1. Barber 1984: 188. [BACK]

2. Dahl 1989: 98. [BACK]

3. The moral philosopher John Rawls (1971) arrives at these principles by presuming that if people did not know what lot they might draw in society, they would want a social system that protected the interests of the least-well-off. Freedom and equality, as he defines them, serve those interests. [BACK]

4. "The more one knows about politics, the more effective—the more instrumentally rational—one’s voice is likely to be" (Delli Carpini and Keeter 1996: 56). On how values and decision logics can go together for individuals and groups, see Macoubrie 1998. [BACK]

5. Delli Carpini and Keeter 1996: 101–2, 133. Moreover, "[c]itizens generally take less interest in … and are less informed about their local governments than they are about their national government" (Burns et al. 1996: 11). [BACK]

7. Lull and Cappella 1981. [BACK]

8. Andrade and Cobb 1996: 229–30. [BACK]

9. Blumer and Gurevitch 1995: 214. [BACK]

10. Cappella and Jamieson 1997 provides detailed evidence of the "cynical" nature of mainstream media coverage of politics. Eliasoph 1998, a more openended study, concludes that the media contribute to a larger process of "political evaporation" that leaves the public sphere barren. Along similar lines, Hugh Heclo argues that the quantity of political news is large but that it is misleading and superficial. "Modern media," he writes, "deluge the public with information" and "give the impression of national problems as always unresolved…. [They] often associate immediacy with importance, and intensity with seriousness. A dull congressional vote, agency announcement, or international agreement may represent important change in the world, but it holds little attraction for the media next to a plane crash or a public clash of personalities…. Even media-sponsored policy debates often merely provide, in place of a one-sided presentation of an issue, two one-sided presentations, albeit from opposite sides" (Heclo 1999: 66). [BACK]

11. See, e.g., Ginsberg 1986 and Parenti 1995. Even those who find limited evidence of business's direct influence on policymaking still find traces of indirect influence through shaping public opinion. Mark Smith, for example, finds that when business reaches consensus on a high-profile issue, its direct influence becomes limited relative to public sentiment, yet he also finds evidence that

12. Zaller 1992: 327. Some scholars have begun to lose their optimism about the fidelity of public reception of elite cues. Kuklinski and Hurley 1996, for instance, argues that citizens routinely misinterpret elite messages. [BACK]

13. Bennett 1993: 109. [BACK]

14. Figures are derived from William Jacoby’s (1991, 1995) analysis of 1984 and 1988 data. Using 1984 national survey data, Sniderman, Brody, and Tetlock 1991 also found that 20 percent of the population to have "intense" ideological feelings. Respondents rated liberalism and conservatism on a 100point "feeling thermometer," and intense ideologues were defined as persons reporting a difference of 25 points or higher in their two thermometer ratings (e.g., giving liberalism a rating of 50 and conservatism as rating of 75 or greater). Only 7 percent of respondents could be categorized as intense liberals and 13 percent as intense conservatives. The most influential writing on ideology and the American public was an essay by Philip Converse (1964), which found that only a tiny fraction of the public engaged in ideological thinking but that, again, roughly one-fifth of the public has "real" attitudes. These estimates are important because they suggest what percentage of the public uncritically filters political messages. [BACK]

15. Conover and Feldman 1984 [BACK]

16. Benjamin Page (1996) shares John Zaller’s belief in the value of learning from elites, but he raises this concern about elite versus mass interests. In chapter 6, I examine Page’s view in detail. Shapiro 1998 raises similar questions about Zaller’s model. [BACK]

17. David Mathews (1994) has promoted the use of the phrases "public perspective" and "public voice" to remind us of the difference between aggregated self-interest and a broader public will. [BACK]

18. Fiorina 1981. [BACK]

19. Ibid.: 208. [BACK]

20. Although the high-information voter engages in elaborate, issue- and party-based candidate evaluation, "the less well informed voter may have the information he needs provided he treats the choice before him as a choice for or against the incumbent; for poorly informed or not, he is in a position to judge if the incumbent’s performance is satisfactory" (Sniderman, Brody, and Tetlock 1991: 178). I believe that evaluation of actual incumbent performance is more difficult than other evaluations, and if low-information voters are making their choices that way, they are likely to make haphazard candidate choices in low- and medium-intensity campaigns. [BACK]

