BELIEF CONFLICT IN THE COMMUNITY: LEADER AND FOLLOWER DIFFERENCES IN POLICY PREFERENCES Thesis for the Degree cf Ph. D. MICHIGAN STATE UNIVERSITY Norman R. Lufl'beg “I965 IHESIS This is to certify that the thesis entitled Belief Conflict in the Community: Leader and Follower Differences in Policy Preferences presented by Norman R . Luttbeg has been accepted towards fulfillment of the requirements for Ph.D. degree in Polificol Science Major professor Date November 24, 1965 0-169 u—t— ‘ ‘ fl. -. x14 ABSTRACT BELIEF CONFLICT IN THE COMMUNITY: LEADER AND FOLLOWER DIFFERENCES IN POLICY PREFERENCES by Norman R. Luttbeg This study focuses on two central themes. First, in the outpouring of research in the community certain hypotheses have been accepted as valid with- out receiving thorough testing. The social and economic backgrounds as well as the basis on which a leader's strength in the community decision-making process is based are both seen as vital factors regarding what pragrams would be adapted in the community if that leader were dominant. This study attempts to assess the importance of distinctions between labor and business leaders as well as between economic and political leaders concerning their policy preference differences. In addition to pattems of policy preferences among community leaders, the study also focuses on the policy preferences of different levels of both political activism and economic achievement. The second focus while related to the first is broader in its concern, since it deals with the functional basis of democracy. The importance of pOpular con- trol of leaders in democracy and the effectiveness of various mechanisms for its achievement are scrutinized . Four possible means for popular control are sug- gested, two of which are empirically evaluated in the study. The question of the sharing of policy preferences among leaders and followers and followers' Norman R. Luttbeg awareness of leaderships' biases is central to the analysis. The attitudes studied are derived from nine issues of concern to two northwest communities, Eagene and Springfield, Oregon . These issues are: attracting new industry to the community, annexation, public parking lots, special education, fluoridation, public housing, urban renewal, a metropolitan park, and the provision of public kindergartens. Distributions of attitudes on these nine issues held by various samples of leaders are compared both with each other and with the publics of the two communities. The sharing or lack of sharing termed the degree of representativeness in the study is the primary data analyzed . Finally, those persons who are most misrepresented as a result of the biases of community leadership are studied to discover their reactions to this situation. A process by which community leadership becomes unrepresentative of the policy preferences in the community is clearly evident. With the exception of the public housing issue, leadership in the community is more favorably inclined to the adoption of the pragrams than the community. This process is not the result of the backgrounds and influence bases of the leaders, rather, it is a phenomenon of which men within each of the various backgrounds become the actual leaders of the community. Labor and business leaders cannot be characterized as being liberal and conservative respectively. But both labor and business leaders who are members of the communities' leadership are more liberal than their counter- parts who are not community leaders. This bias of leadership in the community results in some persons being ”'95" a ‘0 -~ -;- #9 6443.: .rr‘hrwuoub .u («54... NM: '0 -ca .Fr 9 & 't-m- M" (IH'I'M' " A at?" “gym. . . o .1 v, L I’D-“l: .. U-‘t j. * -. "‘ it's.- a)” «We :fi .‘ - . dew—ha A: :;T:s.—-_‘—‘.‘q :a—eé .: vs rig-:4 Norman R. Luttbeg under-represented or not represented on these nine issues. As one might expect from the direction of the bias of leaders, such under-represented persons are disapproving of the adoption of the issues. Such persons show an awareness of! their being under-represented . BELIEF CONFLICT IN THE COMMUNITY: LEADER AND FOLLOWER DIFFERENCES IN POLICY PREFERENCES 3y -. ,3 r J Norman R leuttbeg A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of I DOCTOR OF PHILOSOPHY Department of Political Science I 965 .2 Q U} U (DJ, J ,. ACKNOWLEDGEMENTS I would like to thank those who suffered with me in the pains of secondary analysis and writing one's dissertation 2500 miles from one's department. N.R.L. ACKNOWLEDGEMENTS ................... TABLE OF CONTENTS LIST OF TABLES ........ LIST OF FIGURES ...................... Chapter II. THE CONCEPT OF UNREPRESENTATIVENESS. . . . ,, . . Serving the Public's Will or its Interest POpu lar Control The Effectiveness of Any Means of Leadership Coercion by Fol lowers The Community Context Leadership Vulnerability to Popular Coercion ThrOUgh Elections Conclusion and Presentation of Other Questions THE CAUSES AND THE IMPACT OF UNREPRESENTATIVE- NESS ........ The Representativeness of Leaders Variables Differentiating Between Leaders and Their Effects on Representativeness The Direction of Leadership Policy Preference Bias Subjective Awareness of Being Unrepresented Conclusion and Summary CONCEPTS AND MEASUREMENT ............ The Samples The Data Other Concepts Evaluating the Hypotheses Page vi 29 5i IV. VI. Page TWO COMMUNITIES AND NINE ISSUES ........ 7l Two Communities Attitudes on Nine Issues Involvement in the Issues Issue Involvement and Attitudes Toward the Issues Polarization with Increased Involvement The Generality of Political Involvement General Political Involvement and Attitudes Toward the Issues The Relationship Between Involvement and Social Status Social Status and Attitudes Toward the Issues Summary of the Findings and Conclusion LEADERS AND FOLLOWERS . . ............ I05 Leaders and Followers Vulnerable and Less Vulnerable Leaders Potential and Actual Leaders The Alternative Policy Preferences of Community Leadership Whom Do the Leaders Represent? 4 Bios Toward the Policy Preferences of Those With Higher Social Status Leadership Bias Toward the Policy Preferences of the Politically Active Comments on the Issues of Public Housing and the Attraction of Industry Summary of Findings and Conclusions AWARENESS ON THE PART OF THE UNREPRESENTED CITIZEN ..................... I43 The Unrepresented Citizen Awareness of Being Unrepresented The Accuracy of the Perceptions of the Unrepresented Citizens The Impact of the Realization of One's Unrepresentativeness Another Hypothesis, the Disgruntled Man Conclusions III. I?! (I ‘I [I‘ll ll . III\ I II I l VII. CONCLUSIONS. . . OOOOOOOOOOOO The Bias of Community Leadership The Distinction Between Vulnerable and Less Vulnerable Leaders The Distinction Between Actual and Potential Leaders Awareness of Being Unrepresented Leader-Follower Cleavage as a Source of Community Conflict Limitations of the Study BIBLIOGRAPHY ................... Appendix A. THE QUESTIONS .................. B. THE SAMPLE AND SAMPLE DENSITIES ......... C. IDENTIFYING LEADERS AND ASSIGNING STRATA. . . D. CONTENT SENSITIVITY. 00000000000000 I73 I79 l87 I90 I94 Table 4-4 4-I 0 LIST OF TA BLES Comparison of political economic leaders on the percentage differences between their agreement on the items and that of the publics in the two communities studied by Presthus . . Comparison of the percentage of persons in each social class. . . ............ Comparison of policy preferences in Springfield and EUgene by percentages holding each attitude . ....... Comparison of the involvement in the issues in Springfield and Eugene by the percentages of respondents giving each answer .................... Minimum and maximum percentage of either community on any issue using the various alternative responses for the issue involvement question ..... Means of the distributions of attitudes toward the issues of each of the issue involvement categories Polarization of attitudes with increasing involvement in the issue measured by percentages of the sample in the strongly approve and strongly disapprove categories . Degree of general political involvement for persons active in one or more issues, by percentages . . . Polarization of attitudes with increasing general political involvement, by percentages of the sample in the strongly approve and strongly disapprove categories . Standard deviations of the distributions of attitudes toward the issues for each category of general political involvement . ............ Comparison of Issue divisiveness types on average standard deviation and average increase in standard deviations with increasing involvement .......... vi Page 36 74 76 78 79 80 88 9O 94 4-II 4-I2 5-8 Means of the distributions of attitudes toward issues for each category of general political involvement ....... Social status composition of each general political involvement stratum, in percentages . . . Social status composition of the "Have taken an active part" response of issue involvement, by percentages . Means of the distributions of attitudes toward the issues for each level of social status ........ Summary of discovered findings and relationships . Comparison of leaders and followers in the two communities based on their means ................ Rank order of differences in the means of leaders and followers in the two communities from smallest to largest . . ................... Comparison of leaders and followers in the two communities by standard deviations of their attitude distributions ...................... Comparison of vulnerable and less vulnerable leaders in the two communities on the basis of the differences between their means and those of the communities . . Comparison of actual and potential leaders within selected leadership background types by means of their attitude distributions. Springfield and Eugene leaders are combined . . . .................. Comparison of actual and potential leaders within selected leadership background types by the differences of their means from those of the combined publics ...... Comparison of actual and potential leaders with the combined communities based on standard deviations . . . . . Comparison of selected leadership background types combined from both communities based on the means of their attitude distributions ........... vii Page 96 9'8 98 I00 I02 I06 I07 I09 III II4 II6 II7 I20 5-II 5-I2 6-5 Comparison of selected leadership background types by the rank order of their means, from most approving to least approving . . . ......... Comparison of selected leadership background types by the rank order of their differences from the means of the combined communities. . . ....... Comparison of the direction and degree of social status bias among Springfield and Eugene leaders Comparison of vulnerable and less vulnerable leaders on the direction and degree of their social status bias . ........ Comparison of actual and potential leaders in relation to their social status bias . . . Comparison of the direction and degree of leadership bias toward the preferences of the activists . . Comparison of vulnerable and less vulnerable leaders on the direction and degree of their bias toward the preferences of political activists ........ Comparison of actual and potential leaders on the direction of their political activist bias The attitudes on the issues and the number of persons identified as unrepresented in each of the issues in both communities . . . ..... The number of issues on which the unrepresented citizens are unrepresented . . Comparison of the general political involvement of the unrepresented citizens and the more represented citizens . Comparison of the social status of the unrepresented citizens and the more represented citizens ....... Comparison of how the unrepresented and the represented citizens view the activities of city officials. . . viii Page I2I 124 I27 I28 I29 I34 I35 I37 I44 I44 I45 I46 I47 6-8 Comparison of the unrepresented and the represented citizens' view of the activities of city officials, controlling for the number of issues on which the unrepresented citizen is not represented Comparison of the unrepresented and the represented citizens' view of the importance of voters in the making of key decisions on major policies in the community . Comparison of the unrepresented and the represented citizens' view of the importance of businessmen in the making of key decisions on major policies in the community Comparison of the represented and the unrepresented citizens' vew of the importance of voters in the making of key decisions on major policies in the community, controlling for the number of issues on which the unrepresented citizen is not represented . Comparison of the unrepresented and the represented citizens' view of their personal influence . . Comparison of the unrepresented and the represented citizens' sense of efficacy Comparison of the unrepresented and the represented citizens' cynicism toward local politicians, controlling for the number of issues on which the unrepresented citizen is not represented Number and percentage of sample falling into each category of content sensitivity Comparison of yea-sayers' and nay-sayers' responses to the nine issues with those of the entire public. Page I48 I49 I49 I5I I53 154 I55 I96 Figure I-I LIST OF FIGURES Means by which leaders can serve the preferences of the followers ..... 000000000000 00000000 Miller and Stokes' paradigm. . . .......... Continuum of vulnerability . . . . . . . ...... OOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOO Overlapping between actual leaders and the various leadership background types. . . .......... The possible relationship between involvement in an issue and issue polarization ............. 3 II I2 I3 26 3l 3] 34 57 86 CHAPTER I THE CONCEPT OF UNREPRESENTATIVENESS This study is concerned with persons of differing degrees of involve- ment and influence in the political process. As our understanding of political behavior grows more extensive, it becomes increasingly apparent that society is stratified politically. We know that people vary in their degree of concern about the outcome of elections, in their amount of information about the political process and the actors in it, in the extensiveness of their own partic- ipation, and in their commitment to political beliefs. More importantly, we know that persons who rank high on any one of these attributes also tend to rank high on the others.1 There appears to be a general variable of involve- ment in the political process. A given individual's location on this variable, his political stratum, is indicated by his possession of the various character- istics of political involvement, such as how frequently he votes and how informed he is on the issues. The top political stratum includes those who rank high on each of the characteristics of political involvement. They are sufficiently involved to lSee especially, Robert A. Dahl, Modern Political Analysis, (Engle- wood Cliffs, N.J., I963), p. 57. Also see, Angus Campbell, Philip E. Converse, Warren E . Miller, and Donald E . Stakes, The American Voter (New York, I960), ch. vii; Philip E. Converse, "Information FTow and the Stability of Partisan Attitudes," Public Opinion Quarterl , XXVI (Winter, I962), p. 58l; RobertA. Dahl, Who Govems? (New Haven, l96l), p. 90; and V .0. Key, Jr. , Public Opinion and American Democragy (New York, l96l), ch . viii. I I‘J spend large amounts of their time in political activity. They care a great deal about who wins a given election, perhaps because they are candidates, friends of candidates, or campaign workers . They care greatly whether or not a given policy is adopted by the government; they may, for example, have proposed it or have a self interest in the outcome of the argument over a policy. By their extensive participation in community decision-making, these men are most likely to make the actual choice of future government policy from among various alternative policies. At the local level of government and, probably, at higher levels, this stratum includes men other than those who hold elected or appointed offices. At the opposite extreme are those persons who seldom, if ever, vote, possess little or no information about the political process, and view govem- ment as a factor of little importance in their lives.2 They seem far more con- cerned with what is immediate and personal in their lives rather than with the abstractions of political issues.3 One of the more important findings of the electoral survey studies is that there are a large number of persons within this stratum. Individuals can be ordered along this variable of political involve- ment between the end points of great involvement and little or no involvement. 2Daniel Goldrich, "On the Concept of Politicization," (unpublished paper prepared for the Research Seminar on "Processes of Community Decision- Making and Change and Their Influence on Education," University of Oregon, August, I963). 3Edward Shils, "Primordial, Personal, Sacred and Civil Ties," British Journal of Sociology (June, I959), p. I30. C3 Thus for, these political strata have been defined entirely on the basis of involvement in the political process, but our society can also be stratified in terms of the influence persons have in affecting what is done in their community. A rough correspondence between involvement and partici- pation in the political process and personal influence in that process would seem likely, and has been argued by two authors.4 As yet, this relationship has not been substantiated, and will not be until we develop the concept of influence, understand its dynamics, and develop adequate indicators. The relationship between influence and involvement seems most evident at the extremes of involvement. The opinions of those who are in a position to choose from among the policy alternatives under consideration seem to weigh heavily in that selection . Thus, regardless of how the community feels on an issue, these men can, for at least a short time, choose the policy alternative they themselves prefer. On the other hand, the apolitical have little or no influence on that selection. I will call the top political stratum, leaders. It would be more mnemonic to call them leaders in the community political process, but this is awkward. Also, I hesitate to call them "political leaders" because this implies that they are professional politicians, which many of them are not. The number of possible political strata into which a theorist could divide society is greatly variable and depends on his analytic purpose and the l"Robert E. Agger, Daniel Goldrich, and Bert E. Swanson, The Rulers and the Ruled (New York, I964), p. 705; Key, Public Opinion, p. l84. llllllrlll.‘ Il'[([[l[.{\[l‘[[rt{r‘(([{[lull/Illl . .rl sophistication of his research design . For most of the analysis presented in this paper, I will define two political strata--leaders and non-leaders, non- leaders being persons in political strata other than the top political stratum. Depending on the prevalent term for non-leaders in the literature being dis- cussed or in my analysis, I will use the following terms synonomously: "public," "community, citizens,‘ and "followers." In addition, I will use the term "constituency" when the type of leader under discussion has a defined subset of this public which he is to represent and from which he is elected . In classifying the active, well-informed members of the society with the apolitical members of the society, I may be obscuring many pertinent relationships. At this early stage in the exploration of patterns of policy preferences in society and for clarity in presentation of data, however, limiting my strata to two seems desirable to me. Also, in discussing the primary rela- tionship I will be investigating, theorists have seldom distinguished between subsets of the public; they might argue that no such distinction should be made for normative reasons. I am certain this reasoning will become more apparent later. Later, some variations in the public, theoretically important for the relationship of interest, will be analyzed . People differ greatly in their opinions. One person may approve of public housing but be strongly opposed to fluoridation; another may strongly approve of both. An issue will have X number of peOple approving and Y number disapproving of its content. Furthermore, within these approving and disapproving subgroups, some members will feel more intensely about their opinions than others. Thus, one can speak of the distribution of opinions for a group on a given issue and patterns of distributions of issues within a society. My central concern is with the relationship between political strati- fication, as described above, and the patterns of distribution of policy-related opinions or attitudes within the society. Do different political strata hold greatly dissimilar opinions on the desirability of various policies under con- sideration by government? For example, do thelmembers of a particular political stratum as a group more strongly approve of urban renewal than does the public? To the degree that there is a relationship between political stratification and patterns of distribution of policy preferences in a community and that leaders rely on their personal attitudes in the act of making decisions, the policies of that community are more likely to be those of the more polit- ically involved . Serving the Public's Will or Its Interest In making a decision a leader may act on several sets of information . He may act entirely on the basis of his personal opinions as to the desirability of the various alternatives. Or he may choose one of the alternatives on the basis of its long-term desirability. If the decision requires action by others, the leader must consider the likely actions of these others. They may be cc- decision makers, persons greatly concerned with the issue, or even large aggregates of persons, such as the entire public in the case of the decision maker in a political system. The leader may weigh the opinions of others because he believes that these persons may have important effects on the decision, or he may merely believe he should weigh their opinions. Many factors affect the decision made by a leader. Apart from explaining the behavior of the leader in making decisions, this behavior may be evaluated by the use of various standards. Burke con- ceived of two alternative normative standards by which the behavior of a representative could be evaluated . His insistence that a representative should serve his constituency's interest and not necessarily its will has long been a subject of discussion in relation to representation in democracy.5 In its more recent usage, Burke's idea of the representative serving the public interest has increasingly come to mean that the representative should act on his personal, better-informed opinions. In attempting to better explain behavior, political scientists have con- ceived of variables other than the institutional and legal variables which were formerly used to indicate the distinctiveness of the representative's position from others in the government. The representative is no longer conceived of as the sole member of government to give consideration to the interest or the will of the public. Other men's opinions also affect the policy of government and its ultimate meaning to society.6 Thus, this same argument is applicable to all leaders in the political process in a democracy, no matter what their institutional and legal roles. Both of these norms of proper behavior by the 5The Works of the Right Honorable Edmund Burke, II (Boston, l826), p. I0, cited from Warren E. Miller and Donald E. Stokes, Representation in the American Congress (forthcoming), ch. i. 6See the literature on the dissatisfaction with the distinction between administrative and policy positions. Herbert A. Simon, Donald W. Smithburg, and Victor A. Thompson, Public Administration (New York, I949). representative seem to exist in the behavior of the American leader in the political process . 