Democratic Innovation

Deliberation, Representation and Association

Michael Saward

• Publisher: Taylor & Francis
• Year: 2003
• eBook ISBN: 9781134566617
• Edition: 1
• Topic: Politics & International Relations
• Subtopic: Politics
• Index: Politics & International Relations

Democratic Innovation is an original look at the political future of democracy, exploring the latest ideas aimed at renewing popular power.
Featuring new writings by leading European, American and Australian democratic theorists, this book explores the following themes:
* the importance of public deliberation in democracies
* how effective representation for all might be acheived
* the role that voluntary associations can play in democratic governance

Part I
Deliberative democracy

Advocacy and critique

1 The quest for deliberative democracy¹

James S. Fishkin and Robert C. Luskin

The history of democratic reforms is a history of two closely related trends. The first has been toward more direct public consultation. The public has been given a larger and larger role in determining who makes policy decisions and even in deciding directly on policies. Some obvious examples in the US include the spread of mass primaries and relatively open caucuses in candidate selection, the direct election of US Senators, the atrophying of the Electoral College into a mere vote aggregation scheme, the increasing reliance on referenda and initiatives in legislation, and the evolution of public opinion polls into unofficial, advisory referenda. Although the first few of these examples are specific to the US, the same general trend pervades the democratic world.

The second trend has been toward broader public consultation. The public being consulted more directly has also grown wider and wider. In the US, the relevant changes have included the broadening of the franchise to include non-property owners, women, blacks, and eighteen- to twenty-year-olds; the removal of poll taxes, residency requirements, and other barriers to voting; and positive measures to facilitate registration and voting such as balloting periods of more than one day and motor-voter statutes. Again the rest of the democratic world has been moving along the same curve.

Equality, deliberation and deliberative democracy

Generally speaking, these changes have served political equality, the democratic ideal of weighing all citizens’ preferences equally.² But this pursuit of equality has had the unfortunate effect of undermining another democratic value, deliberation, broadly defined as the serious consideration of arguments and counter-arguments for and against policy alternatives.³

Note, parenthetically, that this notion of deliberation does not absolutely require discussion, face-to-face or otherwise. Theoretically, the arguments and counter-arguments could be raised either through mediated discussion, as by some synchronous or asynchronous on-line exchange, or even without discussion, as by written digests of the full range of arguments that should be considered. Real-world deliberation is a mix – people read, watch, and listen; people ruminate; people discuss. But it does seem safe to say that deliberation quite centrally involves discussion, and indeed that at least some of the benefits of deliberation would be harder to attain without it.

The irony of the institutional changes promoting equality is that they have moved the effective locus of many decisions to a mass public subject to what Anthony Downs (1956) famously termed ‘rational ignorance’. Why should individuals with only one vote in millions spend a lot of time and effort becoming informed about a national policy debate or political campaign? No individual vote or opinion is likely to make any difference.

Where deliberation requires immersion in policy-relevant arguments and counter-arguments, the incentives for rational ignorance promote inattention and non-participation. Thus democratic institutions appear to face a choice between politically unequal but deliberative elites and politically equal but non-deliberative masses. Is there some way of combining deliberation and equality? The quest for institutions to accomplish this is the quest for deliberative democracy.

Existing democratic institutions based on elected representation may be seen as an effort to achieve the ideal of deliberative democracy by stages. They combine equality at a preliminary, electoral stage with deliberation at a subsequent, decision-making one. The division is not absolute. There are elements of deliberation in elections (as citizens, to whatever usually modest degree, learn about and discuss the choices) and of equality in representative decision-making (one legislator, one vote, at least at final passage). But by and large the equality (such as it is)⁴ lies in elections and the deliberation in discussions among elected representatives.

But this strategy of concentrating deliberation among elected representatives is far from ideal. At the decision-making stage, considerations other than those likely to figure prominently in mass deliberations may carry great weight. Elected officials have been known to pursue their own pet interests or those of powerful individuals or organizations above those of the majority of their constituents. The equality achieved by elections is often subverted.

At other times, elected representatives concerned with re-election may simply follow the polls, abandoning any meaningful deliberation. That would not necessarily be bad, if the polls reflected substantial public deliberation. But of course they do not. The results stand in generally unknown and often quite limited relation to the views people would hold if they had the chance of learning, thinking, and talking about the issues (as illustrated below). The decision-making stage, then, is no-win: ignoring mass preferences vitiates equality; following them vitiates deliberation.

At the electoral stage, the paucity of deliberation means that there is no great deal of equality (beyond the merely formal) either. Because citizens are often – but varyingly – unengaged and ignorant, not everyone votes. In the US and Switzerland,⁵ almost half the eligible electorate abstains, even in the most salient national elections. Electorally, non-voters have no voice.

