American Academy of Political and Social Science

The Televised Deliberative Poll: An Experiment in DemocracyAuthor(s): James S. FishkinSource: Annals of the American Academy of Political and Social Science, Vol. 546, The Mediaand Politics (Jul., 1996), pp. 132-140Published by: Sage Publications, Inc. in association with the American Academy of Politicaland Social ScienceStable URL: http://www.jstor.org/stable/1048176Accessed: 30/11/2008 07:02

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/action/showPublisher?publisherCode=sage.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with thescholarly community to preserve their work and the materials they rely upon, and to build a common research platform thatpromotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

Sage Publications, Inc. and American Academy of Political and Social Science are collaborating with JSTORto digitize, preserve and extend access to Annals of the American Academy of Political and Social Science.

http://www.jstor.org

ANNALS, AAPSS, 546, July 1996

The Televised Deliberative Poll: An Experiment in Democracy

By JAMES S. FISHKIN

ABSTRACT: The concept of the televised deliberative poll is outlined. Its first realization-in Britain by the television network Channel Four in May of 1994-is described. Ordinary polls model what the public is thinking even when it is disengaged or inattentive. The deliberative poll attempts to model what the public would think, if it were truly engaged by the issues. A national random sample is brought to a single place where its deliberations, in small group sessions and with competing experts and politicians, can be broadcast nationally and where it can arrive at considered judgments. The 1994 British experiment was held on the issue of crime. The sample was highly representative, both attitudinally and demographically, and its opinions changed dramatically. Implications for future experi- ments are discussed, particularly for the National Issues Convention, the January 1996 PBS experiment based on the same concept.

James S. Fishkin holds the Darrell K. Royal Regents Chair at the University of Texas, Austin. He is author of Democracy and Deliberation: New Directions for Democratic Reform (1991) and The Voice of the People (1995). He is executive director of the National Issues Convention, a deliberative poll on issues in the 1996 election campaign televised by PBS in January 1996.

132

THE TELEVISED DELIBERATIVE POLL

AS democracy has spread around the world, it has brought "power

to the people," but under conditions where the people have little motiva- tion to think about the power they are supposed to exercise. Elections, opin- ion polls, primaries, and referen- dums all provide opportunities for the expression of public opinion. The opinions expressed by ordinary citi- zens through these institutions, how- ever, tend to suffer from at least two major defects. The first is the well- known problem of "rational igno- rance."1 If I have one vote in millions, why should I spend a lot of time and effort pondering the details of public policy or the positions of rival candi- dates? Lamentably, any individual's vote or opinion is unlikely to make much difference. In other words, a person has little effective incentive to behave in the way we would like ideal citizens to behave. Of course, I may have peculiar preferences and enjoy political discussions. Alternatively, I may, as a by-product of other things I do, acquire a great deal of knowledge that is politically relevant. I may, for

example, develop an accurate im-

pression of how the economy is doing or whether budget or tax cuts may affect my life.2 But generally, acquir- ing information in preparation for the voting decision is not, in itself, an effective way to have an influence on the policies I care about, assuming that there even are policies or politi- cal issues I care about.

There is also a second major defect with mass public opinion, at least as

1. See Anthony Downs, An Economic The-

ory of Democracy (New York: Harper & Row, 1957).

2. See Sam Popkin, The Reasoning Voter

(Chicago: University of Chicago Press, 1991).

it is measured in polls or surveys. There is a well-established tendency of polls and surveys to report what Philip Converse has famously called "nonattitudes." Basically, some of the opinions reported in polls or surveys may not exist at all. Citizens are "re- markably obliging" in offering an- swers to the questions posed even when they have no opinion whatso- ever on the particular issue in ques- tion.3 There are many important is- sues on which citizen views seem to fluctuate more or less randomly, even when scrupulous screens are in- cluded in questions to offer the re- spondents the option of saying that they do not know. While there is some controversy about which opinions do not exist and which are simply hard to measure, there is little doubt that a great many phantom opinions are reported in polls and surveys. Fur- thermore, there is little doubt that a great many other opinions are re- ported that are off the top of the head and represent very little real thought or effort.

