The New Party Professionals? An Initial Look at National Convention Delegates in 2000 and Over Time

Transcription

1 The New Party Professionals? An Initial Look at National Convention Delegates in 2000 and Over Time Thomas Carsey Florida State University John Green University of Akron Rick Herrera Arizona State University Geoffrey Layman Vanderbilt University Abstract Party scholarship suggests that the two major political parties have recently gone through a pattern of decline and resurgence in their relevance to American politics. Numerous factors have been advanced to explain this changing relevance of parties, but there is currently no compelling theory of parties that explains this entire cycle. One explanation is the decline of party professionalism among party activists due to the party reforms of the 1970s, and the recent development of a new kind of professionalism, which combines the issues focus of amateur activists with the organizational and electoral concerns of professionals. Using surveys of national convention delegates from 1972 to 2000, and focusing especially on the newly gathered 2000 data, we investigate the emergence of professional ideologues, and find considerable evidence of this phenomenon. Authors listed in alphabetical order. Prepared for delivery at the 2003 Annual Meeting of the American Political Science Association, August 28 - August 31, Copyright by the American Political Science Association

2 Party scholarship suggests that the two major political parties have recently gone through a pattern of decline and resurgence in their relevance to American politics. Although the dimensions, character, and meaning of this change are hotly debated (see Coleman 1994 and Frendreis 1994), the basic pattern is widely accepted: party organizations became less relevant to the conduct of campaigns and to the political behavior of the mass public and government officials after But, by the 1990s, the parties had recovered much (but not all) of their relevance in each of these areas, albeit often in new forms (Cohen, Fleisher, and Kantor 2001). For instance, the development of service parties and innovations in campaign finance allowed national party organizations to pay a significant role in campaigns (Herrnson 1988; Herrnson and Green 2003). Meanwhile, partisanship became more important in mass voting behavior and among public officials (Bartels 2000; Hetherington 2001), such as members of Congress (Rohde 1991; Sinclair 1992; Bond and Fleisher 2000). Numerous factors have been advanced to explain this changing relevance of parties (Cohen and Kantor 2001). However, there is currently no compelling theory of parties that explains this entire cycle of decline and resurgence. In this paper, we sketch the outlines of such a theory with a special focus on the role of professionalism among party activists. The intuition behind this sketch is that party decline was encouraged by the decay of traditional professionalism among party activists after 1960, and that party resurgence may be fostered by the rise of a new kind of professionalism, which combines the issues focus of amateur activists with the organizational and electoral concerns of professionals. We investigate this possibility by reviewing the characteristics of political style among delegates to the national party conventions in

3 Toward a Dynamic Theory of Party Relevance Despite Madison s famous warning in Federalist #10 about the mischiefs of faction, scholars have long argued that political parties play a central role in representative democracy. Political parties simplify and regularize the process of collective decision making both for voters on election day and members of government after the election is over (e.g., Downs 1957, Aldrich 1995, Cox and McCubbins 1993). Seen from this perspective, party organizations are created by rational politicians seeking to resolve several collective action problems inherent in representative democracy. If effective, these extra-governmental institutions bring a high degree of coherence to politics and policy, serving the broader interests of the citizenry. For this reason scholars tend to be highly sympathetic to strong party organizations, even if the reality of party politics often falls far short of the ideal. Thus, party scholars are fond of citing the call of the American Political Science Association in 1950 for more responsible parties (Committee on Political Parties, 1950) and E.E. Schattschneider s statement,... political parties created democracy and that modern democracy is unthinkable save in terms of the parties. (1942,1). Aldrich (1995, 22-24) argues that there are three major collective action problems that prompt the creation of party organizations by politicians: the need to choose candidates from among ambitions rivals, to take positions on issues from among competing alternatives, and to organize campaigns to compete for votes in elections. Each of these decisions is important in its own right, but they are also relevant to one another: candidates and issues can matter in mobilizing voters, issues often motivate candidates, and so forth. The individuals most active in resolving these collective action problems are assumed to be those who will benefit most directly from the resolution. Such benefit seekers include individuals seeking public office, those seeking particular public policies, and those seeking influence in, and benefits from, the 3