In a subtle but important sense, moreover, even voters have unequal voices. Every vote carries equal weight, of course (issues of fraud and apportionment aside). But the paucity of deliberation also means that not everyone votes as he or she would with ‘full information.’ And because information is very unevenly distributed across the electorate, some people approximate their full-information preferences much more reliably than others. Thus while voters’ actual preferences are equally represented, their full-information preferences are not.⁶

These shortcomings of elected representation have made the quest for deliberative democracy into a quest for a more unitary process capable of embodying both equality and deliberation, a process of deliberation among citizens, not just among legislators or other elected representatives. The key problem is how to overcome the barriers to citizen deliberation in communities numbering in the thousands or millions.

The Deliberative Poll (and other deliberative microcosms)

An ancient solution to this problem of scale was the deliberative microcosm selected by lot – in modern terms, by simple random sampling. Random selection gave each citizen an equal probability of participating. The Athenians used this device for citizens’ juries of 500 or more, for legislative commissions, for the Council (which set the agenda for the Assembly), and in a variety of other contexts. By the fourth century BC, legislative commissions of several hundred citizens selected by lot were making the final decisions on legislation (see Hansen 1991 for a systematic account).

These ancient examples have lately inspired a range of deliberative microcosms,including citizens’ juries in the US and Britain (Coote and Lenaghan 1997), planning cells in Germany and Switzerland (Dienel and Renn 1995), and consensus conferences in Denmark ( Joss and Durant 1995). The shared idea is to gather citizens together in small groups to discuss policy issues face-to-face, thus providing both opportunity and incentives for the participants to behave more like ideal citizens. For the participants, at least temporarily, it becomes less rational to be ignorant. Instead of being just one vote among anonymous millions, each has one voice among ten or twenty in the small-group discussions and one among only hundreds even aggregating across groups. In addition, the necessity of talking to and in front of one’s fellow participants creates social incentives to inform oneself and think about the issues.

The point is to offer the participants’ post-deliberation opinions as an input to politics and policy. To the degree the participants represent the whole public – an important stipulation – the deliberative experience gives the final distribution of opinion a certain normative relevance, even as it renders it highly counterfactual.
The wider public remains rationally ignorant, unengaged, ill-informed. In contrast to ordinary polls, showing public opinion as it is, these deliberative microcosms attempt to show public opinion as it would be if its members learned, thought, and talked more about the issues.

Our own version of this general idea is the Deliberative Poll. We take a random sample of some relevant population (country, region, city), interview them, and invite them to a single site for a deliberative weekend. We send them carefully balanced briefing materials, laying out the main policy options and the arguments for and against each. We bring them to the site (all expenses plus an honorarium paid) and set them to deliberating in two alternating modes: small-group discussions led by trained moderators and plenary sessions to put questions composed by the small groups to panels of competing politicians and policy experts. At the end we ask the participants the same questions as at the beginning.

While all deliberative microcosms involve deliberation, the Deliberative Poll goes furthest toward preserving equality – as part, in this context, of representation. Two nested sorts of representation, and of equality, are relevant. The first is pre-deliberation representation of the actual public – demographically, attitudinally, behaviourally, and in terms of knowledge, thought, and involvement. What may or may not be equal here is the probability of being drawn into the deliberative sample. The second sort is post-deliberation representation of a hypothetical, more engaged, thoughtful, and knowledgeable public. What may or may not be equal here, in addition to the probability of being sampled in the first place, is one’s contribution toward determining the post-deliberation results. Good representation of the first sort is a necessary but insufficient condition for good representation of the second.

One advantage of Deliberative Polling for post-deliberation representation lies in the nature of the deliberative process and results. Citizens’ juries, planning cells, and consensus conferences are designed to reach some public, collective decision. The Deliberative Poll, in contrast, is designed simply to help its participants inform and clarify their own thinking. We gauge their opinions with the same privately completed questionnaire as at the outset. No one else knows what responses they give. Their opinions are aggregated only statistically, after the deliberations are done. This is essentially a ‘secret ballot’ and has the same advantage vis-à-vis the social aggregation of opinions in citizens’ juries, planning cells, and consensus conferences as the secret ballot does for elections in the outside world: it minimizes the intrusion of other people’s preferences, and thus the inequalities that creep in when some people, inevitably, have greater economic, social, or informational resources than others.