21. See Lyons, Lowery, and DeHoog 1993: ch. 6. [BACK]

22. Svoboda 1995. [BACK]

23. Arnold 1990: 47. For a shortened version of this argument, see Arnold 1993. [BACK]

24. Ibid.: 72–73, 75, 77–78, 80–81, 272. [BACK]

25. According to Sniderman, Brody, and Tetlock 1991: 178, voters with low information use this up-or-down voting approach, which the authors find to be a reasonable strategy for making voting decisions. [BACK]

26. Popkin 1994: 92–94 [BACK]

27. Statistics are from "A Look at Voting Patterns of 115 Demographic Groups in House Races," New York Times, November 9, 1998, A20. Voter News Service collected the data in this article through election day exit polling of over 10,017 respondents in 1998. It is quite possible that these surveys overestimate party loyalty, but secret ballots permit no better measure of actual voting behavior. Also see Jacobson 1997: 92–93, which finds that roughly onequarter of voters cross party lines in congressional elections. [BACK]

28. Miller and Shanks 1996. The greater predictive strength of partisanship in congressional elections is consistent with the fact that partisanship is a cognitive shortcut. When voters are exposed to tremendous amounts of information in the unique environment of a presidential election, many decide that a candidate from another party would better represent their interests. In other words, these crossover voters made a decision other than the one that the simplistic partisan cue would have suggested. [BACK]

29. Miller and Shanks 1996. Though some misidentification takes place during exit surveys, partisan identity predicts voting preferences even years later. [BACK]

30. Quote is from Wattenberg 1994: ix, which presents a detailed description of the decline of political parties in America. A survey by Princeton Survey Associates found that 75 percent said they vote for different parties in response to this question: "When you vote in an election for national, state or local offices, do you always vote for candidates from one particular political party, or do you vote for candidates from different parties?" A Roper Center / Institute for Social Inquiry survey found that 65 percent reported voting for different parties in response to this question: "When voting in elections do you typically vote a straight ticket—that is for candidates of the same party, or do you typically split your tick—that is vote for candidates from different parties?" See The Public Perspective 9, 2 (February–March 1998), 49. According to University of Michigan surveys, ticket-splitting is far more common now than it used to be: 29 percent reported splitting in 1952 compared to 63 percent in 1996 (Tarrance, De Vries, and Mosher 1998: 35). [BACK]

31. Data are from Keith et al. 1992: 68, 72 [BACK]

32. B. Drummond Ayres, "Political Briefing," New York Times, October 27, 1998, A16. [BACK]

33. Dane Smith and Robert Whereatt, "Ventura Elected Governor," Online Star Tribune, November 4, 1998. [BACK]

34. Delli Carpini and Keeter 1996: 252–54. [BACK]

35. Abramowitz and Sanders 1998: 643–44. [BACK]

36. Arnold 1990: 53. [BACK]

37. Miller and Shanks 1996. Abramowitz and Saunders 1998 provides compelling evidence that partisanship is more ideological than it used to be. In their view, ideological debates between the two major parties made their differences

38. Evidence of the strength of partisanship influence on attitudes is provided by Campbell et al. 1960; Fiorina 1981; Green and Palmquist 1994; Jennings and Niemi 1981; and Miller and Shanks 1996, among others. [BACK]

39. Jacobson 1997: 71 [BACK]

40. George Gastil, a California Democrat and member of the Lemon Grove School Board, has discussed this possibility with me. The open primary does give minority-party voters the real chance of influencing elections, which is a real improvement over their utter powerlessness in solid majority districts. But their influence comes at the cost of closed-party primaries: a candidate’s party affiliation in the general election becomes a less useful cue as to the candidate’s actual stances because members of other parties may be responsible for the candidate’s victory in the primary. [BACK]

41. For the roll-call voting statistics, see King 1997: 167. On the similarities of the two parties, Michael Parenti observes: "Both the Democratic and Republican parties are committed to the preservation of the private corporate economy; huge military budgets; the use of subsidies, deficit spending, and tax allowances to bolster business profits; the funneling of public resources through private conduits, including whole new industries developed at public expense; the use of repression against opponents of the existing class structure; the defense of the multinational corporate empire and intervention against social revolutionary elements abroad" (Parenti [1974] 1995: 181). Parenti also points out that the two major parties have jointly implemented laws across the United States to thwart efforts to form viable third parties. [BACK]

