1. Even when citizens make flawed judgments, random chance should still make their votes "correct" half of the time on "yea" or "nay" policy issues and in two-candidate elections. The goal is to improve judgments above that 50 percent threshold. [BACK]

2. A useful but expensive research project would examine the extent to which citizen panels produced varied judgments under similar information conditions. Such an experiment would be difficult because it would require several days of time from thousands of participants (e.g., 5 days 30 panel members 100 panels = 15,000 participant days). Moreover, the practical constraints of such research would limit group size, the depth of information exposure, the amount of deliberation, and the amount of external critique of preliminary arguments. The last of these processes would probably reduce considerably the amount of variation that would occur in citizen panels that began to diverge after the initial panel deliberation. [BACK]

3. Tindale 1996: 16 [BACK]

4. The establishment of a solid information base is the first function in Gouran and Hirokawa’s 1996 functional theory of small group communication. [BACK]

5. Hirokawa 1990 argues that when information is evenly distributed among group members, communication has little influence on decision making, and following this reasoning, one might question the need for deliberation after listening to witness testimony. The problem with that view is that the understanding, interpreting, and processing of that testimony will not be uniform across panelists, and panelists also have unique life histories that they can only share with one another through group discussion. [BACK]

6. Hastie, Penrod, and Pennington 1983 found that in twelve-person mock juries, there were dramatic differences in participation rates. It was common for the most active speaker to take up as much as a quarter of the group’s discussion time, whereas the quietest participant might say almost nothing. Studies commonly find that demographic and other individual-difference variables cause differences in speaking frequency and duration in small groups, and participation is, in turn, associated with group influence (Bonito and Hollingshead 1997). Baker 1988, for example, found that older and male participants spoke more often in small group discussions. It is important to have procedures in place to ensure adequate participation because the natural tendency is for extreme participation inequality to result from arbitrary social and psychological differences. See also Smith 1999: 43–44. [BACK]

7. As explained in chapter 6, the ability to challenge statements is central to deliberation as envisioned by Jürgen Habermas (1979). [BACK]

8. In a good summary of survey research, Krosnick 1999 identifies the three causes of superficial response ("satisficing") as issue complexity, low skill levels, and low motivation. As presented in chapter 7, the citizen panels not only give participants motivation, they also provide the information and brief training necessary to engage in political debate. In a similar sense, Petty and Cacioppo 1990 found that systematic cognitive processing of information requires a high level of "involvement." [BACK]

9. Mathews 1994: 36 [BACK]

10. Hare 1976: 24. To counter some readers' nightmarish memories of famous studies by Solomon Asch, a meta-analytic study by Rod Bond and Peter Smith indicates that the amount of conformity found in experiments using Asch’s "line judgment task" has "steadily declined since Asch’s studies in the early 1950’s" (Bond and Smith 1996: 124). [BACK]

11. On politics, see Barber 1984. On ambiguity and conformity, see Baron, Vandello, and Brunsman 1996; Bond and Smith 1996. [BACK]

12. On the influence of a single dissension, see Hare 1976: 32. The reversals Hastie, Penrod, and Pennington 1983 observed usually resulted from a change in many jurors' interpretations of the judge’s instructions or a wave of jurors changing their minds about a particular argument. Fishkin and Luskin 1999 and chapter 6 in this book give examples of such shifts in random sample forums. [BACK]

13. Hastie, Penrod, and Pennington 1983: 230 [BACK]

14. Janis 1982. Janis originally published Victims of Groupthink in 1972, but many readers are probably familiar with Groupthink, the expanded 1982 version of his book. [BACK]

15. Janis 1982 recommends this procedure. Neck and Moorhead 1992 stresses the importance of even more methodical discussion procedures as a means of preventing groupthink, but I have decided against rigid, preset discussion procedures. If the moderator has general guidelines to promote a balanced, respectful, participatory, and on-schedule discussion, that should be sufficient to keep the discussion focused and deliberative. If the moderator’s role extends too far beyond that, the potential for charges of moderator bias become too great. [BACK]

16. See Park 1990; Evans and Dion 1991 [BACK]

17. Sanders 1997 presents such a critique and relies primarily on evidence from juries. Peter Berkowitz also raises this concern about different skills at the "deliberative arts" ("The Debating Society," New Republic, November 25, 1996, 36–42). [BACK]

18. The mere presence of a well-trained group facilitator can greatly improve group discussions and might even be warranted in jury trials (Anderson and Robertson 1985). [BACK]

19. Eagly and Karau 1991. Though structuring discussion can reduce status effects on group behavior, it may be more difficult to create neutral percep tions of participants' behavior (Andrews 1992; Eagly, Makhijani, and Klonsky 1992). Mansbridge 1990b raises concerns about the interplay of gender roles and group deliberation. Will women be more attentive to men’s concerns than vice versa? Will men speak more forcefully and dismiss women’s speech as "emotional"? Once again, I think a well-designed discussion context can diminish the impact of gender roles in small group deliberation. [BACK]