The advantage of Deliberative Polling for pre-deliberation representation is still clearer-cut. Equal probabilities of selection require probability, i.e. random, sampling.⁷ But the participants in consensus conferences are self-selected, and while planning cells have frequently employed random sampling, they have done so only within non randomly selected communities, making their pooled samples, aggregated across communities, non random.⁸

But representation of both sorts involves not only equality but sample size. Even assuming random sampling, the number of participants must be sufficiently large to lend some assurance that the sample resembles the population – as it is, before deliberation, and as it might be, after. A random sample of several hundred is very unlikely to differ radically from the population. A sample of only ten or twenty, the size of most citizens’ juries and planning cells, may easily do so. Both citizens’ juries and planning cells are therefore much too small to be especially representative. The pooled samples from several parallel planning cells may be large enough but are non random. The result is that there is no great assurance that another jury or cell of the same size, drawn from the same population, and deliberating in the same fashion on the same topic would yield anything like the same results – or, equivalently, that the whole population, divided up into citizens’ juries or planning cells, would do so.⁹

In Campbell and Stanley’s (1963) terms, the Deliberative Poll is a quasi-experiment, whose one compound treatment consists of everything from the invitation to participate to the last moment of on-site deliberation. In anticipation of the weekend, participants begin to pay more attention to the news, to think more about the issues, and to discuss them more with friends and family. Many of them also read the briefing materials prepared for the event. Perhaps the most important part of the treatment, however, occurs on the weekend itself, in the small-group discussions among the participants and in the dialogue with competing experts and politicians.

When the idea was first proposed, one sceptic questioned the necessity for deliberation, arguing that ‘the process of group interaction will produce substantially different results than if the participants consider alternatives and are asked questions in private’ (Traugott 1992: 27). But we believe it important that the participants have the opportunity of hearing others’ views and questions and considering others’ interests. Our ideal citizens would not simply receive and digest political information in private, then respond by voting or otherwise expressing their views without ever having shared or tested them in conversation. Rather, they would talk with a wide variety of other citizens, hearing and weighing opposing points of view in the process of coming to their own considered judgments. They would deliberate.

One other important feature needs mentioning: In every Deliberative Poll so far, the proceedings or results have been broadcast to the rest of the community (nationally, in the case of the national Deliberative Polls; regionally, in the case of the regional ones). They have also received coverage in radio and newspaper stories, on internet websites, etc. Such broadcasts and stories give the Deliberative Poll the possibility of making a contribution to the actual practice of democracy. Both citizens and policy-makers can be made aware of the conclusions of the Deliberative Poll and the considerations that weighed with the microcosm in reaching those conclusions. They may help set an example for representative and informed citizen consultation in an age of self-selected viewer call-ins and televised ‘town meetings’ based on convenience samples.

Deliberative Polling and democratic theory

The key theoretical and empirical question to do with Deliberative Polling is how a counterfactual world of informed and thoughtful citizens discussing policy issues would differ from the very different one we know. No mass public anywhere in the world has been continuously and seriously engaged in public issues.¹⁰ Between such an ideal public fulfilling the aspirations of democratic theory and real-world publics mired in rational ignorance yawns such a chasm that we have trouble telling much about what the ideal case would be like. Theorists can only attempt to imagine it. The point of Deliberative Polling is to approach these questions in an entirely different way, moving from the thought experiments of democratic theory to the actual experiments (and quasi-experiments) of social science.

From the standpoint of democratic theory such experiments help us approach a problem central to the 2,500 year old dialogue about the desirability of democracy. Advocates of democracy have long faced the challenge that the whole public might not be sufficiently competent to rule. Why not have rule by meritocratic or technical elites instead? Or by only the more educated and the higher strata of society? Scepticism about the competence of ordinary citizens famously motivated Plato’s Republic (although he later qualified this scepticism by his willingness to rely on random microcosms of the citizenry in The Laws). It also motivated J. S. Mill’s (1991) proposal for ‘plural voting’, in effect allotting extra votes to those who could demonstrate their greater wisdom or intelligence.

Every society that has invoked the notion of democracy or popular rule has wrestled with this challenge. And every known democracy has employed some combination of public consultation and decision-making by elites. The issue of the appropriate mix, however, remains. In employing a representative microcosm, deliberative polling sheds light on the public’s potential for competent decision-making. How far could we trust ordinary citizens, even under unrealistically favourable conditions, to inform themselves and discuss and consider the issues?

Deliberative Polling also illuminates a related issue very much at the intersection of democratic theory and political behaviour: namely, what differences it would make to the distribution of policy and electoral preferences if the citizenry knew and cared more about politics. There is now little controversy about claims that most people know little about the details of politics and policy (Price 1999; Luskin 2000). There remains, however, a contention that ill-informed citizens manage by cognitive short-cuts or ‘heuristics’ to approximate their ‘full-information’ preferences (Popkin 1991; Lupia and McCubbins 1998). Such ‘low-information rationality,’ it is claimed, leads to roughly the same preferences as the ‘high-information’ kind.¹¹ A less-charitable but aggregatively equivalent contention is that while actual and full-information preferences commonly differ, the differences cancel out, with some citizens erring in one direction and some in the other, so that the distributions of actual and full-information preferences are nonetheless the same (Page and Shapiro 1992).

We do not doubt t...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contributors
• Series editor’s preface
• Editor’s acknowledgements
• Introduction
• Part I Deliberative democracy: Advocacy and critique
• Part II Representation and deliberation
• Part III Associations and democracy
• Conclusion
• Bibliography