4 marshalling of campaign resources. Most such formulations focus on office-seekers as a driving force behind parties and party change, presumably because holding office is a principal source of benefits, especially for those seeking office, but for others as well (Schlesinger 1991). Indeed, office-seekers need the help of other benefit-seekers in order to obtain office, and presumably, other activists need office-holders to obtain the benefits they seek (Aldrich 1993). Thus, the motivations and characteristics of the diverse activist corps are critical to resolving the collective action problems, and hence the form party institutions take. By the same logic, changes in the motives and characteristics of activists will be associated with changes in party institutions. From this perspective, party institutions are the product of adaptation, whereby the activist corps develops rules and norms for making necessary decisions that maximize the provision of benefits, given the legal and social context in which it operates. Of course, one important critical context is the constitutional structure of the polity, which in the United States fosters the two-party system which tends to funnel political conflict into a single dimension. Social and economic circumstances are also important contextual factors to consider. Indeed, given the dynamic nature of American society, some steady, incremental adaptation is necessary. However, sometimes major changes occur that require more dramatic adaptations that may take longer to achieve. In the simplest terms, such adaptations are successful if they resolve the collective action problems among activists well enough to produce a minimal level of the desired benefits. There are, of course, many factors at work in party success, but all else considered, party organizations are likely to be more relevant to politics when they foster a high degree of coherence among the activist corps (Herrera 1993). Popularly known as party unity (Sullivan ), such 4

5 coherence involves a high degree of agreement on candidates, consensus on issues, and cooperation in campaigns. In fact, the major parties have frequently achieved such coherence for extended periods of time, but they have also often lapsed into periods of incoherence. This theoretical sketch highlights several features of a dynamic theory of party relevance to politics. First, activists are central to the formation, functioning, and change of party organizations. Second, the motivations and characteristics of the activists are the raw materials of institutional adaptation. Third, the rules and norms of party decision-making allow for the maximum benefits to the activist corps in a given context. And fourth, parties are most relevant to politics when the rules and norms foster a high degree of coherence in the activist corps. We can now apply this sketch to the pattern of decline and resurgence of party organizations over the last forty years. National Conventions and Party Relevance There is a long scholarly tradition of studying national convention delegates to assess the operation and success of party institutions (see Miller and Jennings 1986, and Herrera 1993). After all, the national conventions are the formal decision making bodies for the Democratic and Republican parties in national politics. Although their exact functions have changed over time, the conventions can claim the fullest representation of the party activist corps at any given moment. Indeed, national conventions are still a good place to assess the operation and success of party institutions today as they were fifty years ago. Since the 1960s, the national conventions have undergone dramatic rule changes. For one thing, conventions no longer deliberate over the choice of presidential nominees: such decisions are effectively made in primary elections, where competition between rival candidates influences 5

6 the selection of the convention delegates. The development of such candidate-centered politics was already underway in the 1960s, but it was confirmed and extended with changes in formal party rules beginning in The primary impetus for these changes were issue disagreements within the activist corps, but such disputes were exacerbated by a host of changes outside of party institutions, including new campaign technologies, an expanded role of money in campaigns, increased power of interest groups, and the politicization of a new set of cultural issues (Cohen and Kantor 2001). Thus, the party reforms of the 1970s can be understood as an attempt by the activist corps to adapt to these new realities and maximize the flow of benefits (Aldrich 1995: ). However, the initial evaluations of this adaptation were not positive: many scholars believe the reforms were not a success because they reduced the capacity of party institutions to develop high levels of coherence (Ceaser 1982; Polsby 1983). In this regard, two related problems have been identified. First, the new rules opened the national conventions to more diverse activist corps, making it more difficult to make effective decisions on issues. Bitter debates ensued over the dimensions of political conflict, with the new activists seeking to displace the social welfare dimension with a cultural and foreign policy dimension. A second and related problem was a change in the norms in the activist corps. The reformed party included a large portion of activists who combined a strong interest in issue benefits with an uncompromising style the legendary amateurs of the party literature. The amateurs battled the professional activists, who were more concerned with other kinds of benefits. This clash of norms made it especially difficult for party institutions to cope with the substantive disagreements over the dimensions of political conflict. These disagreements over issue dimensions and party norms produced something of a 6

7 scholarly consensus on the reformed parties: the relevance of parties to national politics had declined. However, this perspective has been challenged by the politics of the 1980s and 1990s, when a variety of evidence revealed a resurgence in the relevance of party. While far from complete, this resurgence begs an important question: if party reform contributed to a decline in the political relevance of party by changing the character of the activist corps, could resurgence result from a new set of changes in the character of the activist corps? We think this possibility is worth exploring. It could be that the activist corps has adapted to the reformed party institutions in increasingly successful ways. After all, the reformed parties are now several decades old old enough for rational politicians to find new ways to resolve the collective action problems in the new context. Two adaptations seem promising: a shift from conflict displacement to conflict extension with regard to the dimensions of conflict, and the development of a new professionalism in the activist corps. Conflict Extension. Conflict extension describes a process whereby the two political parties become increasingly polarized across multiple issue dimensions (see Layman and Carsey 2002a; 2002b). The idea that multiple divisions or cleavages define mass beliefs about politics and policy has a long-standing tradition (e.g. Lipset and Rokkan 1967; Sundquist 1983; more recently Layman and Carsey 2002a; Miller and Schofield 2003). However, the standard view has been that in a two-party system, only one political cleavage can and will dominant political conflict at any particular point in time, thus defining the party system. Party change occurs when conflict on the dominant cleavage or issue dimension is displaced by conflict organized along a new (Sundquist 1983) or re-emerging (Miller and Schofield 2003) dimension. In contrast, under conflict extension, multiple dimensions of conflict are rationalized, so that parties issue positions become more consistent ideologically. The emergence of crosscutting issues may well 7