42. On the increasingly ideological nature of party identity, see, e.g., Abramowitz and Saunders 1998, King 1997, and Miller and Shanks 1996. Unfortunately, this rigidification of the two major parties is further eroding public trust among those ideologically moderate Americans most distant from the two major parties (King 1997: 176). [BACK]

43. The Greens have recorded electoral victories in New Mexico, California, and elsewhere. In a special 1997 congressional election in New Mexico, the Green Party gained national notice by winning 17 percent of the vote. As a result, an obscure Republican minister was able to win a seat in a safe Democratic district by a narrow margin. The Democrats took the seat back in 1998 when they ran a less objectionable candidate. On the 1997 election, see Benjamin Sheffner, "Dems See Green in Special Election Loss," Roll Call, May 15, 1997. On the 1998 election, see Karen Peterson, "Johnson, Udall Win," Santa Fe New Mexican, November 4, 1998, A1. For an optimistic assessment of the prospects for progressive third parties in the United States, see Reynolds 1997. [BACK]

44. On the history of nonpartisan reforms, see Lee 1960. For criticisms of nonpartisan elections, see Sherrill 1998 and Hawley 1973. [BACK]

45. Given the relatively low turnout in primary elections, most voters do not choose to exercise their right to vote, which often makes primaries unrepresentative of the views of the average party member. Since the most active voters are more ideological, primaries often result in the election of extremist candidates who do not reflect the views of the nonvoting partisans. This is

46. "The incumbent clearly has the upper hand in dealing with the local party," which "no longer controls renomination or reelection" (Alford and Brady 1993: 154). [BACK]

47. Jacobson 1997: 13. In one such exception, officials in the Democratic Party had the opportunity to name their candidate in a special election for Congress in northern New Mexico in 1997. The chosen candidate, Eric Serna, went on to lose the general election because of unprecedented Democratic defection to a Green Party candidate. This simply illustrates that existing party organizations are not able to guarantee the nomination of a candidate whom even their own party members believe to be representative of their interests. See Benjamin Sheffner, "Dems See Green in Special Election Loss," Roll Call, May 15, 1997. [BACK]

48. Citizens are more likely to be influenced by their immediate family and friends and known neighbors than by the perceived norms of their neighborhood or community (MacKuen and Brown 1987). On the other hand, Robert Huckfeldt and John Sprague (1995: 180–84) found that among a person’s friends, the "closest" ones have less political influence. [BACK]

49. Inglehart 1997: 309. Ronald Inglehart’s surveys found that political discussions among friends were more common in Canada, West Germany, Norway, Sweden, Iceland, and the Netherlands, but in both the 1981 and 1990 surveys, people in the United States discussed politics with friends more often than did respondents in Britain, France, Belgium, and Ireland. [BACK]

50. Yum and Kendall 1995: 135. On a more theoretical level, MacKuen 1990 provides an explanation for why people tend to avoid political conversations with others who hold different views. With some exceptions, people tend to avoid "embarrassing" conflict and use subtle cues to determine the views of conversants before launching into sustained political exchanges. [BACK]

51. On influence in couples, see Huckfeldt and Sprague 1995: ch. 9. On the political influence of parents on children, a good example is Jennings and Niemi 1981. [BACK]

52. This process in American politics was recognized several decades ago, though little research addressed the issue again until the 1980’s. For a good example of earlier writing on this subject, see Katz and Lazarsfeld 1955. [BACK]

53. For the results of the focus-group study, see Delli Carpini and Williams 1994: 793; on the Pittsburgh study, see Mondak 1995: 82. [BACK]

54. Huckfeldt and Sprague 1995: ch. 9, but see Kenny 1998 for a critique of this interpretation. See also Yum and Kendall 1995; Scheufele 1999. [BACK]

55. Knoke 1990. McLeod et al. 1999 also finds a link between conversation and participation in public forums. [BACK]

56. Luskin 1990: 332. Also see research on political expertise (e.g., Krosnick 1990). [BACK]

57. The estimate of the size of the high-sophistication population is loosely based on Jacoby 1991 and 1995. It is convenient to further subdivide the public into low- and medium-sophistication in roughly equal proportions. For