20. Mansbridge 1983 effectively contrasts the adversarial and unitary democratic traditions and suggests the need for a better balance. Barber 1984 also argues along these lines. [BACK]

21. Edelman 1988 documents the "symbolic uses of politics" for both generating conflict and building consensus, but in an adversarial system, the tools of the trade are used most often in the service of oppositional politics. [BACK]

22. On the media’s role in obstructing deliberation and civic participation, see Entman 1989; Rosen and Taylor 1992; and Cappella and Jamieson 1997, which found that the bulk of news coverage of elections frames stories in terms of campaign strategy rather than substantive issues; readers and viewers "learned" more cynical perceptions from those strategic framings. [BACK]

23. On the destructive nature of political discourse in the United States, see Barber 1984, 1998b; Gastil 1992; Mathews 1994; Pearce and Littlejohn 1997; and Tannen 1998. Ryfe 1998: 11 provides a review of the relevant literature and summarizes this concern concisely: "Social fragmentation challenges models of good public discourse in three ways. First, it suggests that isolated social groups will have fewer interests in common…. Second, levels of trust may be much weaker between groups that have little contact with one another…. Finally, social fragmentation exacerbates longstanding traditions of discrimination" and stereotyping. [BACK]

24. Hart and Jarvis 1999: 81. Because it does not permit face-to-face discussion and some people maintain anonymity, conventional Internet discussions among citizens tend to be more adversarial. Nonetheless, in a study of political exchanges on computer bulletin boards, Thomas Benson found that even those "uncivil" discussions are valuable: "Amidst the name-calling, the flaming, and the ideological demonization common to both sides, there is a demonstrable faith of some sort in the power of argument and passionate advocacy" (Benson 1996: 374). In a more structured face-to-face setting, those underlying discourse norms become even more visible, and debate tends to be more productive because of the relative rarity of rudeness. [BACK]

25. Writings on criminal and civil juries suggest that even divided juries normally develop some camaraderie, and jurors sometimes even become friends as

26. The citizen conferences and citizen juries described in chapter 6 all reached such agreements on most critical points of discussion. See Crosby 1998 on the jury experience, and Gastil and Jenkins-Smith 1997 on the citizen conferences. [BACK]

27. Hans and Vidmar 1986: 129. [BACK]

28. League of Women Voters, "Charting the Health of American Democracy," June, 1997, available at http://www.lwv.org/report.html. [BACK]

29. Newport 1996: 7. [BACK]

30. The survey results were reported in Richard Morin, "Is Anyone Listening?" Washington Post National Weekly Edition, February 15, 1999, 34. [BACK]

31. Hans 1993: 254–57. [BACK]

32. Ibid.: 259. Valerie Hans also found that judges generally agree with jury verdicts and have great confidence in the jury process. [BACK]

33. Crosby 1998 [BACK]

34. Public opinion surveys, summarized by Hans 1993, have found that the public knows more about the role of juries than the court system in general. This is partly due to media coverage of juries, citizens' identification with jurors, and direct personal experience serving on juries. This suggests that the public might also readily learn about the panel process, and it might be able to extend jury-related concepts (e.g., a "hung jury") to the citizen panels. [BACK]

35. In the long term, the impact of the panels would be self-reinforcing. "When citizens assume that speaking in public is a source of power, public speech magically can become a source of power," Nina Eliasoph (1998: 263) explains. At least with regard to the panels, this process is not magic: if voters respond to panel recommendations during elections, the panels will become a powerful political cue. The importance of favorable panel evaluations will increase in proportion to the willingness of voters' to incorporate panel judgments into their own voting decisions. [BACK]

36. Hirschman 1970: 2 [BACK]

37. See Lyons et al. 1993: ch. 3. These authors drew the concept of "neglect" from Farrell and Rusbult 1981 and Rusbult and Lowery 1985, which apply the exit, voice, and loyalty model to dissatisfaction among private and public employees. Rusbult 1987 has also extended this same revised model to the maintenance of close personal relationships. I do not discuss the fourresponse model developed by these authors because it changes the meaning of the term loyalty. In the original model, loyalty was a variable that influenced the selection of voice versus exit (e.g., Hirschman 1970: ch. 7, esp. fig. 1 on p. 87). For Lyons et al. and Rusbult, loyalty is one of four responses. They create a two dimensional table, with "active-passive" as the vertical axis and "constructive destructive" as the horizontal axis: voice is active-constructive, exit is activedestructive, loyalty is passive-constructive, and neglect is passive-destructive. I find their notion of neglect useful, but I do not wish to treat loyalty as a response to decline instead of a force that promotes the use of voice versus exit. [BACK]