8 create disruption within parties, but they are eventually absorbed into a broader dimension of conflict, restoring stability. The result is greater issue consensus within parties and greater issue differences between the parties. While unpacking the process of conflict extension among activists is outside the scope of this paper, there is considerable support for it in the literature (see Layman and Carsey 2002a, 2002b; Abramowitz and Saunders 1998). The resulting outcome is important for party adaptation: conflict extension allows activists to struggle less over the issue benefits within their own parties, while at the same time raising the costs of losing to the opposition on election day. The New Professionalism. Conflict extension could push even amateur activists to become increasingly concerned with professional goals. The parties may need to maintain consistent ideological stances across a range of issues precisely in order to assemble and mobilize a coalition of electorally valuable activists. In fact, what earlier scholars viewed as the competing goals of securing issue and process benefits may have become increasingly complementary. The prime carriers of this new combination of norms might be labeled as professional ideologues: activists who maintain a strong interest in issues, but who also seek benefits from a the process of mobilizing resources and winning elections. Indeed, professional ideologues would have largely resolved the tension that existed between amateur and professional activists inflamed by party reforms of the 1970s. The remainder of this essay will investigate the possibility that this new professionalism has developed, but first we need to briefly review the concept of party professionalism. Investigating Party Professionalism Party professionals are central to the lore of American politics, especially in the 8

9 golden age of parties in the late 19 th century, when local party leaders dominated politics (Silbey 1994). Professionals are still recognized in party affairs, although in recent times the term has broadened to include full-time office-holders, political consultants, and the technical staffs of state and national party bureaucracies (Green 1994). But the essential meaning is the same: the element of the party activist corps that seeks tangible benefits from the process of resolving the collective actions problems for which parties are created. Although professionalism has been both praised and reviled, it has received little scholarly attention in its own right. Instead, party professionalism has been studied in contrast to what has been thought of as its opposite: amateurism. The study of party amateurs and professionals began in earnest when scholars noticed a shift in the norms of party activists in the 1950s and 1960s. Focusing on the Democrats, Wilson (1962) identified a rivalry between amateurs and professionals in local parties. This distinction was principally about political style, namely, the norms for making party decisions, rather than issue positions or ideology. So, amateur Democrats were not so much liberals as they were advocates of greater rank-and-file participation in party decision-making, and thus a greater stress on issues and principles. Indeed, Wilson argued that there were likely to be amateurs among Republicans as well. Shortly thereafter, Wildavsky (1965) found evidence of this proposition at the 1964 Republican convention, where the conservative purists supported Goldwater, while the moderate politicians or pragmatists did not. These insights prompted a series of studies of local party activists and national convention delegates (for a useful summary, see Hershey and Beck 2003), most of which found some value in the amateur/professional distinction before the party reforms of the 1970s. Studies of the convention delegates in 1972 and after (Kirkpatrick 1976, Miller and Jennings 1986, 9

10 Herrera 1993, 1995; Layman 2001) solidified amateurs and professionals as ideal types in the literature (see for example, Beck and Hershey 2003). But as with any useful distinctions, the amateur/professional distinction has proven to be problematic in its application. Miller and Jennings describe the literature as a quagmire noting great confusion over the way scholars measured this concept, its elements, and referents (1986:90). Indeed, a thorough attempt to validate and test the concepts by Hofstetter (1971, 1973) found three distinct elements: incentives, norms, and campaign commitments. All three concepts are part of the definition of amateurs and professionals used in the discipline and scholars often use these definitions interchangeably. The most basic concept is incentives for activism. From a theoretical perspective, incentives are interesting because they are a direct measure of the benefits activists seek. In the party literature most scholars have followed Clarke and Wilson s (1961) tripartite distinction between purposive, material, and solidary incentives (Wilson 1962; Eldersveld 1964; Conway and Feigert 1968; Hofstetter 1971; Kirkpatrick 1976; Abramowitz, McGlennon, and Rapoport 1983; Costantini and King 1984; Miller and Jennings 1986; Miller, Jewell, and Sigelman 1987; Shaffer and Breaux 1998), although some scholars have identified more than these three types of incentives (Kirkpatrick 1976; Costantini and King 1984; Miller and Jennings 1986). Indeed, this typology informed Clarke and Wilson s view of the amateur Democrats. Purposive incentives focus on benefits tied directly to the manifest purpose of an organization, and in the case of amateurs, the purpose was promoting issues and principles in politics. For this reason, amateurs have long been seen as more ideological than professionals. In contrast, the professionals were characterized more by solidary incentives, namely benefits from social interaction, such as personal recognition, friendship, and the excitement of 10