When George Gallup first devel- oped the rationale for the public opin- ion poll as we know it, he argued that it would "restore the democracy of the New England town meeting" to the large nation-state. His idea of the democracy of the New England town meeting was very simple. Citizens gather together in a single room, they hear the arguments for and against various proposals, they come to some conclusions, and they vote. Gallup's notion was that through the mass

3. The classic discussion is Philip E. Con- verse, "The Nature of Belief Systems in Mass Publics," in Ideology and Discontent, ed. David E. Apter (New York: Free Press, 1964), p. 245.

133

THE ANNALS OF THE AMERICAN ACADEMY

media, which in his day were radio and newspapers, the competing views of political leaders are dissemi- nated to the people, and the people's views are brought back to the elites through the public opinion poll. It is as if"everyone is in one great room."4 He did not take account, however, of the problems of rational ignorance and nonattitudes. In effect, the room was so big that no one was listening, and some of the opinions voiced in it were not real.

The televised deliberative opinion poll is intended to overcome these two problems. By itself, it is not a panacea, but it offers a new method for stimulating serious citizen en- gagement with the issues in a manner that is statistically representative of the entire country. It changes the in- centives for citizens to seriously de- liberate on public issues. It stimu- lates a microcosm of the country to overcome the effects of rational igno- rance. Instead of reporting nonatti- tudes, it stimulates the development and measurement of considered judgments.

THE CONCEPT

The basic idea is simple. Take a national random sample of the citi- zen voting-age population and trans- port its members to a single site for several days where they can deliber- ate about political issues, discussing them face-to-face with each other and with political leaders. Broadcast por- tions of the debate on national televi- sion, and then, at the end of the de- bate, poll the participants on their

4. George Gallup, "Public Opinion in a Democracy" (Stafford Little Lectures, Prince- ton University, 1938), p. 15.

views of both the candidates and the issues. The result is a deliberative opinion poll.

An ordinary opinion poll models what the public thinks, given how little it thinks, how little it knows, how little it pays attention. A delib- erative opinion poll, by contrast, models what the public would think, if it had a better chance to think about the questions at issue.5

A deliberative poll overcomes the conditions that foster rational igno- rance. Instead of one vote in millions, a participant in the deliberative poll has one vote in several hundred. He or she has every reason to invest in political information and to be atten- tive to the conflicting claims in politi- cal debate. Each member of a na- tional sample of several hundred will have a big part to play in a major national event, one watched by view- ers around the country.

The point of an ordinary poll is descriptive. It depicts the actual dis- tribution of public opinion at a snap- shot in time. The point of a delibera- tive poll is prescriptive; it has the recommending force of what the country, in all its diversity, would think if it were better informed and had a better opportunity to debate and reflect on the questions at issue. In a sense, a deliberative poll is an actual sample from a hypothetical so- ciety-a hypothetical version of our actual society transformed by the op- portunity to become more informed

5. For more on the concept of the delibera- tive poll, see my Democracy and Deliberation: New Directions for Democratic Reform (New Haven, CT: Yale University Press, 1991). For historical and theoretical background, see my Voice of the People (New Haven, CT: Yale Uni- versity Press, 1995).

134

THE TELEVISED DELIBERATIVE POLL

and more thoughtful on the issues in question.

The deliberative poll is the most ambitious realization of changes in the role of television in politics, changes that became most obvious in the last American presidential elec- tion. In 1992, there were dozens of televised forums that permitted eclectic collections of voters to ques- tion the candidates directly, unfil- tered by the media professionals. These electronic town meetings, by and large, lacked the two essential ingredients of the deliberative poll: they were not representative, and they were not deliberative. Either through self-selected samples ofviewers calling in questions or through hand-picked studio audiences, such variations of talk show democracy fail to be statis- tically representative of the entire country as judged by the scientific methods of public opinion research. Such efforts also fail to be delibera- tive in that they stimulate off-the- cuff questions, not questions that re- sult from a sustained effort of face-to-face discussion.

Citizens in a deliberative poll are chosen in a national random sample, and their participation is subsidized so that the poor can participate on the same basis as the more wealthy. The citizens have briefing materials on the issues, meetings in small groups to work through the questions of greatest concern, and opportunities to pose those questions directly to the relevant political leaders on national television. A full-scale, statistically representative microcosm of the country is gathered in a single place and empowered to seriously con-

front the key issues in an electoral campaign.