1/. While this normative argument has endured for nearly 200 years, only recently have we turned to the factual underpinnings of the argument. Are the leaders capable of representing the will or preferences of their followers? In part, this long wait was necessary until methodoIOgy developed to allow exploration of the descriptive questions . My study will attempt to answer this question . I hasten to point out that the investigation of the capability of leaders to represent the will of the public need not reflect adherence to this normative position. A proponent of leaders serving the public's will may be appalled at the existence of unrepresentativeness and stirred to further efforts on the behalf of his normative position. Or an opponent of such behavior by leaders may be gratified to find leaders taking stands contrary to the preferences of the public; but nevertheless, the degree of representativeness, exists. Similarly, from a discovery that leadership is not able to serve the public's will, one can draw two apposite conclusions. First, given the stabil- ity and present desirability of the United States as an example of democracy, the normative requirement is unnecessary for, or perhaps even detrimental to, the functioning of democracy. Second, one could conclude that democracy in the United States is challenged by the failure of leadership to satisfy this normative standard which would then call for corrective action .' Because I 7John C. Wahlke, Heinz Eulau, William Buchanan, and LeRoy C. Ferguson, The Legislative System (New York, I962), p. 28] . cannot demonstrate the superiority of one of these conclusions over the other, no normative conclusions will be attempted as a result of the empirical findings. A word of caution at this point: Ever present are the dangers of reifying concepts such as "the public's will" and "public opinoin'I by giving them content other than the distribution of the preferences of the public . In this paper the "public's will" is merely the distribution of the public's policy preferences on various issues, and, being a distribution of preferences, it is unlikely to be monolithic. However, the degree of consensus will be one of the research interests in this paper. I persist in the use of this phrase, "the public's will," because it is short and, I believe, meaningful. The distinction between serving the public's will and serving its interests also appears to be meaningful to leaders in our society as shown by Wahlke et al . in their study of state representatives. Since the selection from among various altemative policies on a given issue is ultimately reduced to being for or against one or more alternatives presented as a bill or bills before government, the public's will is served only iflthe decision on these bills is consistent with the attitudes of the majority of the public. Thus, if one were dealing with the voting performance of leaders on bills offered before government, whether or not they are serving the public's will would be judged with reSpect to the majority policy preferences of the public. No such majority rule assumption or value is needed when working with the sharing of policy preferences between leaders and followers, as in this study. I will be concemed with leadership as a policy preference microcosm of the public, and not with the voting behavior of the leaders and its consistency with the policy preferences of the majority of the public. Popular Control One relationship between leaders and followers, within the context of the leader serving the followers, is that of popular control--how the fol- lowers can control the policy actions of the leader. The concept of popular control has two components. The first is the standard of judgment by which the leader's performance is evaluated. As i noted above, I am focusing on the public's will or the policy preferences of the followers. This then is the standard of judgment used in the following analysis. The second component of papular control is the means of coercion available to followers for use against leaders who fail to meet the standard of judgment. Most commonly, the device of elections is conceived of as the means of coercion. But other means, such as pressure from interest groups and party discipline, can also be employed . I will return to this point later. Those who argue that the leader should serve the public's interest also concede that a leader may be unfaithful to those he serves, that is, he may fail to satisfy a standard of judgment. The component of coercion by the people to assure performance is, however, an anathema of this latter position. if the public's interest is known only to the leader, how can the public remove him for having failed in his attempt? Of course, other leaders may not share a given leader's conception of what is in the public's interest. Within the leadership stratum settlement of this disagreement is a question of interaction between leaders, a point beyond the concern of this study. And should this lO disagreement be carried to the public, that public could only reSpond in terms of its personal preferences as to which interpretation of its interests is more reasonable and preferable. This evaluation by the public seems to me to be indistinguishable from the expression of its will. The leader's freedom from popular control means that he may follow his beliefs as to what is in the public's interest. The figure below shows four alternative means of satisfying the public's will. Only two of these means posit the need for a functioning means of coercion to be made available to the public. I have avoided calling these Leader acts consistently with his personal preferences Leader acts on what he believes to be the pref- erences of those he leads Coercion of some sort used to assure performance No means of coercion necessary to assure performance Men whose preferences are preferred by the followers are made leaders Because leaders and followers share many experiences and pref- erences, leaders in voting their own pref- erences also vote the preferences of the followers. Leaders vote the pref- erences of their followers for fear of being removed from leadership. Leaders vote what they believe to be the pref-4 erences of their fol- lowers and even anti- cipate their prefer- ences because the leaders believe they should do so . Figure 1-] Means by which leaders can serve the preferences of the followers means by which leaders can serve the preferences of the followers, models of representation. With the exception of Hobbesian usage of the term, all models ll of representation are concemed with the question of followers controlling leaders. Two distinct dynamics are included in such models. In the schema offered above, each of the rows SUggeSl' possible explanations as to how the preferences of the followers are, or can be included within, the highest political stratum. The schema deals only with the leaders as individuals and, more particularly, with their preferences or perception of follower preferences relative to the actual preferences of the followers. Given the distribution of preferences, perception of followers' pref- erences, or other means by which the public's will can be communicated to the leadership stratum, the second dynamic, the interaction between members of this stratum, becomes of concem in the resolution of policy actions or outputs. This second dynamic is beyond the interest of this study, but this second dynamic is greatly dependent on the first. For example, if the distribution of preferences within the top political stratum relative to that of the populace on a given issue were that shown in Figure 1-2, numerous inter- action dynamics among the members of the top political stratum can be sug- gested by which the policy actions taken by this stratum could be consistent with the preferences of the followers. Public Percentage of the group holding a Leaders given opinion - Attitude Dimension + Figure l-2 lg! Elections may place the subgroup of leaders most accurately reflecting the preferences of the followers in an institutional position which facilitates the articulation and enforcement of their policy decisions. Or because of the belief that the policy decisions should be supported by the followers, one of the subgroups may alter its preferences to make them more like those of the public and appeal to the public for support. Figure 1-3 shows another distribution of preferences within the top political stratum which precludes the operation of these dynamics. It is hard to conceive of any interaction between the members of a stratum of leaders holding these preferences which could result in policies consistent with the preferences of the public. The public's will would need to be served either by the leader acting on other sources of information about the public's will, or by the leader serving it by accident. This distribution differs from that in Figure 1-2 in that the central tendency of the leadership stratum differs greatly from that of the public, the dispersion of preferences within the stratum is not sufficient to overlap or include all preferences within the public, and it is unimodal, thus not likely to lead to competition among the members of the leadership stratum for public support. Numerous distributions falling between these two distributions are possible varying on these three characteristics of Public Percentage of the group holding a / Leaders given opinion "- Attitude Dimension + Figure l-3 distributions. The point is, however, that we know very little about the nature of these distributions. The interplay of leadersI policy preferences and the concept of popular control are dependent on the availability of several alternatives on a given issue. If situational, technical, or bureaucratic limitations define only a single action for the decision-makers, clearly, neither their preferences nor the preferences of their followers play a part in the selection of government policy. It may be that government policies are increasingly so defined, especially at the local level of government.8 But even a cursory survey of the decisions before local govemment shows that many policies are not so defined. In their study for the House of Representatives, Miller and Stokes have offered a paradigm quite similar to that shown in Figure 1-] . They are con- cerned with the paths by which the Representative's constituency controls him. Their paradigm is shown in Figure 1-4. Representative's Attitude Constituency's Attitude \ Representative '5 / Roll Call Behavior Representative's Perception of Constituency's Attitude Figure l-4 Miller and Stokes' Paradigm 8Robert E. Agger, et al. comment on the 'pluralistsI dependence on this limitation on community decision-making. Agger, Rulers, p. 76. 9Warren E. Miller and Donald E. Stokes, "Constituency Influence in Congress," The American Political Science Review, LVll (March, I963), p. 50. 14 They conceive of two major pcths by which the constituency can control the policy actions of their Representative. "The first of these is for the district to choose a Representative who so shares its views that in following his own convictions he does his constituents' will . . .The second means of constituency control is for the Congressman to follow his perceptions of district attitude in order to win re-election."lO These paths compare to those listed in my schema in the first column (see Figure l-l). Although Miller and Stakes, in this article, deliberately exclude from their consideration means of constituent control other than those involving some form of coercion, the use of the path of control thrOUgh the Representative's attitudes could or could not be the result of coercion.H Also, finding extensive use of the path through the Represent- ative's perception of his district's attitudes might be explained by his belief that he should strive to vote according to his constituency's attitudes. Miller and Stokes anticipate that different paths will explain the Representative's behavior within different issue areas. Indeed, they find that for issues of domestic welfare, the path thrOUgh the Representative's attitudes explains the greater share of the Representative's roll-call behavior.12 The policies with which I will be working are entirely domestic policies. I will, therefore, be concerned only with the upper row of means of popular control, leaders acting on their personal preferences. My focus of inquiry becomes: lO|_l_3_igl., p. 50. ”Enid” p. 50n. 1th1?” p. 53. 