11 party politics. Because it taps into a sense of group identification and loyalty, attachment to a political party and commitment to its success also has been classified as a solidary incentive by numerous researchers (Conway and Feigert 1968; Kirkpatrick 1976; Shaffer and Breaux 1998; Layman and Carsey 1998). Professionals are also characterized more by material incentives, tangible benefits to individual activists, such as patronage jobs, political preferments, business or professional contacts, and developing a career in politics. Because solidary and material incentives are often directly connected to electoral success, professionals were seen as more committed to winning elections. Of course, victory at the polls could be thought of as a central purpose of parties and thus a purposive incentive. Indeed, some studies have found that purposive incentives were associated with a desire to win elections among party activists (Abramowitz, McGlennon, and Rapoport 1983; Abramowitz and Stone 1984). A second concept is norms of party decision-making. From a theoretical perspective, norms are interesting because they reveal activists views on appropriate decision-making in parties. Here the party literature has tended to use Wildavksy s (1965) terminology of purists and pragmatists. Purist norms include the belief that the party should pursue a coherent ideological agenda and nominate candidates who uncompromisingly support that agenda, and that individual party members should stick to their ideological guns rather than compromise to encourage party unity or further the party s electoral chances. In contrast, pragmatic norms include the beliefs that electoral victory as more important than ideological purity, that broad electoral appeal is more important than ideology in the selection of party candidates, and that party members have a responsibility to support the party and help unify it regardless of whether it supports ones preferred policy positions. Indeed, many of the most useful empirical findings in the literature have employed 11

12 measures of such norms (Soule and Clarke 1970; Hofstetter 1971; Soule and McGrath 1975; Roback 1975, 1980; Jackson and Hitlin 1976; Kirkpatrick 1976; Stone and Abramowitz 1983; Miller and Jennings 1986; Prysby 1998). Amateurs are thought to be purists and professionals, pragmatists. But as DeFelice (1981) demonstrated, pragmatism and professionalism are not necessarily the same thing, especially if one defines professionalism as commitment to party organizations. Or as Roback put it, party leaders could be as guilty of a passionate professional purism as of passionate amateurism (1975:466). The third concept might be called campaign commitments, namely, what activists seek to achieve through their participation in party affairs, including working for issues, interest groups, party organizations, or candidates (Miller and Jennings 1986). Although campaign commitments have not employed in studies of convention delegates, the concept is commonplace in the literature. Beginning with Wilson s amateur Democrats, many scholars have believed that incentives and norms were important precisely because they motivated party activists toward different ends, thus producing very different kinds of organizations. So, for example, the tensions between holding to issue positions and winning elections are not just norms, but also behavioral commitments by activists (Hofstetter 1973). Of course, amateurs are assumed to be committed to pursuing issues and professional to supporting the party. The rise of candidate-centered politics complicates the picture, since both amateurs and professionals might well support candidates. From a theoretical perspective, campaign commitments are interesting because they reveal the preferences of activists in a particular election. The variety of concepts associated with the amateur-professional distinction helps explain an important finding: although the distinction is often portrayed as a sharp contrast, researchers have regularly found activists to share a mix of amateur and professional traits. 12

13 Indeed, our professional ideologues have precedent in the literature, such as Soule and Clark s (1970:891) semi-professionals. To begin sorting through these concepts, we will investigate incentives, norms, and campaign commitments, and use them to test the major findings of the literature on the relationship between amateurs/professionals and the traits of activists. Major differences between our findings and the literatures may well suggest changes in the characteristics of the party activists and/or how these characteristics fit together. Taken as a whole, the literature suggests the following expectations in this regard. Demography and Group Memberships. Overall, the literature suggests that amateurs are younger, better educated, and have higher incomes than professionals, and that women are more likely than men to be amateurs. Among the Democrats, Jews and seculars are likely to be amateurs, while among Republicans, committed evangelical Protestants are likely to be so as well. Finally, amateurs are more likely to belong to ideological or policy-oriented groups, and professionals to belong to community or professional groups. Support for Party. Amateurs are less attached to and less supportive of their parties than are professionals. Political Participation. Amateurs are political short-timers or occasional activists who have less political experience than professionals; are less likely than professionals to remain active in the party once the candidates advocating their particular agenda leave the political stage; are less likely than professionals to participate in the activities that help the party and its candidates win, particularly in the less-glamorous, lower-stimulus campaigns below the presidential level; and, for all of these reasons, are less likely than professionals to hold elected office or official party positions. Issue Positions. Amateurs tend to have more ideologically extreme issue positions than 13