The results of such a poll provide a critical supplement to the thinking of ordinary citizens. By issue area, the poll would give voice to the rec- ommendations of a deliberative mi- crocosm of the country in a dramatic event available to all viewers who cared to tune in. During a campaign season in which the rationally igno- rant ordinary citizen is wrestling with impressions of headlines, shrink- ing sound bites, and staged photo op- portunities, a televised deliberative poll could be of major assistance in bringing substance to the campaign and to the public's decision making.

Around the world, the combina- tion of television and opinion polling has introduced an informal plebi- scitary element into democratic prac- tices. Some political scientists have characterized the combined effect of polling and television as an "echo chamber": impressions of television sound bites are reflected by an almost daily stream of opinion polls, whose results are reported, in turn, in the mass media, to be reflected back once again.6 But given an inattentive, eas- ily distracted, and rationally igno- rant public, the resulting echoes of the "will of the people" fall far short of democratic aspirations.

The deliberative opinion poll would harness the power of these same technologies-opinion polling and television-to a new and con- structive purpose. Television would attract the political leaders and citi-

6. V. 0. Key, Jr. with Milton Cummings, The Responsible Electorate (Cambridge, MA: Harvard University Press, 1966), p. 2.

135

THE ANNALS OF THE AMERICAN ACADEMY

zens to participate; the sampling methods of survey research would en- sure that the participants repre- sented the country in all its diversity. The result would be an occasion for the country, acting through an en- gaged microcosm, to offer itself ad- vice at a moment when it can make a real difference-before a primary, referendum, or general election.

THE BRITISH EXPERIMENT

This new form of opinion polling became a reality on national televi- sion in Britain on 8 May 1994. It was broadcast by Channel Four, produced by Granada Television, and cospon- sored by the Independent newspaper. The sample was selected by Social and Community Planning Research (SCPR), the independent research in- stitute based in London, from a stratified list of constituencies, with probability proportional to the elec- torate. Forty randomly chosen poll- ing districts were then selected, 1 from each of the 40 constituencies. We first conducted a baseline survey of 869 citizens, randomly chosen from the electoral register in the 40 polling districts. This survey had a high re- sponse rate: 74 percent. It gives an excellent picture of the public's atti- tudes on the issue in question, what can be done about rising crime. It is highly representative of the entire country in age, class, geographical representation, gender, education, and every other important dimen- sion. But this baseline survey was not the deliberative poll. It was only the beginning of the process.

The site chosen for the partici- pants' deliberation was in Manches-

ter. Voters were invited to the Man- chester event only after they com- pleted the baseline survey. The 300 who completed the weekend in Man- chester were, in every important re- spect, indistinguishable from the 869 who took the baseline survey. In terms of class, education, race, gen- der, and geography, the weekend mi- crocosm was fully as representative of the entire country as the baseline survey sample.

Some critics had contended that women would not come or that par- ents with children would not partici- pate. Both surveys, however, had 50 percent men, 50 percent women. Par- ents with children were not in the least underrepresented. Another speculation was that either the old or the young would be seriously under- represented. Once again, the 300 are virtually indistinguishable from the 869 in terms of age.

Perhaps the most persistent criti- cism was that we would get a dispro- portionate share of those who were politically informed. We assessed this both with specific knowledge ques- tions and with newspaper reader- ship. In both cases, it was obvious that we got an excellent microcosm to come to Manchester. Newspaper readership provides a useful sum- mary statistic, because specific Brit- ish newspapers clearly delineate the class and political orientations of their readers. The baseline poll sam- ple and the weekend group matched up almost perfectly with respect to the readership of particular papers.

Furthermore, in their attitudes about crime, and in their political positions more generally, the week- end microcosm was just as repre-

136

THE TELEVISED DELIBERATIVE POLL

sentative of the country as the base- line survey. One claim of critics was that participants in the poll would be precisely the people who were more interested in the specific issue-the people who were more worried about crime.7 Once again, the weekend sample turned out to be an almost perfect microcosm.