15 Can leaders, on the basis of their personal policy preferences, act consistently with the policy preferences of th :ir followers? The Effectiveness of any Means of Leadership Coercion by Followers Descriptive theorists have spent a great deal of time and conceptual effort seeking a means by which there can be public coercion of leaders. The first models of representation pitted the public, as individuals with certain institutional powers such as elections, against leaders. If the individual citizen needs to have extensive information about his own personal preferences, the various candidates' policy positions, and their voting records in order to perform this function of coercion, it is evident that he cannot do so.'3 With the growing realization of the average citizen's lack of involve- ment in politics, several authors offer secondary organizations which they believe assist the citizen in controlling leaders. Noting the predominance of groups in the United States, both in the process of government and as a characteristic of American society, several theorists suggest that individuals of common beliefs join together to proselytize those beliefs. '4 The implica- tion is also made that personal preferences derive from the groups to which an ”Numerous studies could once again be cited, but see especially Bernard R. Berelson, Paul F. Lazarsfeld, and William N. McPhee, Voting (Chicago, I954), ch. xiv. 14David B. Truman, The Governmental Process (New York, I960). I6 individual belongs.'5 Given the representation of all or most preferences among existing groups, the competition among these groups for control of government is seen as serving the public's will.16 Empirical research, however, has found that group membership is characteristic of a minority of the public .'7 The inclusion in these groups of all preferences within the public is therefore questionable, and without total inclusion, the effectiveness of this means of affecting leaders' actions becomes a means of biasing that action contrary to the public's will. Secondly, granting that group membership is a minority phenomenon, Presthus finds the activity of groups qua groups in the environ- ment of the community to be quite limited .'8 At the state and national levels of government, however, organizations are both active and somewhat successful.19 I conclude that, like other means of control, competition among groups is an effective means of leadership control but is used only by a minority of activists and thus, may bias the policy actions of leaders toward the I55ee Stanley Rothman's criticism of this implication that groups are both the source of and the result ofattitudes. Stanley Rothman, "Systematic Political Theory: Observations on the GroupApproach, " The American Political Science Review, LIV (March, I960). 16Truman qualifies this on page 51 but, nevertheless, sees group competition as vital to the performance of govemment. See page 502. l7Gabriel A. Almond and Sidney Verba, The Civic Culture (Princeton, I963), p. 320. '8Robert Presthus, Men at the Top (New York, I964), p. 28I . 19An extensive literature exists on the subject of group activity and success at these levels. V.O. Key, Jr., Politics, Parties, and Pressure Groups (New York, I958); LesterW. Milbrath, The Washirflon Lobbyists (Clficago, I963); and Harmon Zeigler, Interest Groups in American SocieDI (Englewood Cliffs, N.J., I964). l7 preferences of this minority. Empirical work has not yet been done on the policy-preference bias inherent in the various means of leadership control by activist minorities. It has also been squested that the political party may control leaders. In their pursuit of electoral victory, the parties serve the public's will by seeking out the policy preferences of a majority of citizens and, by one of two proposed means, inducing their elected members to enact these preferences into policy.20 The two general means by which a political party induces its elected members to enact its policies are party control and selective recruitment. Party control means that the party has available to it means of coercion by which it can force its elected members into compliance with desired policy positions. Theorists, believing selective recruitment important, argue that, by means of various organizational dynamics within the party, only men holding certain preferences rise in the party hierarchy. Thus, the party's elected members'are men who hold policy preferences close to those of the party leadership, making coercion unnecessary, inasmuch as the elected members favor and enact the policies of the leadership.22 Regardless of the ability of parties to enact their 20\/.O. Key, Jr., American State Politics (New York, I956). Downs' argument requires that the political parties be able to define their issue positions so as to optimize their appeal to the greatest number of voters. Thus, some means of membership control to achieve ideoloQical purity is necessary. Anthony Downs, An Economic Theory of DemocraCJ (New York, I957), ch. viii. 2'E.E.Schattschneider, Party Government (New York, I942). 22Samuel J. Eldersveld, Political Parties: A Behavioral Analysis (Chicago, I964), pp. 20I-204. I8 policy positions, this means of popular coercion fails to be effective for more than a minority of the public because people do not reSpond to the party on the basis of its policy positions.23 Also, the stability of party voting in the United States and the evidence that less than one-half of the public is aware of which party controls Congress (despite the fact that a guess would be correct 50 per cent of the time), denies the effectiveness of this means of coercion of leaders by followers .24 Faced with this apparent absence of means of public coercion, yet unwilling to conclude that leader's beliefs alone can explain the present functioning of democracy or be relied on to preserve it in the future, many theorists have reduced their requirements for popular coercion .25 They argue that control by coercion need not be so specific as to apply to performance by the leader on every policy issue, but may be an overall evaluation.26 Also, 23Campbell , ch . x . 2A'Philip E . Converse, "New Dimensions of Meaning for Cross-section Sample Surveys in Politics," International Social Science Journal, XVI (No. I, I964), p. 2i; Fred I. Greenstein, The American Party System and the American Peop_le (Englewood Cliffs, N.J., I963), p. 32; and Warren E. Miller and Donald E. Stokes, "Party Government and the Saliency of Congress, " Public Opinion Quarterly, XXVI (Winter, I962). 25Robert Dahl argues the need for more than dependence on leadership indoctrination. Robert A. Dahl and Charles E. Lindblom, Politics, Economics, and Welfare (New York, I953), p. 290. 26Dahl, Who Govems?, p. 305; H .8. Mayo, An Introduction to Democratic Theory (New York, I960), p. 77. I9 they argue, the expression of this ability to control need not be used so frequently as to occur every tirre the elected leader comes up for election .27 Noting that an activist minority does seem capable of utilizing the various means of coercion, several theorists have also argued that these activists, being themselves broadly representative of the general public in their policy pref- erences, do coerce leadership into compliance with their preferences and thus, into consistency with the public's will .28 The trend in the writings of democratic theorists seems to be to reduce the dependence on public coercion of leaders. More recently, several theorists, agreeing that the public is capable of only a vague and general supervision of leaders, have looked more benignly on the relative autonomy of the leader in democracy.29 This autonomy allows the leader to consider the long-run needs of society, the needs of a more inclusive public than that of his constituents alone, and the practice of diplo- macy in international affairs. Burke's desire for representatives to serve the public's interest is satisfied by this argument, and Lippman's criticism of the fickleness of public opinion seems unwarranted in the excesses he attributes to 27Key, Public Opinion, p. 553; E.E. Schattschneider, The Semi- Sovereigir Peeple (New York, I960), ch. viii. 28Berelson, p. IIO; Dahl, Politics, p. 313; and Schattschneider, Semi- Sovereign, ch. viii. Schattschneider puts faith in activists offering meaningful alternatives to the public in the socialization of conflict to include those who are normally apolitical. 29Almond, p. 476; Key, Public Ojinion, p. 555. 20 it.'30 The leader need not be guided only by the whims of the public. Not all research findings demonstrate the lack of effectiveness of coercion by the followers. In the study cited earlier, Miller and Stokes interview the candidates who failed to win election against the incumbents. They find that the incumbents act for more consistently in accord with the preferences of their constituencies than their opponents seem capable of doing .3] By means of the electoral process, the constituency, on the average, appears capable of choosing the candidate whose preferences are most consistent with its own. Costantini also finds evidence of the effectiveness of public coercion by means of the electoral process. He finds that the party functions to exclude from party leadership positions those persons who hold immoderate opinions. Thus, persons in the higher levels of party leadership hold policy preferences 32 more in line with those of the public. The threat of the electoral process, whether effective or not, causes the parties to select for top leadership men whose preferences are most consistent with those of the public. And they, in 33 tum, choose candidates even closer to the preferences of the public . Using a similar research design in his study of party ideolon in 3(hlValter Lippmann, The Public Philosophy (Boston, l955). 3lMiller and Stokes, p. 50. 32Edmond Costantini, "lntraparty Attitude Conflict: Democratic Party Leadership in California," The Western Political Quarterly, XVI (December, 1963), p. 972. 33519., p. 973. 2i ’3 Detroit, Eldersveld does not find the relationship noted by Costantini .04 In fact, he finds competition causes the parties to offer ideologically immoderate persons as candidates, men whose preferences differ greatly from those of the districts in which they are running.35 Research into the public's demonstrated ability to utilize available means of selecting leaders who most exactly reflect their policy preferences is thus inconclusive. The dependent variable in all but Chapter VI of this study is the sharing of policy preferences between different political strata. A major independent variable is the availability of means of coercion to one stratum, the followers, to force another, the leaders, to share their policy preference by excluding those who do not. Without the availability of such means, do leaders hold policy preferences not shared by the followers? Leadership Representativeness l have been using the term "policy preferences" to denote the opinions or attitudes of individuals towards government issues. The distribution of these preferences in the society or community is the content of the public's will concerning policy alternatives before government. These policy preferences are but a small part of the individual's beliefs relevant to the process of govern- ment. Other relevant beliefs are that the individual believes the laser in an electoral contest should accept the will of the majority and that he believes his opinions have an impact on government. These beliefs are also important 3‘lEldersveld, p. 193. 