14 do professionals, especially on cultural and foreign policy issues. The 2000 National Convention Delegate Study Most of the analysis in this paper is based on a survey conducted from the fall of 2001 through late spring of delegates to the 2000 Republican and Democratic national party conventions. The survey was modeled after the Convention Delegate Studies (CDS) conducted by Warren Miller and others in 1972, 1980, 1984, 1988, and The universe of delegates in the study was 6,033; 4,284 Democrats and 2,049 Republicans. 2 Our response rate from the 2000 delegates was 39%. Included in our sample are 1,547 Democrats (38% response rate) and 780 Republicans (40% response rate). 3 Though our response rates are respectable and comparable to previous surveys of delegates, we nonetheless dealt with setbacks that undoubtedly negatively affected our response rate. We proceeded as follows. We mailed advance letters to the delegates in early October 2001 followed by the first wave of questionnaires at the end of October (Republicans) and beginning of November (Democrats). Our first reminder postcard was mailed on November 30. The second wave of questionnaires was sent two weeks following the reminder and a final letter 1 This survey was funded by a grant by the Ray C. Bliss Center for Applied Politics at the University of Akron. The 1972 CDS was conducted by Warren E. Miller, Elizabeth Douvan, William Crotty, and Jeane Kirkpatrick. Miller was joined by M. Kent Jennings for the 1980 CDS, by Jennings for the 1984 and 1988 studies, and by Richard Herrera for the 1992 study. 2 We also collected data from respondents to the 1992 Convention Delegate Study creating a panel of 923 respondents. There were a total of 1888 respondents to the 1992 survey for which we had correct address information resulting in a response rate of 49%. Some of the respondents in the panel were also delegates to the 2000 conventions and are included in the analysis. Respondents to the 1992 and 2000 surveys who were not delegates in 2000 were not included in this analysis. 3 The response rates are computed taking into account delegates to the 2000 conventions who were deceased at the time of our survey and delegates from whom we did not have updated address information. 14

15 and questionnaire was sent in January. As is evident from the dates of the survey, we were in the field in the aftermath of the September 11 tragedy. In fact, we were ready to send out advance letters to the delegates on that date and postponed the mailing for obvious reasons. Soon after the hijackings, the threat of anthrax in the mail struck the nation s postal service. Cognizant of the potential dampening effects the anthrax scare might produce, we took steps to alleviate the potential fears of the respondents, including altered advance and cover letters and notice on the envelope indicating the contents. Owing to the precautions instituted by the postal service to deal with the possibility of anthrax being transported through the nation s mail, delivery was slowed considerably. For us, that meant that returned surveys were arriving a month after their postmarks. Recipients of the survey were phoning us to complain about receiving reminders when they claimed they had already returned their completed surveys. The length of time between mailings and returned questionnaires was much longer than in previous years, no doubt a result of the slowed postal service. We also received many surveys that were unusable due to the radiation treatment they received by the postal service to insure against contaminants. These surveys arrived in plastic packets and in crumbled pieces. Despite all these problems, the only bias in the sample that we have observed is a lower response rate from the Washington, D.C. area. This means a lower than optimal number of African American Democrats in the sample. The data collection ended in June Measuring Amateur-Professional Traits in 2000 Incentives. The 2000 Convention Delegate Study (CDS) included the same incentive items that have been included in all of the previous CDS surveys in addition to three new items designed to tap into the purposive motivations of advancing issue positions, values, and group 15