As a starting point on the issue of crime, the weekend sample was an almost perfect representation of the nation gathered together in a single place. The challenge for the experi- ment was whether they would change over the course of the week- end. Of course, it is probable that they had already begun changing from the moment they received our invitation. Knowing that they would be on national television, they prob- ably began discussing the topic with family and friends, they probably be- gan to read newspapers and listen to the media with more care, they prob- ably read the briefing materials we sent them to prepare for the event. In all these ways, their views became unrepresentative of public opinion in the conventional sense. But those views also became representative in a new and different sense. They be- came representative of the views the entire country would come to if it were populated by ideal citizens, by people who took up the same oppor- tunity as the members of our sample did, to engage the issues and debate them over an extended period. In short, their new, consideredjudgments represented what the public would

7. See, for example, comments of Robert Worcester on the deliberative poll on the NBC Nightly News, 7 May 1994.

think if it actually had a better oppor- tunity to think about the issues.

Once they had a chance to think, would the results be any different from those of an ordinary survey? Comparing the "after" survey with the "before" (baseline) survey, there were a number of significant changes in the 300. I can present only some summary statistics here. More de- tailed and technical reports are in preparation.8 First, the respondents showed an increased sense of the limitations of prison as a tool for fighting crime. Support for "sending more offenders to prison" as an "effec- tive way of preventing crime" went down 19 points, from 57 to 38 per- cent. Support for the notion that "the courts should send fewer people to prison" went up 15 points, from 29 to 44 percent. Support for "stiffer sen- tences generally" as an effective way of fighting crime went down 13 points, from 77 to 64 percent.

A second major change was an in- creased willingness to support alter- natives to prison, particularly for ju- venile offenders. For example, the percentage strongly against sending a first-time burglar, aged 16, to an ordinary prison went up from 33 to 49 percent. The percentage agreeing that such a juvenile should receive a strict warning but should otherwise be left to the parents to sort out went up from 49 to 63 percent. For "crimi- nals who are not a big threat to soci- ety," the percentage favoring compul- sory training and counseling went up from 49 to 63 percent. There was also a drop in the focus on punishing

8. My collaborators Robert Luskin, Roger Jowell, and Rebecca Gray and I are at work on two systematic reports on the experiment.

137

THE ANNALS OF THE AMERICAN ACADEMY

rather than reforming criminals. For the question, "If the government had to choose, it should concentrate more on punishing criminals or it should concentrate more on trying to reform criminals," the percentage choosing punishing went down from 55 to 45 percent.

Perhaps even more dramatically, the respondents showed an increased sensitivity to the procedural rights of defendants. For example, the per- centage strongly disagreeing that the police should sometimes be able to "bend the rules" in order to get a conviction went up from 36 to 46 per- cent, and the percentage believing it is "worse to convict an innocent per- son" than "to let a guilty person go free" rose from 60 to 70 percent. Fur- thermore, on a specific procedural right, the "right to silence"-the right to say nothing under police question- ing without adverse effects at trial- there was a dramatic shift. On the question, "If a suspect remains silent under police questioning, this should count against them in court," the per- centage agreeing went down 16 points, from 57 to 41 percent. The percentage agreeing that "suspects should have the right to remain silent under police questioning" went up from 36 to 49 percent.

Despite this increased sensitivity to the procedural rights of defen- dants, the respondents remained very tough on crime. The percentage agreeing that "the death penalty is the most appropriate sentence for some crimes" remained unchanged at 68 percent, and the percentage agree- ing that "prison life should be made tougher and more unpleasant" also remained unchanged, at 71 percent.

The respondents also showed a move- ment toward family values as a strat- egy for dealing with crime. In terms of dealing with the root causes of crime, there was an increase of 11 points, from 66 to 77 percent, in agreement that "a very effective way to help prevent crime" is to "teach children the difference between right and wrong." There was a similar change in support, from 54 to 65 per- cent, regarding the notion that "par- ents spending more time with their children" is a "very effective way to help prevent crime."

It is worth noting that these re- sults are all net changes. Many more respondents changed than indicated by these figures because on many questions change in one direction was canceled out by change in the other. For example, on the idea that "the courts should treat suspects as innocent until proved guilty," there was virtually no net change, but only about half-46 percent-gave the same answer both times. On support for the death penalty, which re- mained unchanged at 68 percent, 41 percent changed their answers on a five-point scale, and 20 percent changed their basic position.

