
Having discussed the many advantages of enclave deliberation among the disadvantaged, how to avoid its potential dangers, and some principles for integrating it effectively into more representative political processes, we conclude with some practical considerations about how organizers of public deliberation can discern whether and how to incorporate affinity group discussions, the contexts in which they are most likely to strengthen equity, and some ideas for future research.

Questions for Organizers
When deciding whether and how to design enclaves into processes of public deliberation and consultation to make them more equitable, organizers can ask themselves the following questions.
 Relevant publics and enclaves: Whose interests or values are most affected by the issue?Of that group, who are the most politically disempowered, whose voices are least well-represented elsewhere in the political system?Is there anyone else who is situationally disempowered in relation to this particular issue?Is there anyone else who is likely to be disempowered by the act of deliberation itself?What sorts of enclaves might be formed to help them participate on more equitable footing? Role of the deliberation: Where does the deliberation fit in the political process, and what will it contribute to the larger political system?How could the results of enclave discussion be formally connected with or integrated into existing institutions of public decision-making?For example, could enclave deliberation lend greater legitimacy to choices about matters of distributional equity?Could such an enclave help to better prioritize scarce resources?Could an enclave help to clear a logjam elsewhere in the political system or prevent against a hasty decision that would harm the interests of a disempowered group?Could the group help inform and generate solutions to emerging problems? Inclusion: How can the perspectives of these disempowered groups be included in sufficient numbers so that they can talk among themselves as well as with others?How many of the marginalized need to be included to form a critical mass, rather than token representatives in small group discussions?How will the process uncover diversity of opinion within the enclave? Integration: How will the process promote interaction with other members of the public, stakeholders, experts, and officials?At what stages will such interaction occur?How will the process empower enclave participants' voices in the larger process? Audiences: Who will be the audience for the enclave's conclusions?How will these conclusions be decided and communicated effectively and fairly?What will the audience need to know about the aims and process of the enclave deliberation in order to see it as legitimate?What institutional links might persuade the audience to listen? Actions: How will the enclave be invited to take constructive action in their own groups and with other community members or institutions?

Contexts
While we think that well-designed enclave deliberation among the less powerful could enhance equity under most conditions, there are several contexts that seem especially ripe for it because other equalizing strategies are likely to be insufficient.
 Triple disempowerment: In some situations, groups that are among the most affected by the issue at hand have been politically, situationally, and deliberatively disempowered.For example, equity may especially be achieved by including enclaves of undocumented immigrants in deliberations over immigration policy, of youth in discussions of educational policy, and of low-income neighborhoods in discussions of city budgets. Extreme disempowerment: Some groups are highly disadvantaged with regard to one dimension of power.For example, organizers might especially consider whether to integrate enclaves of prisoners in discussion of criminal justice reform (to counteract their political disempowerment), the elderly in regard to nursing home care (to address their situational dependency on caregivers), and the developmentally disabled or deaf in discussions of education and social services targeted to their communities (to counteract deliberative challenges). Lack of alternatives: Enclaves may be especially important in deliberations that address issues in which the marginalized have few other avenues to empowerment elsewhere in the political system, particularly when associations and movements of the disempowered are weakest, least democratic, or simply have not yet recognized an emerging issue.An example is a set of forums in the U.S. that recruited people of color to solicit their policy recommendations for guarding against discriminatory uses of medical data in genetics testing (Bonham et al., 2009)an issue that emerged from the medical community rather than from grassroots associations of people of color. Relational issues: Enclaves may also be most useful when deliberation focuses on relations between unequal social sectors, such as relations between members of majority and minority faiths or ethnicities (as in the Facing Racism example).When intergroup relations are highly strainedfor example, between the police and communities of color after episodes of violence between thembeginning discussions in enclaves may be especially necessary for preparing people for mixed group discussions on civil as well as equitable terms.By extension, enclave discussions may also help more privileged and powerful groups reflectively consider (or reconsider) their perspectives and practices in preparation for discussion in mixed settings. Need for new solutions: Enclaves of the least powerful may be especially necessary and useful to fill in professionals' and officials' blind spots, drawing on the enclave's situated knowledge to generate new ideas by allowing disempowered participants to add their own proposals to the agenda rather than restricting discussion to a closed menu of policy options provided by others.As an example, recall how participants in the broadband forum discussed earlier identified privacy and security as critical components of digital inclusion programs.