16 goals (see Table 1 for the incentives items and a designation of the items new to 2000). Most of the items fall clearly into the traditional categories of purposive, solidary, and material incentives. But some of the items, such as party attachment and fulfilling civic responsibility, do not fit as obviously into these categories and have been placed differently by different researchers. To help determine the appropriate place for these items, we first conducted an exploratory factor analysis of all of the incentive items. Then, based on the results of that analysis, we conducted confirmatory factor analyses of just the items we denoted as purposive, just the solidary incentives, and just the material incentives. The measures of each type of incentives are the factor scores from that second round of analyses. [Table 1 about here] Table 1 shows the results of both types of factor analyses. 4 The exploratory analysis of all of the incentive items produced only three factors with eigenvalues greater than one, and the three factors are very consistent with the Clarke-Wilson typology. The first factor clearly taps into solidary motivations, with social incentives such as politics as a way of life, friendships and social contacts, and the fun and excitement of electoral politics loading very strongly. Some researchers have placed attachment to party in a separate category of partisan or allegiance incentives (Costantini and King 1984; Miller and Jennings 1986), but the item s very strong 4 All of our analyses with the 2000 CDS are for delegates to the 2000 national conventions only. The exploratory factor analysis of all of the incentives items uses principal components extraction and varimax rotation. The factor loadings in the confirmatory analysis are from unrotated principal components factors. Our measures of purposive, solidary, and material incentives are the scores from these unrotated factors. Readers familiar with the CDS series will notice that we did not include the incentive of wanting to see particular candidates elected in the analysis presented here or in our incentive measures. We excluded it for two reasons. First, it is a very important incentive to nearly all activists (more than 96 percent of both parties delegates labeled it as quite important or extremely important), and thus does not help us to distinguish between various types of incentives or activists. Second, it is far from clear in which category it belongs. Electing candidates has often been labeled as a purposive motivation (cf. Miller and Jennings 1986), but also taps into professionals and party loyalists primary goal of winning elections. Thus, it may not be surprising that in some of the CDS surveys, including 2000, this item loads more strongly with solidary incentives than with purposive incentives in factor 16

17 loading with solidary incentives supports the work that places it in that category. The literature also lacks clear guidelines for the placement of fulfilling civic responsibility as a motivation for participation. Some work labels it as a purposive incentive (Kirkpatrick 1976; Miller, Jewell, and Sigelman 1987), while other scholars have placed it in a partisan or allegiance category (Costantini and King 1984; Miller and Jennings 1986). However, a sense of civic or community obligation clearly taps into the sense of collective identification and belonging that is at the heart of solidary incentives, and here it loads very strongly with the solidary items. One surprise is that the desire to be active in politics because friends and family members are active loads more strongly with the material incentives than with the solidary incentives. However, because this item does not load that strongly on any of the three factors and because it clearly should elicit solidary considerations, we include it with the solidary incentives in the confirmatory factor analyses. Each of these separate analyses of solidary, purposive, and material incentives produces only one factor with an eigenvalue greater than one and nearly all of the items in each analysis loads very strongly on the single factor. Purposive incentives such as a desire to get the party and its candidates to support one s policy positions, working for issues, and advancing one s values and group goals load strongly on the second factor. The material motivations of making business and professional contacts, having a personal political career, being close to important people, and gaining recognition have strong loadings on the third factor. Norms. To measure purist-pragmatist norms, we conducted a principal components factor analysis of levels of agreement with six statements about the appropriate behavior of the party and its members, three of which are worded so that agreement signifies purist norms and three of analyses. 17

18 which are worded so that agreement signifies pragmatist norms (see table 2). The analysis produced two factors with eigenvalues greater than one. After rotation, the items worded in a purist direction loaded strongly on one factor and the items worded in a pragmatic direction loaded strongly on the other factor. However, because all of the statements were designed to capture the single underlying concept of purist-versus-pragmatist norms, nearly all of the literature measures purist-pragmatist norms along a single continuum, and the eigenvalue of the second factor was only 1.06, we discarded the second factor. 5 Our measure of purist-pragmatist norms is the factor score from the first unrotated factor, and ranges from most purist to most pragmatic. [Table 2 about here] In addition to the questions about delegates reasons for being involved in politics in general and their norms about party and activist behavior, the 2000 CDS asked respondents the extent to which their involvement in the 2000 presidential campaign was motivated by commitments to the party, to the candidate, to an issue position, or to a specific group. 6 Table 3 shows the percentage of delegates to the Democratic and Republican national conventions from 1972 to 2000 who said that a lot of their activity in the presidential campaign of that year was 5 Green and Citrin show that nonrandom measurement error in a series of positively and negatively worded questions "may attenuate the observed correlations between items worded in opposite evaluative directions" (1994:278), and lead to the conclusion that favorable and unfavorable attitudes toward the same object are not strongly related. We have a slightly different case here with some items worded in a purist direction and others in a pragmatic direction, rather than positively and negatively worded items. However, it is very similar to the one that Green and Citrin identify. As a check, we constructed a factor model allowing for measurement error in our norms items (computed with AMOS 4.0). This model suggested that two-factor solution fit somewhat better than a single factor solution, but it also revealed significant correlations between the measurement errors for the items worded in a purist direction and between the measurement errors for the items worded in a pragmatic direction. 6 Over 95 percent of both parties delegates said that they were involved in the 2000 presidential campaign. 18