Some critics have argued that in- teraction in small groups must mean that the more educated and articu- late will unduly influence the pro- cess. First, it should be noted that we collected the responses through con- fidential questionnaires, so partici- pants would not feel any social pres- sure in giving their private, considered judgments. Second, the moderators were trained to involve everyone in the discussion, and to make sure that no individuals domi-

138

THE TELEVISED DELIBERATIVE POLL

nated the discussion. Here we drew on the experience of the National Is- sues Forums, a network of small group discussions around the world supported by the Kettering Founda- tion in the United States that, in a completely nonpartisan way, at- tempts to facilitate citizen discussion of public policy issues.9 The Kettering Foundation helped train the modera- tors provided by the Independent, and it offered invaluable advice on the briefing materials (on the basis of focus groups conducted by SCPR with British citizens). Third, it should be remembered what we are trying to accomplish. If, somehow, the entire society were magically trans- formed so that everyone were as en- gaged by the issues as the partici- pants in our sample, then there would also be opinion leadership, there would also be cases where more informed and articulate people would be more influential. The moderators and the deliberative atmosphere kept such influence within limits, but it cannot, and need not, be eliminated entirely.

About a year after the first delib- erative poll, Channel Four, Granada Television, and SCPR conducted a second experiment, a nationally tele- vised deliberative poll on the future of Britain in Europe, broadcast the evening of 11 June 1995, the twenti- eth anniversary of Britain's June 1975 referendum on European is- sues. Once again, a nationally repre- sentative sample was gathered in

9. For an excellent summary of the ap- proach of the National Issues Forums in facili- tating small group discussion of issues, see David Mathews, Politics for People: Finding a Responsible Public Voice (Urbana: University of Illinois Press, 1994).

Manchester, and once again they changed dramatically after a week- end's deliberations. There were mas- sive, statistically significant changes in favor of closer links with the Euro- pean Union. Once again, all the changes were not in one direction, as the citizens also moved dramatically against the Common Agricultural Policy and in favor of retaining bor- der controls (to keep out drugs and crime). From a baseline situation where 84 percent had said that they did not have enough information to vote in a referendum, they devel- oped considered judgments that, on balance, made them skeptical Euro- enthusiasts.

Imagine what would happen if the deliberative poll, now conducted suc- cessfully twice in Britain, were in- serted into the beginning of the American presidential selection pro- cess. The two defining features of the deliberative poll, that it is repre- sentative and deliberative, would of- fer a dramatically different start to the process, a dramatically different version of the so-called invisible pri- mary, the period before the first offi- cial events when momentum for both issues and candidacies is born.

Rather than small, self-selected electorates in unrepresentative states, we would have a national ran- dom sample. Rather than sound-bite campaigning or the ritual of endlessly repeated stump speeches, we would get a deliberative and in-depth ex- amination of the candidates and the issues at the moment when it could make a difference-the beginning.

The idea is not to replace the pri- mary process as we know it but to provide a deliberative prelude where

139

THE ANNALS OF THE AMERICAN ACADEMY

the people can come to grips with the issues facing the country during the period when momentum is born. This deliberative poll would simply be a supplement to the many conven- tional polls that will undoubtedly be held during the period. But the early conventional polls do not do much more than measure name recognition of candidates even though they pro- vide the basis for fund-raising in our increasingly front-loaded primary system. With California, New York, and other major states moving up in the queue, it should be obvious that the opportunity for the people to have representative and deliberative input will come only at the very beginning.

The British have already an- nounced that they will make the de- liberative poll the centerpiece of one network's television coverage of their next general election.10 With luck, the process they have pioneered may be

10. "Channel 4 to Introduce Radical Polling Method," Broadcast, 6 May 1994, p. 3.

used to bring deliberative democracy to our next venture in presidential selection.1" The hope is that we can use the two technologies, polling and television, that have, thus far, com- bined to give us a superficial form of mass democracy. Through survey re- search, we can select the sample and formulate and tabulate the ques- tions. Through television, we can at- tract the citizens and the candidates and disseminate the conclusions of the deliberative microcosm. The re- sult may be a new method for bring- ing power to the people, but under conditions where the people can think about the power they are to exercise.

11. Since this article was written, I have served as executive director of the National Issues Convention, a deliberative poll with presidential candidates nationally televised on PBS. It was broadcast in January 1996. The sample was selected by the National Opinion Research Center, and the broadcasts were pro- duced by MacNeil/Lehrer Productions.

140