58. Election research routinely grounds itself in studies of presidential elections such as The American Voter (Campbell et al. 1960) and The New Amer ican Voter (Miller and Shanks 1996). Perhaps tiring of such studies, election scholarship is turning increased attention to House, Senate and even gubernatorial races, but other state and local races still receive little attention. [BACK]

59. See McDermott 1997; Plutzer and Zipp 1996 [BACK]

60. Figures come from Burns et al. 1996: 171. These low-intensity races fall below the radar of almost all political science research on campaigns and elections. If the conventional wisdom is that a typical House race is a "low-information" race, then these elections are close to "zero-information" races. One of the few studies looking at near-zero information elections is Dubois 1984. [BACK]

61. See Champagne and Thielemann 1991 for Texas information and data on judicial name recognition. They found that between elections, fewer than 5 percent of people who say they vote in judicial elections could name the elected office held by some incumbent justices. [BACK]

62. For the Truman anecdote, see Brown and Giorgetti 1992. Gastil 1998 discusses how voters typically make decisions in these near-zero-information elections. [BACK]

63. Fishkin 1995: 10 [BACK]

64. My own mother, Janet Gastil, was one of many public officials who lost reelection bids in 1990 and subsequent local elections. She and her peers did not recognize the threat that the unknown candidates on the ballot posed until it was too late. Ralph Reed, the president of the Christian Coalition at the time, boasted of waging stealth warfare "under cover of night" where you "shimmy along on your belly." On Janet Gastil’s 1990 defeat and the stealth campaign phenomenon, see Seth Mydans, "Evangelicals Gain with Covert Candidates," New York Times, October 27, 1992, A1. Barry Horstman of the Los Angeles Times also wrote two detailed stories in 1992: "Crusade for Public Office in 2nd Stage" (March 22, 1992, B1) and "Christian Activists Using ‘Stealth’ Campaign Tactics" (April 5, 1992, A1). [BACK]

65. This figure is not meant as a formal model of voting, but rather as a means of clarifying some of the ways in which voter and campaign characteristics change how voters select candidates. Voting research has developed many detailed models, such as those advanced by Fiorina 1981 and Sniderman, Brody, and Tetlock 1991, but those tend to focus on the presidential election and other high-visibility federal races. By splitting elections and voters into

66. Incumbents concerned with reelection devote much of their time to meeting voters or even just inviting voters to such meetings. After shadowing congressional incumbents, Richard Fenno decided that many voters thought of "representation" in terms of access, familiarity, and trustworthiness. The average citizen appeared more concerned with officials' readiness and willingness to talk and "extra policy behavior" than with actual policy decisions (Fenno 1978: 240). In this sense, voters who select a candidate because they have met him or her during a precinct walk or a community forum might reason that the candidate will prove more accessible in public office than the opponent. Even minimal familiarity—the mere recognition of a candidate’s name—might suggest to a voter that the candidate will make as much effort to stay in contact while in office as he or she did through yard signs and political mail during the campaign. [BACK]

67. In a review of campaign advertising research, Ansolabehere and Iyengar 1996 found that decent ads can give a typical candidate a 7–8 percent boost. [BACK]

68. Morris Fiorina (1981: 208) estimates that 40 percent of voters in congressional races can give meaningful appraisals of congressional candidate voting records. [BACK]

69. Some might argue that ideology is a factor only in high-intensity, partisan elections. I argue that for as much as a third of the electorate, it does play a role in other races. Paul Raymond’s (1992) study of the 1985 Lexington, Kentucky, city council elections provides support for the influence of perceived candidate ideology, along with name recognition and candidate personality, background, and competence as district representative. [BACK]

70. On the philosophy of adversary democracy in America and its alternative, see Mansbridge 1983. [BACK]

71. "Negative advertising" has been "a major feature of recent presidential elections" and "is now a factor in state and local campaigns" as well (MacGregor et al. 1996: 88). See Jamieson 1992 for a brief history of modern negative campaigning. Scher 1996 argues that name-calling and attack advertising ("mudslinging") are old American traditions. [BACK]

72. On the advantages of congressional incumbency, see Jacobson and Kernell 1981; Alford and Brady 1993. [BACK]

73. Jacobson 1997: 100. The power of the "personal vote" that incumbents develop through both mere exposure and actual constituent service is discussed at length in Cain, Ferejohn, and Fiorina 1987, which stresses the value of "warm, humane interventions" by elected officials but concludes: "Without constraints … such behavior can corrode the conduct of democratic government by undermining the ability of that government to act in ways that improve the lot of its citizens" (229). Cultivating and maintaining the "personal vote" can make elected officials servants of particularistic interests rather than true public representatives. [BACK]