19 motivated by a commitment to party, to a candidate, or to an issue or some specific group. 7 [Table 3 about here] Campaign Commitments In 1972, there is a clear contrast between the highly amateur Democratic delegates, who reported a relatively strong commitment to issues and a relatively weak commitment to party, and the comparatively professional Republican delegates, who were highly committed to party and largely uncommitted to issues (see Soule and McGrath 1975; Roback 1975; Kirkpatrick 1976 for similar observations about the 1972 convention delegates). However, by 1976, those differences between Democratic and Republican delegates had virtually disappeared. From 1976 to 2000, there is a further decline the distinction between issue and party commitments. Among Republican delegates, there was a sharp increase in the importance of issue and group commitments, with some increase in candidate commitments, but no noticeable decrease in party commitments. The pattern is even more striking among Democratic delegates, where there was a similarly sharp increase in the importance of commitments to issues and groups, but also an increase in commitment to party, along with some increase in candidate commitments. By 2000, both sets of delegates were more issue-oriented than in any previous year and were more candidate-oriented than in most previous years, but were no less party oriented than they had been in the past. This disappearing distinction between campaign commitments is highlighted in the correlations among these items in Table 4. There was considerable tension between party and candidate commitments up until 1984, as the correlations between these factors were negative 7 In the CDS surveys from 1972 to 1992, issue and group commitments were combined into the same question. In the 2000 CDS, we asked separate questions about issue and group commitments. To make the results for 2000 comparable to those in earlier years, we show the percentage of delegates saying that either a lot of their activity was motivated by a commitment to an issue or a lot of their activity was motivated by a commitment to a group. 19

20 and statistically significant for both parties. At times, a similar tension between party commitment and issue and group motivations was present for Democrats, as the correlation between those factors was negative and significant for Democratic delegates in 1972 and 1984, and never strongly positive. By 2000, however, such tensions appeared to have vanished, as the correlations both between party commitment and issue and group commitments and between party commitment and candidate commitments were positive and statistically significant in both parties. In short, it appears that the parties national convention delegates increasingly have become professional ideologues, characterized by both commitments to issues and to party. [Table 4 about here] Amateur-Professional Traits and Demography Tables 5 and 6 show the results of regression analyses in which our measures of incentives, norms, and campaign commitments are the dependent variables. The independent variables are activists demography and membership in various types of groups. The demographic variables are gender, race, education, income, age, and religious affiliation and religious commitment. Our measure of religious affiliation is membership in religious traditions, 8 and we have divided the major religious traditions into high and low commitment groups. 9 The regression models include dummy variables for eight of our nine religious groups, with high-commitment mainline Protestants serving as the comparison category. 10 Group 8 We use the religious affiliation battery in the 2000 CDS patterned after the affiliation batteries in recent National Election Studies and follow the recent literature (Kellstedt and Green 1993; Kellstedt et al. 1996; Layman 2001) to identify the five largest American religious traditions (white evangelical Protestants, white mainline Protestants, Catholics, black Protestants, and Jews), and the secular, or non-religious, population. 9 Religious commitment was measures by a factor score including frequency of worship attendance, religious salience, and attitudes toward the authority of the Bible. High and low commitment groups were defined as above and below the median of the commitment sore. 10 Due to very small numbers of observations, black Protestants and Jews were dropped from the analyses for Republican delegates. 20

21 memberships are measured as the number of groups in four categories in which delegates described themselves as active members. The four categories are politically-liberal groups such as pro-choice groups, environmental groups, NOW, the NAACP, and the Interfaith Alliance; politically-conservative groups such as pro-life groups, the NRA, and the Christian Coalition; labor and teachers unions; and other relatively non-ideological groups such as fraternal organizations, service clubs, and the PTA. 11 [Tables 5 and 6 about here] The overall relationship between incentives, norms, and campaign commitments, on the one hand, and demography and group memberships, on the other hand, is relatively weak, as indicated by the general lack of explanatory power of the regression models. However, a number of the results conform to the expectations from the literature. Democratic women adhere to less pragmatic norms and have stronger issue commitments than do Democratic men. Bettereducated Republicans are motivated less by material and solidary concerns and more by commitment to issues in the 2000 campaign than are their less-well-educated counterparts. Older delegates in both parties are more pragmatic and more committed to party than are younger delegates. Committed evangelical Protestants in the GOP seem to fit the amateur profile, as they attach more importance to purposive incentives and less importance to material and solidary incentives, have less pragmatic norms, and have stronger issue commitments than do other Republican delegates. Secular and Jewish Democrats are motivated more by purposive concerns and issue commitments, and are less pragmatic than their fellow Democratic delegates. 11 There are significant numbers of Democratic delegates who are active members of politically conservative groups (104) and of Republican delegates who are active members of politically liberal groups (166) to include these variables in the analyses. However, because only 2.2 percent of Republican delegates described themselves as active members of teacher s unions or labor unions, we do not include the union memberships variable in the Republican analyses. 21