351339., p. 203. l\) k) to the functioning of government. The concept of the public's will has its greatest relevance in evaluating the policy output of a govemment and the consistency of that policy output with the preferences of the peOple. There- fore, my analysis will be limited to the policy preference subs.:t of the indi- vidual's total belief system having a direct or indirect impact on government. I have Spoken frequently of the sharing of policy preferences between political strata . Thus, I am speaking of distributions of policy preferences within strata and the comparability of these distributions between strata. All measures which describe the congruence between distributions, then, are indicators of the sharing of policy preferences between strata . More will be said of such indicators in Chapter III . l have spoken of groups of leaders as "sharing the policy preferences" of their followers. To avoid use of this lengthy phrase, leaders will be said to be representative of their followers if they share the follower's policy pref- erences. Using this term also has the desired implication of the leaders responding to the preferences of the followers rather than vice versa . The Community Context Community studies have shown that there are leaders in the community political process other than those formally elected or appointed. The presidents or managers of large corporations, main street businessmen, labor union leaders, and neWSpaper publishers are often found among the community's leadership in the political process. What most clearly distinguishes community leadership from the remaining members of the community is that community leaders, as a 23 group, are more affluent and better educated than the public.36 One of the most extensive literatures in any field of political science has been developed concerning the implication of such findings. One group, composed primarily of sociologists, concludes that the community's leadership does not serve the public's will, but rather, serves the interests of the wealthy, with political leaders playing a subservient role in the selection of policy alternatives.37 Dahl, as the exponent of the other group, concludes from his study of a New England community that few men exercise influence in more than one area of policy, and that, if any leader is found to do so, he is probably the mayor of the community, a leader subject to coercion by means of elections.38 The underlying assumption of this literature seems to be that elected leaders serve the public better inasmuch as they are more subject to public coercion to assure that they serve the public's will . With the uncertainty of the effectiveness of coercion in popular control and with the growing conceptual dependence on means of satisfying the public's will without using public coercion, this assumption should be tested . Are leaders who are not subject to popular coercion less representative of the public than elected leaders? Unless they are, is there any purpose in evaluating 36Numerous studies can be cited beginning with the studies by the Lynds' of Middletown and Floyd Hunter, Community Power Structure (Chapel Hill, N.C., l953). More recent studies have found the same phenomena. See: Agger, Rulers; Presthus; and Dahl, Who Governs?. 37'Hunter, ch. vii. 38Dahl, Who Govems?, p. 183. 24 whether or not economic leaders dominate the political process in a corn- munity? Much of the difficulty arises from the oversimplified conclusion that persons from the same social background share the same attitudes. The correctness of this conclusion is dependent on the distinctiveness of the dif- fering social backgrounds. In a divided society one might expect the relation- ship to be strong, at least on the issues that divide the society. The United States, however, is not a greatly divided society. I do not deny the agglutina- tion of certain attitudes within groups sharing similar social backgrounds, but to use social background data as equivalent to attitude differences between groups seems to be too indirect and to risk unnecessary error. While numerous researchers have shown that elected leaders are educationally and financially "better off" as a group than is the public, numerous researchers have found party leaders and elected leaders to be representative of the public in the sense that they shared the same range of opinions, although they were more extreme in their opinions.39 What is more, leaders were more extreme in their opinions not only in the direction one might assume to be the preference of the wealthy, i.e. , against Medicare, etc . , but also in the opposite direction, that is, contrary to the supposed preferences of 40 the wealthy. Measures of social background are a very imperfect predictor 39Agger,‘ Rulers, p. 335; Eldersveld, p. 52; and Presthus, p. l83. 40Eldersveld, p. 193; Herbert McClosky, Paul J. Hoffman and Rose- mary O'Hara, "Issue Conflict and Consensus Among Party Leaders and Fol- lowers," The American Political Science Review, LIV (June, l960), p. 422. 25 of policy preferences .4' By categorizing community leaders according to the degree of their subjection to popular coercion through the electoral process, I hope to test the importance of coercion in assuring popular control. Leadership Vulnerability to PoBular Coercion Through Elections I noted earlier that persons neither elected nor appointed to public office are often found among community leaders. The various methods of identifying persons of great influence in the community indicate that many elected and appointed officials are not among the community's leaders. Men seem to gain influence in the community by achievement in varied pursuits. In the case of many of these non-elected and non-appointed leaders, the public cannot assert any coercion if the leader fails to serve the public's will in the community's decision-making . Perhaps if the leader's occupation is the production of some commodity, the public can seek to coerce him by refusing to purchase this commodity from him; but for the most part, such men are beyond coercion by the public. Other leaders, such as party leaders, are less vulner- able to coercion than the elected leader but, over a period of time, can be reached by coercion on the part of the public. Thus, the man who gains influence through his ability to control a party's selection of candidates and 4'Rossi makes a different but supporting argument: ". . .it is open to question whether for many issues there are clear and consistent differences among class groups, ethnic groups, age levels, and so on, which could manifest themselves in different decisions dependent on what kind of decision-maker holds office." Peter H . Rossi, "Community Decision-Making," Administrative Science Quarterlj, l (March, 1957), p. 422. 26 the office holders of that party, may be "reached" by the public merely by defeating that party's candidates. It is possible to conceive of a variable of leadership vulnerability to coercion by the public. Figure l-5 shows a roth positioning of different leaders on this variable. The measurement of this variable is very crude; thus, the placement of different leaders along this continuum is somewhat arbitrary. For the purposes of the following analysis, leaders are dichotomized into more vulnerable and less vulnerable classes on the basis of their subjection to popular elections. Economic Labor Party Elected Candidates Leaders Leaders Leaders Officials for Office lnvulnerable Vulnerable Figure l-5 Continuum of Vulnerability One of the means by which one can test the importance of leaders' subjection to coercion to assure their acting consistently with the public's will, is to identify leaders in the political process who are and are not subject to such coercion and measure which is most capable of serving the public's will . Vulnerable leaders are subject to such coercion, and invulnerable leaders are less subject to such coercion. Are vulnerable leaders more repre- sentative of the public than less vulnerable leaders? Conclusion and Presentation of Other Questions Is the leadership stratum in the American community a policy-preference microcosm of the community? I assume the answer to this question to be "somewhat. " My analysis, then, is an attempt to examine the effects of 27 several variables on the policy-preference representativeness of leadership. My primary concern is with those characteristics of leadership which affect its representativeness. Are certain leaders more representative than others? In the first chapter, I outlined one major variable, often stressed by political scientists, which is supposed to affect the representativeness of leaders--being subject to coercion by the public, primarily by the use of elections. Since various theories of politics and democracy rely heavily on this variable, it is of great importance in my analysis. Nonetheless, 1 hope to do more in this paper. I wish to explore the distribution throughout the com- munity of different policy preferences. Other important independent variables affecting representativeness, as well as a variable dependent on represent- ativeness will be offered in the next chapter. The second variable thotht by some scholars to affect leadership is that of recruiting available potential leaders into the stratum. Is a potential leader more likely to become a leader if he holds certain policy preferences? Does this selective recruitment improve the representativeness of leadership? What are the policy-preference boundaries of available leadership in the community? Could changes in the leadership recruitment process improve the representativeness of community leaders? If leaders are biased in their policy preferences, what is the direction of that bias? I have argued that society is stratified according to individuals' involvement in the political process. For each of the above relationships, society has been dichotomized into only two strata, the leaders and the fol- lowers. But are the relationships of political involvement and patterns of policy preferences continuous between strata? Are leaders more representative of the more active members of the public? Another bias attributed to community leadership is that, being more affluent and more educated themselves, they must give expression to the policy preferences of this segment of the community. Are the policy preferences of community leadership more representative of the better educated and more affluent? Finally, if a bias is discovered in the policy preferences of the leader- ship, it is possible to identify persons whose policy preferences are unexpressed in the leadership stratum on any given issue. Is there a group of individuals who might be labeled the "unrepresented citizens," persons who have few or no members of the leadership stratum expressing their preferences on several issues? And to what extent do such persons manifest awareness of their state? Furthermore, what is the nature of their awareness? In this exploratory study the selection of the above variables is based on generalizing to community leaders, previous findings on specific subsets of leaders at various levels of government. The next chapter will attempt to formalize these relationships into hypotheses and to show their origin in the existing literature . CHAPTER ll THE CAUSES AND THE IMPACT OF UNREPRESENTATIVENESS Stating explicit and formal hypotheses has the advantage of focusing one's research endeavors, but doing so also has disadvantages. First, there is the implication that a sufficient understanding of the human behavior being examined is at hand to yield alternative theories of that behavior, and that the more valid theory can be identified by acceptance or rejection of a single, crucial hypothesis. Hypotheses #l and #2 have long been assumed to be correct, thus their rejection would demand extensive revisions of the theoreti- cal offerings of many men, but no alternative theory is intended to be thereby substantiated. But correct knowledge of the fundamental relationship considered in these hypotheses would permit the develoPment of less assailable theories. With these and other hypotheses, I hope to gather information vital to the subsequent development of a theory or theories of leader-follower linkage in democracy. The second disadvantage of offering formal hypotheses is related to the first. The advantage of focusing research endeavor can be carried too far and obscure unanticipated relationships. Sensitivity to such relationships and serendipity are vital at this early stage in our devel0pment of political theory. Being aware of these disadvantages, I will