38. University of Michigan data used by Ginsberg and Shefter 1990: 189–90 show that in the 1984 presidential election, 80 percent of eligible upper-middle-class adults reported voting, compared to only 70 percent of the lower middle class and 55 percent of the working class. As the authors acknowledge, survey data overestimate voter participation rates, but the overestimation should be rather evenly distributed across different social groups. Rosenstone and Hansen 1993 found similar results in analyzing voting trends from 1952 to 1988 broken down by income percentile. Only 39 percent of eligible voters 18–20 years old participated in the 1992 election, as compared to 70 percent of Americans 65 and older. Sixty-four percent of voting-age Anglos, 54 percent of African-Americans, and 29 percent of Hispanics participated (see League of Women Voters, "Charting the Health of American Democracy," June 1997, available at http://www.lwv.org/report.html). Differences in average external efficacy are taken from National Election Study data, available at http://www.umich.edu/~nes/nesguide/2ndtables/t5b_4_1.htm. [BACK]

39. Entman 1989: 26–27. Bennett and Resnick 1990 illustrates some of the differences in the political opinions of voters and nonvoters. [BACK]

40. Putnam 1995b: 664–65 [BACK]

41. See, e.g., Bennett 1998 and The Public Perspective 7, 4 (June–July 1996). [BACK]

42. Galston and Levine 1998: 36. Comparison of 1990 and 1994 surveys by Gallup and the Yankelovich Group show declines in public trust in everything from the local media to religious organizations. For a summary, see League of Women Voters, "Charting the Health of American Democracy" (June 1997), available at http://www.lwv.org/report.html. [BACK]

43. On the positive effect of deliberation on public trust, see Rasinski, Bradburn, and Lauen 1999; Tom Smith 1999; Tyler, Rasinski, and Spodick 1985. [BACK]

44. Levi 1996: 50 makes this point in a review of Putnam et al. 1993, Mak ing Democracy Work: Civic Traditions in Modern Italy.Sabetti 1996 goes even further, arguing that Putnam et al. 1993 underemphasizes the importance of government as a result of an oversimplification of Italian history. [BACK]

45. Warren 1996b: 259–60. See also Warren 1996a. On an equally philosophical level, Weithman 1995 argues that deliberation can provide a strong justification for authority in a democracy. [BACK]

46. The Pew Center conducted the study in association with the National Journal. "A Survey of Members of Congress, Clinton Appointees, and Senior Civil Servants" surveyed 81 members of Congress, 98 presidential appointees, and 151 members of the Senior Executive Service. Results can be found at http://www.people-press.org/leadrpt.htm. [BACK]

47. Ibid. The public opinion survey mentioned was a 1998 Pew study, "Deconstructing Distrust: How Americans View Government." [BACK]

48. This argument comes from David Broder, "Catatonic Politics," Wash ington Post, November 11, 1997. [BACK]

49. According to the opinion ballots completed by NIF participants, typical forums have relatively few participants with low levels of formal education (Gastil 1994b). [BACK]

50. Mark Warren argues that "a theory of democracy should be oriented toward creating institutional environments that encourage the self-examination of preferences that brings them closer to needs" (Warren 1992: 16). The panels may create such an environment by modeling self-criticism. [BACK]

51. Warren 1992: 12. On the effects of NIF, see Gastil 1994b, Mathews 1994, and Pearce and Littlejohn 1997. On the role of mutuality in small group democracy, see Gastil 1993. [BACK]

52. For example, Adam Simon (1997) does this in his study of candidate discourse during campaigns. Rational actor models are common, not only in economics, but also in political psychology and other social sciences. In reality, much behavior is "irrational," but altruistic irrationality is a rare commodity in many settings (Schwartz 1986). My purpose here is to demonstrate how the panels might compel even those "rational" candidates who seek only electoral success to behave in a way that promotes more deliberative elections. [BACK]

53. This partially meets Jack Knight and James Johnson’s call for deliberation to establish "common understandings of what is at stake in a given political conflict" (i.e., an election) after surviving "a process of reasoned debate sustained by fair procedures" (Knight and Johnson 1994: 289). [BACK]

54. Simon 1997, a study focusing on the 1994 California gubernatorial race, suggests that most contemporary reform proposals would do little to increase clashes among candidates, noting that it is exceptional for candidates in U.S. elections even to discuss the same general issue, let alone to challenge one another directly. [BACK]

55. The idea that panels invite different kinds of leaders comes from Bob Kraig, who suggests that the notion of a public process rewarding certain kinds of leaders and discouraging others resembles a notion found in the political philosophy of Woodrow Wilson (November 3, 1998, telephone interview). [BACK]

56. Leah Ceccarelli first brought this concern to my attention, albeit without using such a monstrous metaphor. [BACK]

57. This same argument could be applied to nonlegislative officials, and my argument against the view would be roughly the same. In the case of candidateselection panels, there is still the chance that the particular unpopular action in question might not come up, and if it does, it might not prove decisive or the panel might even see it as a positive action. [BACK]

58. Gastil 1994a. [BACK]

59. Barber 1998a: 118. [BACK]