22 Membership in groups with a clear ideological agenda particularly those in with the dominant ideological sentiments of the party (liberal groups and unions in the Democratic party and conservative groups in the GOP) is associated with having stronger purposive incentives and adhering more to purist norms. However, other results run counter to the conventional wisdom. Higher education levels are not associated with stronger purposive incentives or issue commitments among Democrats, nor are younger delegates in either party motivated more than older delegates by purposive concerns or issue commitments. Unlike in past studies, membership in non-ideological civic or community groups is not negatively related to amateur political orientations. It is positively associated with material and solidary incentives in both parties and to party commitment in 2000 for Republicans. However, it has no relationship with purist-pragmatist norms or issue commitments in 2000 in either party, and is positively related to the importance of purposive incentives for Republican convention delegates. Amateur-Professional Traits and Party Support Table 7 shows the correlations between the various incentives, norms, and campaign commitments and party support, measured by seven-point scale ranging from not very strong to very strong support for the delegate s party. Some of the results conform to the expectatition that professional orientations are related to stronger party loyalty while amateur orientations are associated with weaker attachments to party. The importance of solidary incentives and party commitment in 2000 is associated with stronger party attachment in both parties, and material incentives and pragmatist norms are positively related to party support among Democratic delegates. Other results, however, run counter to expectations. Purist-pragmatist norms are not related to party attachment among Republican delegates, and purposive incentives and issue 22

23 commitments in 2000 are not related to party support in either party. Moreover, it appears that there is no longer a tension between candidate commitments and support for the party among Democratic and Republican activists, as the importance of candidate commitment and party attachments as motivations for 2000 campaign activity are positively and significantly correlated. These data also suggest the emergence of professional ideologues in both parties. [Table 7 about here] Amateur-Professional Traits and Political Participation Table 8 shows the correlations between incentives, norms, and campaign commitments and several measures of political participation, including political experience, activity, and office holding. For the Democrats, the first entry is for the super delegate slots that the Democratic Party has reserved for party and elected office-holders (whether or not delegates currently hold public or party office). For both parties, the measures of political experience are the number of past conventions to which respondents were delegates and the number of years that they have been active in politics. Two of our measures of activity record the number of volunteer and leadership activities, respectively, in which delegates were involved 2000 presidential campaign. 12 A third measure of activity is respondent s level of involvement in sub-national political campaigns in 1996 and [Table 8 about here] 12 Leadership activities include working as a planner or consultant, supervisor, or staff member. The specific volunteer activities that we asked about were voter registration, door-to-door canvassing, telephoning, distributing literature/putting up signs, polling, fund raising, mass mailing, media production/publicity, poll watching, campaign organization/working in a campaign office, rallies and campaign events, small meetings or teas/coffees, get out the vote drives, and other activities. 13 We asked respondents to state the extent of their involvement in 1996 senatorial and congressional campaigns, 1996 gubernatorial, state legislative, and local campaigns, and in both of those types of campaigns in Responses range from not at all involved (1) to very much involved (3). The level of sub-national campaign involvement is each delegate s average response to those four questions. 23

24 Some of the patterns in Table 8 support expectations that amateurs are less politically experienced, less active in party campaign activities, and less likely to hold political office than are the professionals. However, the empirical support is fairly weak, and there is other evidence that runs counter to that conventional wisdom and supports our argument about the emergence of professional ideologues in party politics. As other work has shown, political experience in the form of either past convention attendance or length of political activity is, in both parties, positively related to pragmatist norms and the importance of solidary incentives and party commitment, and negatively related to issue-oriented campaign commitments. However, the relationship between political experience and pragmatist norms is fairly weak for Democratic delegates, and in neither party is experience related to the importance of purposive concerns for general political involvement. Democratic super delegates and members of both parties who hold government office are slightly more pragmatic than other delegates. Super delegates place less importance on issue and group commitments as reasons for their 2000 presidential campaign activity. Government officials in both parties are motivated more than other delegates by material and solidary concerns. However, neither super delegates nor government office-holders are motivated less than other delegates by purposive incentives, and government officials were not less committed than other delegates to issues in the 2000 presidential campaign. Not surprisingly, delegates who hold official party positions are more likely than other delegates to attach importance to party as a campaign commitment and to solidary incentives (one of which is party attachment). However, party office-holders are neither less motivated by purposive incentives or issue commitments nor more likely to adhere to pragmatist norms than are their fellow delegates. In fact, the negative relationship between holding party office and pragmatism or positive 24