74. Jacobson 1993: 119. The 1997–98 figures are official numbers compiled by the Federal Election Commission, available at http://www.fec.gov. Earlier

75. On incumbent intimidation tactics, see Jacobson 1997: 43–47. All spending data are provided by the Federal Election Commission. For congressional spending totals in every Senate and House election, plus the presidential and gubernatorial races, see Barone and Ujifusa 1998. [BACK]

76. Dan Morain, "Wealth Buys Access to State Politics," Los Angeles Times, April 18, 1999, A1. [BACK]

77. "The cost of running for local office is much lower than running for Congress or the governorship," but "campaigns are becoming increasingly expensive at all levels" (MacGregor et al. 1996: 88). The New Mexico race was for State Senate District No. 39: the incumbent, Senator Liz Stefanics, the only "out" lesbian in the legislature, raised $71,000, and her opponents spent $63,800 and $149,000. Stefanics's sexual orientation became an issue in the race (Mark Oswald, "Griego Edges Stefanics," Santa Fe New Mexican, June 5, 1996, A1). [BACK]

78. Jacobson 1993: 131. I witnessed this cost firsthand while assisting my mother with her 1992 and 1994 congressional campaigns. In effect, she had to campaign continuously for three years, and few citizens can afford to take off so much time from work. [BACK]

79. Rachel Zoll, "Congressional Races See Little Competition," Chat tanooga Times, August 5, 1998. The artful drawing of district boundaries is also a common explanation for entrenched incumbencies, but it does not explain why incumbents handily win their party primaries. [BACK]

80. Jacobson 1993: 132–33. [BACK]

81. Burns et al. 1996: 90. [BACK]

82. Barone and Ujifusa 1998. [BACK]

83. Ibid.: 906–7. [BACK]

84. One might argue that high reelection rates partly reflect the "forced retirement" of members who would have been defeated had they not resigned. On the contrary, analysis of congressional behavior suggests that resignations have more to do with thwarted agendas and waning interest in holding office than imminent defeat. It is the cost of campaigning weighed against the declining perceived value of holding office (rather than simply the fear of defeat) that motivates voluntary departures (Jacobson and Kernell 1981; Moore and Hibbing 1998). [BACK]

85. Alford and Brady 1993: 155 [BACK]

86. Ibid.: 154. On victory margins, see Jacobson 1997, ch. 3. Although the incumbency advantage remains strong, Jacobson argues that it has actually waned compared to the 1980s. [BACK]

87. For example, Krebs 1998 found that incumbency offers a tremendous advantage to Chicago city council candidates. [BACK]

88. See, e.g., Dionne 1991; Mathews 1994: chs. 1, 4; Tolchin 1995; Sanders 1990. [BACK]

89. Mitchell 1996: 49, 77. Data are from the General Social Survey, conducted by the National Opinion Research Center. [BACK]

90. Schudson 1998: 302. Later, Schudson adds that he doubts "trust" is an "intelligible indicator of anything" and focuses instead on "social capital," which also proves to be an unwieldy concept (307). [BACK]

91. For 1964–88 data, see Rosenstone and Hansen 1993. Complete National Election Study data are available at http://www.umich.edu/~nes/nesguide/toptables/tab5c_1.htm and http://www.umich.edu/~nes/nesguide/toptables/tab5c_2.htm. Asked why they distrusted government, eight in ten members of a national sample group cited its inefficiency ("wastes money," "spends on wrong things"), roughly two-thirds cited the undue influence of special interests, and over 60 percent said that "politicians lack integrity" (Blendon et al. 1997: 250). [BACK]

92. Hibbing and Theiss-Morse 1995: 45, 166. Approval ratings are the percentage of respondents who "strongly approve" or "approve" of Congress, as opposed to those who "disapprove" or "strongly disapprove." In-depth interview research by Arthur Sanders (1990) also found that a small sample of citizens in Utica, New York, distinguished between trust in leaders versus trust in the system as a whole. [BACK]