offer in the form of 29 30 hypotheses several variables which I feel affect the representativeness of leaders and the direction and result of any misrepresentation discovered. Where possible, these hypotheses are derived from extant literature and theory. Althoth a hypothesis necessitates some statement of the direction of the relationship between variables, my primary concern will be with the discovery of the direction of any existing relationship and the strength of that relation- ship. Thus, in some of the hypotheses the direction of the relationship, as stated , is arbitrary . The Representativeness of Leaders The underlying hypothesis leading to this research is that leaders in the decision-making process are not an attitudinal microcosm of the society for which they make decisions: Hypothesis #l: Leaders are not representative of the community. Since complete consensus on attitudes is improbable among leaders or fol- lowers, measuring the leaders' representativeness depends on a comparison of the distributions of their attitudes toward given issues with the distributions of these same attitudes among the followers. Thus, in addition to comparing the most characteristic attitude or the central tendency of the two groups, it is necessary to compare another measure of the distribution-the consensus or degree of homageneity of each group. A leadership group whose characteristic opinion disagrees greatly with the public's characteristic opinion of an issue is even more unrepresentative of that public if the leaders are consensual in their opinion. On the other hand, 3] if these leaders lack consensus in their opinion or policy preference, with a minority of the leaders sharing the position of the public, there would seem to be a greater likelihood that the leaders would enact a policy consistent with the public's will. Figures 2-l and 2-2 schematically present this argument. Both the leadership groups in Figure 2-l and Figure 2-2 are unrepresentative, Percentage of ' the group holding the opinion ‘ - Attitude Dimension + Figure 2-l Percentage of the group holding a given opinion ‘ Attitude Dimension + Figure 2-2 as judged by their characteristic opinion shown by line A in comparison with the characteristic opinion of the public shown by line B. But the leadership group shown in Figure 2-2 is less consensual, and though still unrepresentative, it is more capable of representing the public than the group in Figure 2-l in that more of the spectrum of public opinion is reflected within the leadership stratum. The effectiveness of such unpopular opinions among leaders would be strengthened by their public popularity. Comparing the representativeness of 32 two or more groups necessitates a consideration of both measures of distribution comparison . Although unimodal curves such as those shown in Figures 2-l and 2-2 are common, it is possible that one or both of the distributions of the leaders and the followers on a given issue may be bimodal or polymodal. Thus, this additional distribution characteristic needs to be noted in order to properly evaluate the measures of central tendency and homageneity. Variables Differentiating between Leaders and Their Effects of Representativeness Vulnerable and Less Vulnerable Leaders. Many theorists have argued that the distinction between elected political officials and non-elected leaders, who are usually called economic or business leaders, is a most important distinction which affects their ability to serve the public in a democracy . In a form of government where men stress public participation and are desirous of serving the public's preferences, devices utilizing this participation, such as elections, are developed to assure public control of the leaders. This distinction between persons subject and not subject to such control, and theories dependent on this distinction were discussed extensively in the first chapter. A variable of vulnerability to public control was conceived to distinguish between leaders. This variable was said to be continuous. Thus, certain political leaders, such as candidates for public office, are more vulnerable than other political leaders because public coercion in response to their unsatisfactory performance is immediately available. Among the non-elected leaders, some are more vulnerable than others. For example, a labor union leader who is accountable to a membership which may remove him from office or not follow his leadership may be more vulnerable than an owner of a business who is not so accountable to a membership. ls it true that vulnerable leaders more accurately reflect the policy preferences of the public? The hypothesis of greatest concern in this study is: Hypothesis #2: Vulnerable leaders are more representative than less vulnerable leaders. Little research dealing directly with this hypothesis has been completed, but vulnerable leaders have been the subiect of a fairly extensive research literature. Herbert McClosky et al . were the first to utilize the attitude questionnaire to compare samples of leaders and followers, and their research design has been employed extensively in more recent investigations of party leaders and the public.] The method used in these studies is essentially that used in my study. A universe of certain institutionally identified party leaders and a random sample of citizens are asked a battery of policy preference questions.2 McClosky et al . were most concerned with whether the American political parties gave a public of varied policy preferences little ideoloQical choice. On the contrary, they found that parties offered a consensual public lHerbert McClosky, Paul J. Hoffman and Rosemary O'Hara, "Issue Conflict and Consensus Among Party Leaders and Followers," The American Political Science Review, LIV (June, 1960), pp. 406-427. Others have used his design, see: SamuETJ. Eldersveld, Political Parties: A Behavioral Analysis (Chicago, 1964), pp. 183-219, and Edmond Costantini, "lntraparty Attitude Conflict," The Western Political Quarterly, XVl (Dec., 1963), pp. 956-972. 2McC|osky, p. 407. 34 a greater choice than the public seemed to desire. If anything, the individual whose opinions were relatively middle-of-the-road might not find any party expounding his policy preferences. Furthermore, they found the delegates no more consensual in their policy preferences than the public.:3 Figure 2-3 schematically presents their findings. Using my terminology, Republican Pu bl ic D emocratic Percentage of Leaders Leaders group holding \ a given opinion “ Attitude Dimension + Figure 2-3 the characteristic positions taken by the leaders in each party are not repre- sentative of the public, but the absence of consensus on these positions means that most of the public have some leaders in each party who expound their positions or preferences. Furthermore, combining leaders in both parties into a sample of vulnerable leaders yields a policy preference distribution which is even more representative of the public, althoth still less consensual than the public's. Later research has focused on the effects of party hierarchy in moder- ating the extremeness of party leaders. Are presons holding higher offices in the party more moderate in their opinions and thus, more representative of the public than convention delegates? Though this development of their research 313319., p. 424. 35 will be referred to again in the literature relevant to later hypotheses, what is important at this point is the relative representativeness of vulnerable leaders. Vulnerable leaders are more divided in their policy preferences than is the public, but their preferences broadly cover the spectrum of opinions held by the public . Other researchers also deal with values and policy preferences held by leaders and followers. Miller and Stokes are interested in the "paths" by which a Congressman's district can control his voting performance.4 In evaluating the path of control through the Representative's perSonal preferences, they find that Congressmen are greatly unrepresentative of their districts. Miller and Stokes control for the Representative's district, thus Congressmen in general would be representative only if they each were representative of their respective districts--a more rigorous definition than that used in my study . Although controlling for the Representative's district is vital in evaluating the performance of representative institutions and of given representatives, only the policy preference representativeness of the whole leadership group is important in serving the public's will. None of this literature deals with the community, nor is there any effort made to compare the representativeness of these vulnerable leaders with less vulnerable leaders. Presthus is the only researcher who has attempted such a comparison using attitudinal data.5 By means of decisional analysis, he 4Warren E. Miller and Donald E. Stokes, "Constituency influence in Congress,'I The American Political Science Review, LVll (March, 1963), pp. 45—46. 5Robert Presthus, Men at the T93 (New York, 1964). 36 identifies political and economic leaders in two New York communities. He uses a "conservatism" scale in evaluating both these leadership groups and random samples of the publics.6 The items do not scale, but one can compare . , . . 7 groups on an Item by Item bCiSlS. Table 2-1 shows the percentage of agreement with each of the items by each of the leadership types in the two communities, and the discrepancy Table 2-1: Comparison of political and economic leaders on the percentage difference between their agreement on the items and that of the publics in the two communities studied by Presthus. Edgewood Riverview Political Economic Political Economic Leaders Leaders Leaders Leaders l. That government which 78 (+54)* 57 (+33) 40 (+23) 53 (+36) least governs best. 2. We have moved too far 33 (-7) 57 (+17) 50 (+18) 53 (+21) away from those funda- mental principles that made America great. 3. One of the biggest 55 (+14) 57 (+16) 60 (+22) 63 (+25) problems with the world is that peOple don't work hard enough any more. 4. Democracy depends 89 (+21) 93 (+25) 80 (+19) 84 (+23) fundamentally on the existence of free enter- prise. 5. On the whole, labor 44 (+12) 50 (+18) 80 (+29) 47 (-4) unions are doing a lot of good in this country. AVERAGE DlSCREPANCY 21.6% 21.8% 22 .2% 21 .8% *The percentage dif-ferences shown in parenfiretie's— are the percentage of leaders agreeing less the percentage of the public agreeing. 61133., p. 326. 71939., p. 323n. 37 between this percentage and that of the public on this item is shown in parentheses.8 Although the publics and the political leaders in the two communities differ greatly in their agreement to the items, the economic leaders are in remarkable agreement. But if one averages the absolute values of the discrepancies between the percentage of agreement on the five items of the two groups of leaders with the publics in the two communities, one finds that the political leaders are no more representative of the publics than are the economic leaders. Generalizing from these data, if the attitudes on this "conservatism" scale are related to, or precondition, the attitudes of leaders toward specific policies that are before their communities currently or may be in the future, the economic leaders could serve the public's will as capably as could the political leaders! Presthus' data contradict Hypothesis #2. Another possible attitudinal comparison between individuals is suggested by two groups of researchers who have concentrated on the concept of ideOIOgy. 9 Although Presthus' "conservatism" scale may be considered a measure of ideOIOgy, the work of these two groups of researchers has been to discover new dimensions of ideOIOgy in the community and not to impose the common liberal-conservative dimension. A person's ideOIOgy is more basic to his belief system than his specific policy preferences, and it preconditions and 83339., p. 326. 