Next Steps for Researchers
Despite our optimism, we recognize that there is still much to learn about the judicious and effective use of enclaves.Evaluations and other research could take several directions.First, in evaluating enclaves and mixed-group discussions, researchers should routinely compare individual-level data to test whether enclave deliberation helps less privileged participants to think, talk, and form their views on equal footing with more advantaged participants in the wider forum or public sphere.Do enclaves indeed allow the less privileged to expand organizers' initial issue framings and policy options?Do they empower enclave members to set the agenda, speak, and influence others equitably in mixed groups too?
Given concerns about group polarization, we should continue to study whether opinions in enclaves are based on mutual reasoning rather than social coercion and inability to consider diverse views.We need to know not only whether enclaves polarize, but why or why not.We should study the role of exposure to competing arguments and facilitation in the forum on enclaves' ability to explore both their commonalities and their differences.Are there particular skills that facilitators need most in order to help enclaves of the marginalized to freely explore their shared and divergent interests?Research can tell us more about whether exposure to balanced arguments and information about an issue allows enclaves to elude undesirable forms of polarization.In addition, are there particular issues that are more and less likely to induce polarization or autonomous deliberation in enclaves?We should continue to conduct experimental research on the effects of different decision rules on enclaves of the disadvantaged.Are we right to hypothesize, in contrast with some group polarization research, that consensus decision making may best protect enclaves of the disempowered from coercive polarization by protecting minority viewpoints within the group?
We can learn more about the quality of deliberation and opinions forged in enclaves.Observation, interviews, and other creative research designs can do more to uncover deliberators' motives for transforming their views and the development of public-minded thinking in enclaves, as participants consider how their group interests relate to the interests of others.In addition, while we still need to ask how much citizens learn in order to assess how well-informed their opinions are, we should also want to know what officials and experts learn from their encounters with enclaves of the less privileged.What kinds of lay knowledge and policy recommendations that emerge from enclaves do experts and officials value most and least, and what kinds ought they to value?Comparative research should test our hypotheses that enclaves of the disempowered are more likely than mixed groups to discuss the situated knowledge of the marginalized and to evaluate expert knowledge more extensively from disempowered perspectives.Studies should also routinely disaggregate participants' learning gains to show whether members of enclaves achieve equitable levels of understanding with more privileged participants to measure whether their final policy preferences are equally well-informed.
Comparative studies of enclave deliberation could also tell us whether the many elements of deliberative design can be disentangled well enough to show which most effectively foster equity.For example, are enclaves most likely to deliberate well and make valuable contributions to debate over issues that are more or less established, local, fact-driven or value-driven, and accessible to personal experience?At the same time, research needs to consider whether these many design factorssuch as group composition, issue, the format of the deliberation, and its degree of empowerment over decision makingmay interact to reinforce one another.The whole experience may be more than the sum of its parts.
We should investigate further the most effective ways to transition enclave deliberators to broader discussion with more privileged deliberators and officials.For example, are the disempowered likely to speak and influence others more if they shift from an enclave to a mixed group in which the less advantaged are the large majority or to a mixed group in which they are no longer the majority?And how does the decision rule used by the group affect this assessment?We noted above that experimental research suggests that when majority rule is used, peopling groups with majorities of women may be enough to promote their voices.Requiring consensus may also protect women enough that they can deliberate on equal terms with a majority of more powerful community members.Will this also be the case for deliberators who are also politically disempowered within the larger democratic system or situationally disadvantaged in relation to the topic of the deliberation?
Transitioning to mixed-group conversation also raises important questions about how to prepare more privileged participants to engage with members of less powerful enclaves, especially when inequalities are large or conflictual, as in many conflict transformation and peace building efforts.We certainly do not want to suggest that enclaves of the less powerful should be a substitute for challenging the privileged to examine their own positions through, for example, anti-oppression and anti-racism trainings.In thinking further about how to design and evaluate mixed-group deliberation that accompanies enclave discussions, we need to integrate insights from the research on intergroup conflict and communication, which generally supports many organizers' practice of moving from dialogue that builds affective ties and trust among unequals to deliberation among them about solving social injustices (e.g., Demoulin, Leyens, & Dovidio, 2013;Ellis, 2010).However, a recent strand of experimental research on relations between immigrants and the native born and between differently-advantaged racial groups demonstrates that "intergroup harmony per se does not necessarily lead to intergroup equality" (Saguy, Tausch, Dovidio, & Pratto, 2009, p. 120).This work shows how positive contact with an advantaged group can cause members of a disadvantaged group to underestimate group inequality and be less willing to take collective action to address it if their interaction focuses only on appreciating their commonalities with the privileged (Saguy et al, 2009;Dovidio, Gaertner, & Saguy, 2009;Glasford & Calcagno, 2012).These findings suggest that to promote resolve for egalitarian social change, not just positive interpersonal communication and feelings, mixed-group discussions need to focus on structural inequalities as well as shared values or goals, acknowledge the unfairness of dominant groups' privileges, and generate respect for members of subordinate groups.Experimental and field research is needed to help us understand more about how to structure those mixed-group conversations to avoid cooptation of the less advantaged.
Finally, research can do more to uncover when and why conclusions and arguments that emerge from enclave deliberation are taken up by other parts of the political system.This research needs to compare affinity group deliberation with other political opportunities for opinion formation and decision making, such as interpersonal and social media communication, involvement in associations, traditional forms of public consultation, polls, and elections.This could help us understand when it would be most effective for the less powerful to devote themselves to deliberation or, alternatively, to other kinds of political action.Which offer opportunities for more autonomous participation and influence?The most powerful way to address doubts about civic deliberation, especially among the least advantaged, may be to evaluate whether it can fulfill democracy's need for political equity and public judgment better than other means of political expression and decision making.

Figure
Figure 1.Some Ways of Integrating Enclave Deliberation

5
Figure 2. Overview of Facing Racism in a Diverse Nation