93. Craig, Niemi, and Silver 1990: 296 [BACK]

94. "[V]oting has declined for the last thirty years," David Mathews (1994: 29) argues, for example. The League of Women Voters also sees a "long-term downward trend in voter participation that began in the early 1960s." See "Charting the Health of American Democracy," June 1997, available at http://www.lwv.org/report.html. One reason that voting is perceived to be in decline is that comparison usually begin roughly ten years prior to the lowering of the voting age in 1972. That is why I present figures starting in 1972. Everett Ladd (1996: 5) has pointed out that turnouts at recent presidential elections do not seem particularly low in comparison with elections going back to 1932. The 1992 election, for example, falls right in the middle of the range. [BACK]

95. It is difficult to calculate precise turnout figures, but the Federal Election Commission numbers are as reliable as any. Estimates of the overall voting-age population sometimes include resident aliens and felons, who are not eligible to vote, but they also normally make no correction for the U.S. Census's undercounting of minorities and the homeless population. For a brief discussion of these issues, see "The Not So Indifferent Voter," Wilson Quarterly 22 (Winter 1998). To the extent that turnout has declined, this drop may be due to a generational shift, as the steady voters of the New Deal era pass away (Miller and Shanks 1996). [BACK]

96. Engstrom and Caridas 1991 found compelling evidence that black voters drop-off in New Orleans judicial races at a greater rate than white voters. National drop-off figures are from the U.S. Bureau of the Census's Statistical Abstract of the United States, 1997, table 464. On low turnout in state and local elections, see Burns et al. 1996. [BACK]

97. Such a view can be found in early pluralist theories of democracy, which found that despite nonvoting and other alleged "problems" with American democracy, the political system was accurately incorporating the public’s interests into public policy. See, e.g., Dahl 1961. [BACK]

98. Ironically, nonvoting is a powerful form of encouragement for elected representatives who believe that they will not be held accountable for their

99. It is reassuring that Sellers 1998 found that campaign ads are more effective, on average, when the candidates can actually substantiate their claims. Backed up claims, both positive and negative, appear to have more influence on voters than do unwarranted attacks. If campaigns acted on these findings, it might improve campaign discourse, but it would not make focused debate any more likely. Though accuracy has merit, one can be honest even while discussing irrelevant issues and avoiding a genuine exchange of viewpoints with an opponent. [BACK]

100. As an illustration of this point, campaign handbooks devote relatively little attention to substantive policy debate with opponents (e.g., Beaudry and Schaeffer 1986; Shea 1996; Grey 1994). Simpson 1996 treats the subject briefly under the subheading "Staged Events." [BACK]

101. March and Olsen 1995: 45. Along similar lines, Jackman and Miller 1996 argues that political institutions structure citizens' incentives so powerfully that they can prove more influential than culture on individuals' political behavior. I share this view, and that is why I recommend institutional changes in chapter 7. [BACK]

102. Freie 1997 found that even direct participation in conventional American political campaigns has been shown to cause many volunteers to become more alienated from the process. In other words, participation in a flawed system fails to restore confidence in the system. [BACK]

103. Quotation from John Anderson in the foreword to Barber 1995: x. In support of this view, Junn 1991 found that political participation tends to have a positive effect on political knowledge. In other words, when less sophisticated voters stop voting, their nonparticipation is likely to stunt the growth of their political knowledge. [BACK]

104. There is considerable debate about whether the nonvoting population differs from the voting population in terms of average attitudes (Bennett and Resnick 1990; Verba, Schlozman, and Brady 1995). Following Zaller’s (1992) public opinion model, this correspondence is not surprising because nonvoters tend to have lower levels of political sophistication and thus report attitudes mirroring the average messages conveyed in the media. What is less controversial is whether the voting public demographically matches the nonvoting public: compared to nonvoters, average voters are older, more affiuent, and have more formal education (Verba et al. 1995). Polling data notwithstanding, it is reasonable to expect these demographic differences to make nonvoters more concerned about class issues—social welfare, job training, and other government programs that aid the economically disadvantaged. Consistent with this view, Parenti [1974] 1995 points out many ways in which the U.S. government overrepresents the interests of the wealthy, and Hill, Leighley, and Hinton-Andersson 1995 shows that social welfare programs fare better when lowerclass voters turn out to vote in greater numbers. Thus, states with more liberal party elites tend to encourage greater voter registration and turnout because it serves their interests (Jackson, Brown, and Wright 1998). [BACK]