9Robert E. Agger, Daniel Goldrich, and Bert E. Swanson, The Rulers and the Ruled (New York, 1964); and Oliver P. Williams and Charles R. Adrian, Four Cities (Philadelphia, 1963). constricts his responses to present and future policies.10 The preconditioning attitudes are not entirely derived from the person's economic interest, and the authors argue that, indeed, ideOIOgy is independent of the concept of interests. ll But the distinction can only be made when the two concepts are in opposition, such as when a small businessman who opposes the expansion of government opposes the construction of a parking ramp near his business establishment even though it would bring him more customers. His ideOIOgy overrides his interest. Community leaders, including vulnerable and invulnerable leaders, have been shown to hold differing ideoloQies. Agger et al . have found that the ideOIOgy of "community conservationism" is growing more dominant in the four communities they studied .12 But Williams and Adrian have found three different ideoloQies dominating three of the four communities they studied, and 13 in the other community they found a conflict of ideoloQies. Though the con- cept seems to have great potential, neither group has successfully operation- alized it, nor substantiated their conclusions in a systematic, empirical manner. M This difficulty has also precluded the study of whether the leaders' 10Agger, p. 16. “reigns 16. 1212c}, p. 648. ”Williams, p. 160. 14See Philip E. Converse and William A. Scott for vital concepts and an empirical solution to this difficulty. Philip E . Converse, "The Nature of Belief Systems in Mass Publics," in IdeOIOgy and Discontent ed. by David E. Apter, (New York, 1964), and William A . Scott, "Empirical Assessments of Values and ldeoloQies, " American SociOIOgical Review, XXlV (June, 1959), pp. 299-310. 39 ideoloQies are shared by their followers. A discovery that two persons share a policy preference gives no informa- tion as to whether they hold this attitude because they have similar ideOIOgies or interests. The more numerous their shared policy preferences, however, the more likely it is that they share similar ideolOgies or interests. The concept of serving the public's will does not require the sharing of ideoloQies. It is difficult to reach any conclusions from this varied research . Presthus found economic leaders responded in a remarkably similar manner to the items on his "conservatism" scale. In contrast, political leaders and the public varied greatly. This would seem to indicate that economic leaders hold attitudes of great similarity, regardless of the attitudes of the community in which they live. But the researchers who have studied the ideolOgies of community leaders have found ideOIOgical conflict between leaders, including both political and economic leaders. The literature developing from McClosky's study shows vulnerable leaders to be unrepresentative only in the extremeness of their policy preferences. But Miller and Stokes find Representatives to Congress to be unrepresentative of their districts; and Presthus finds political leaders no more representative than less vulnerable, economic leaders, thOUgh differing widely in their attitudes. in large part, this confusion is the result of trying to compare findings using policy preferences, attitudes towards conservatism, and ideOIOgies as indicators of the representativeness of different types of vulnerable and less vulnerable leaders at different levels of government . At the same time, this difficulty of reaching a more general conclusion demonstrates the need to deal more systematically with the question of 4O representativeness. The research necessary to evaluate Hypothesis #2 is my attempt to begin a systematic evaluation of this question. Biased Selection of Leaders. Although a citizen can usually move from one political stratum to another merely by varying the amount of effort and time he expends in political participation, entrance into the top political stratum is not voluntary. Influence, more than any other of the complex of variables indicating one's political stratum, seems to be most important in identifying this stratum. Being elected to public office or being the president of a large corporation does not guarantee that one will be a member of this stratum. Not all men who wish to have the final say on an issue are in a position to. One might expect to find many more men who possess all of the attributes of leaders yet are not leaders because they lack influence. 1 will speak of these less influential members of the highest political stratum as "potential leaders." The possibility exists that the actual leaders in a community are a biased selection from among these potential leaders. A biased selection of leaders may result from three possible processes. First, a selective recruitment process may exist through which only potential leaders who conform to the policy preferences of the present leaders are awarded influence or admitted into the leadership stratum. Second, great influence may be associated with certain achievements in society, and persons most likely to achieve in society are also most likely to hold atypical policy preferences. An aample of this process is the type of individual who is likely to become the top executive in a large, important corporation and might, therefore, be expected to be unfavorable to government regulation . 41 The third process is the socialization of leaders after they have become leaders. No matter how men may have become leaders, once they are leaders, they may be enticed into conformity or, thrOUgh the realization of problems previously unperceived, they may change their policy preferences to the point that they conform with the older leaders. Whatever the means of achieving this bias, is there a bias in the selection of leaders from among those with leadership potential ? Hypothesis #3: Leaders are less representative than potential leaders. The literature on political party structure and policy preferences developing from McClosky's article is relevant to this hypothesis. Eldersveld conceives of a dynamic process within the parties serving to maintain them as "viable ideological entities."15 By the processes of attracting persons who adhere to the party's ideoloQical positions and rewarding higher offices to those who conform to these positions, he argues that the party maintains its ideolOgical purity. Supporting evidence for this conclusion is found in the more extreme ideological positions of persons higher in the party hierarchy and of persons who persist in their activity in the party.16 Political parties have a biased leadership-selection process which produces a leadership stratum that is more extreme in its policy preferences than the potential party leadership. In his study of the California Democratic party, Costantini finds the ”Eldersveld, p. 201. “$19., p. 193 and p.215. 42 same selective process; but its purpose is opposite to that found by Eldersveld. l7 He finds a selective process operating to make high party leaders more representative of the public in that they are less extreme in their policy preferences than lower level leaders in the party.‘8 A biasing process of some sort, seemingly selective recruitment, is operative in the selection of higher officials in the party hierarchy. But it is uncertain whether this process improves the representativeness of leaders. ls such a process characteristic of entrance into the top political stratum of the community, and if so, does it improve the representativeness of leaders? ln their discussion of conformity, each of these authors referred to the conformity by different groups of leaders at different levels in the hierarchy of the party to the ideOIOgical position assumed to be associated with that party. Thus, a Democrat is more of a conformist if his policy preferences are those of the Democratic party, that is, he conforms to a "liberal" position. With the exception of McClosky et al . , the researchers make no effort to evaluate the internal homageneity of these groups. One good indication of conformity is a high degree of consensus within a group. Notably, McClosky et al . find ‘no greater consensus among the delegates to the national conventions of the parties than they find among the followers of the parties within the random sample of 19 the public. As l have stressed before, each of my hypotheses on ”Costantini, pp. 966 and 967. 1813319., p. 971 . l9McClosky, p. 424. 43 representativeness entails an evaluation of central tendency as well as the spread of the distribution. Thus, conformity, as iudged by consensus, will be important to the evaluation of Hypothesis #3. By combining the distributions of the policy preferences of the leaders and the potential leaders, one can speak of a distribution of the policy preferences of available leadership in the community and of persons who define the extremes of that distribution. The popular wisdom is that labor union leaders and Democrats are the "liberals" in policy preferences and that the business and economic leaders and Republicans are the "conservatives" in our society. More recently, the distinction between local government and higher levels of government has been added because local labor often takes "con- servative" positions on local issues. The association between new governmental services and tax increases is more visible at the local level; therefore, the anti -tax-increase attitudes of the less educated, lower incomed, and often unionized individuals force labor to resist extension of costly services by local government. Do labor and business leaders define the extremes of the distri- bution of policy preferences of leaders and potential leaders? And similarly, do Democrats and Republicans within available community leadership define the extremes of the distribution of policy preferences? Hypothesis #4: Labor union leaders and economic leaders define the extremes of leadership policy preferences. Hypothesis #5: Democratic and Republican party leaders define the extremes of leadership policy preferences. Eldersveld found that Republicans did take policy positions which were 44 conservative, that is, against medical aid by government, foreign aid, and government involvement in civil rights; while Democrats took the liberal . . 20 . . . . positions. But are economrc leaders more conservative than the Repuolrcan party leaders? And are labor union leaders more liberal than the Democratic party leaders? Very little research has been completed which places different types of leaders within the context of attitudes towards policies at the local level. The Direction of Leadership Policy Preference Bias More Representative of the "Better Off. " Thus far, the standard with which the distribution of policy preferences of the different leadership groups has been compared, is the public's distributions of policy preferences, the public's will. But if leaders are less than truly representative of the public, then whom do they represent? Judging from the literature of community studies, one would expect their policy preferences to be more like those of the better educated and more affluent members of our society. These members of our society will be called "better off" in the following discussion. Certainly community leaders do tend to be heavily drawn from these better off segments of our society.” But Presthus found that in Riverview, his poorer, lower class city, the political leaders more close ly approximate the average income and education of the public, but these leaders held attitudes no closer to those 20Eldersveld, p. 188. 2lPresthus, p. 183. of the public than any of the other leadership groups (see Table 2-1).22 Riverview did not benefit from the fact that its political leaders more accurately reflected the social background of the community. The inadequacies of such social background variables as indicators of policy preferences have been discussed at length in Chapter I, but to what extent does the selection of leaders from among the better off bias the policy preferences of the leadership stratum towards this minority? Though the path to becoming a community leader through public election may be more frequently closed to the poorer members of our society because of increasing campaign costs and time restrictions, this path is more open to such persons than is achieving influence in the community through success in a business hierarchy. Because of the possibility of recruiting vulnerable leaders from other than the more affluent and better educated seg- ment of society, are such leaders less representative of the better off than less vulnerable leaders? Similarly, does the biasing processes of the selection of actual leaders from among available leaders further orient the policy preferences of leaders toward those of the better off? Hypothesis #6: Leaders are more representative of the better off members of the community. Hypothesis #7: Less vulnerable leaders are more representative of the better off members of the community than are vulnerable leaders. Hypothesis #8: Leaders are more representative of the better off members of the community than are potential leaders. 2211351., p. 326. 46 More Representative of the More Politic<..lly Active. My hypotheSes of the relationship between stratification by involvement in the community political process and patterns of policy preferences has been limited to three strata--the two very small leadership strata of leaders and potential leaders and the vast maiority of the community, the public . As I have stated earlier, combining the active and the apolitical members of the public may obscure important differences in the public and give no insight into the continuity of discovered relationships. The potential of error in assuming linear relationships of variables with political involvement has been demonstrated by Converse and Costantini .23 Are the previously hypothesized relationships continuous? if so, leaders would be more representative of the more politically active fol- lowers, and potential leaders would be more representative of the politically active than actual leaders. Also, as seeking public office might well be an indication of the greatest degree of involvement in the political process, one would expect vulnerable leaders to be more representative of the more politically active than the less vulnerable leaders who achieve influence by other means. Hypothesis #9: Leaders are more representative of the more politically active members of the community. Hypothesis #10: Vulnerable leaders are more representative of the politically active members of the community than are the less vulnerable leaders. 23Philip E. Converse, "Information Flow and Stability of Partisan Attitudes," Public Opinion Quarterly, XXVI (Winter, 1962), p. 589, and Costantini and Eldersvefirs discovery of the need to differentiate between higher and lower levels of leadership. 47 Hypothesis #1 I: Leaders are less representative of the more politically active members of the community than are the potential leaders. SubEctive Awareness of Being Unrepresented Numerous authors have offered similar hypotheses to explain why people vote against school-bond issues, fluoridation, metropolitan reorganiza- tion, and vote for the Communist party. The political acts of such persons are said to derive from their perceptions of their government as a body which does not act in their interest and from their belief that they are powerless to control it. They perceive the leaders of the government as a conspirational clique, united by some value or values not desired by the public. This clique may be perceived of as an aristocracy, a big-city machine, or an influential elite in a community.24 The vote of these alienated voters is, therefore, said to be a protest against what they perceive to be true in their government. Researchers are divided in their belief in the validity of the alienated voters' perceptions. The theory dovetails neatly with the findings of elite power structure literature. The perceptions, therefore, are seen as correct in that the power structure is an elite acting on its own higher social status beliefs.25 Most of the researchers, however, seem to view the perceptions 2“*Arnold Simmel, "A Signpost for Research on Fluoridation Conflicts: The Concept of Relative Deprivation," The Journal of Social Issues, XVII, No . 4 (I961), p. 34; Hadley CantriI, The Politics of Despair (New York, I958), p. 22]; Murray B. Levin, The Alienated Voter (New York, I960), p. 62; Wayne E. Thompson and John E. Horton, "Political Alienation as a Force in Political Action," Social Forces, XXXVIII (March, I960), p. 195. 25Levin sUQgests this, p. 58. 48 of these persons as the irrational product of anomia in modern society.26 But is there objective reality in the perceptions of the alienated voter; are his perceptions of the local power structure essentially correct? Is the local power structure unrepresentative of his desires for government action, and does this lead to his alienation? Agger et al. report a growing consensus among the leaders of the community.27 If this is true, is there a comparable developing consensus among the public; or are major segments of the community left without vocal leadership for their policy prefemces? Various dimensions of the syndrome of the alienated voter have been . . . . 28 . . offered including distrust of pOllI’lCOl leaders, 0 low sense of political efficacy,29 and a general view that one is powerless to influence the holders 30 . . . . . of power. The unrepresented Citizen IS a member of the community wrth few or no leaders expressing his policy preference on a given issue. Such a person, having no leader to whom he can offer his support, would be expected 2(()Edward L. McGill and Jeanne Clare Ridley, "Status, Anomia, Political Alienation and Political Participation," The American Journal of SocioIOgy, LXVIII (September, I962), p. 206. 27Agger, p. 648. 28Thompson and Horton, p. I90. 29McGiII and Ridley, p. 206. 30Levin, p. 62. 49 to be very negative in his evaluation of community leadership if he were aware that he is unrepresented . One measure of an individual's evaluation of community leaders is his cynicism towards local politicians, a variable offered by Agger et al .3] To what degree is the distrust of political leaders by the alienated voter an indication of his awareness that he is unrepresented by community leadership? Given the fact that leadership is closed to his policy preferences, the unrepresented citizen might evaluate his influence in the community as low if he were aware of his unrepresented state. Thus, asking the individual to evaluate his influence in the community could aid in evaluating whether or not the citizen is aware that he is unrepresented. To what degree is the sense of powerlessness of the alienated voter an awareness that his policy preferences are not expounded nor considered by more influential men? Finally, if the unrepresented citizen were aware of his state, he might realize the futility of his political act and lose his sense of efficacy.32 Is the inefficacy of the alienated voter the result of his awareness that community leadership disregards his policy perferences. Two important dimensions of politics in the community, the degree of unrepresentativeness and the awareness of being unrepresented, have been combined in a dynamic which is a possible source of community conflict. The 3lRobert E. Agger, Marshall N. Goldstein, and Stanley A. Pearl, "Political Cynicism; Measurement and Meaning," Journal of Politics, XXIII (August, I961), pp. 477-506. 32Angus Campbell, Gerald Gurin, and Warren E. Miller, The Voter Decides (Evanston, Illinois, I954), pp. 187-199. 50 systemic impact of these variables cannot be assessed in this study because neither community shows either great unrepresentativeness or great conflict; but it is most probable thatin any community unrepresented citizens can be identified and the correctness of the dynamic evaluated. Conclusion and Summary The evaluation of these hypotheses should give insight into the place- ment of community leaders within the context of the attitudes of the community towards present policy issues. Leaders, like all persons, have attitudes. It seems likely and has been partially substantiated that, given an opportunity to have a say in the methods of resolving issues, they act, in part, on their own policy preferences. The questions of whose and what policy preferences are articulated by leaders then assumes great significance. Two maior distinctions between leaders have been SUgg-ested: being successful in achieving or not achieving influence and being more or less vulnerable to public ire or pleasure as a result of one's actions in the community political process. Three standards of evaluating whose policy preferences are articulated have also been offered: the public's policy preferences, the policy preferences of the better off, and the policy preferences of the community members more deeply involved in its affairs. Finally, the variable of representativeness has been utilized as an independent variable to assess the accuracy of the alienated voter's perceptions of local government as not being reSponsive to his attitudes and values. CIIAPTER lII CONCEPTS AND MEASUREMENT Numerous concepts and hypotheses have been offered in the previous chapters. In this chapter my concern will be to show how these concepts can be measured and how the hypotheses can be evaluated. The Samples The Followers. Three distinct samples of individuals are used in this study, and numerous distinctions can be made within these samples. The largest sample consisting of I,226 individuals. is a probability sample of the two com- munities, Springfield and EUgene, Oregon, and their surrounding suburbs.1 This sample is the followers, citizens, and other non-leader groups referred to in the text of the following chapters. The usual definition of a community in community studies is the geo- graphical confines of the incorporated municipality under study. This definition is generally satisfactory for the small cities under study. The discussion of the definition of community and the problem of defining its boundaries becomes necessary and more difficult when larger cities with suburbs and neighborhoods are to be studied . This is especially true in the large metr0politan areas.2 The lSee Appendix B for a discussion of the selection of individuals for this sample. 2John C. Bollens, Bploring the Metropolitan Community (Berkeley, Calif., I96I), chapters xi, xii, xiii. SI 52 definition in terms of legal boundaries is used in this study because the close proximity of the two communities under study makes it difficult to decide which suburb is part of which community. In many ways the two cities, as well as their suburbs, are so much alike that I am tempted to call the entire urban area one community. The two cities are, however, institutionally distinct, and some problems are more critical to one city than to the other. Thus, I will treat them as distinct communities, making community synonomous with city. In dealing with a number of the hypotheses, it is impossible to tell to which of the two cities a certain category of leaders belongs. Whenever this is the case, the samples of the two cities and their surrounding suburbs are pooled to form a single community. The Leaders of the Community. If I were to interview the entire pop- ulations of both cities, I would be able to identify the politically most involved group of individuals; and assuming a strong relationship between involvement and influence, I would be able to identify leaders in the cities‘ political processes. Given the uncertainty of the involvement-influence assumption, I might apply an additional criterion such as success in getting their preferred policies adopted or demonstrated involvement in the actual process of deter- mining government policy. It is unnecessary to interview so many non-leaders to find out the policy preferences of the community, however, and the cost of such over-interviewing would be prohibitive. Then the problem is how to oversample leaders in the community political process. Three methods of 3See Appendix B for a discussion of the combining of these various samples of followers caused by their varying sample densities. 53 oversampling leadership have been used by researchers. First, one can con- centrate on the process of selecting policies by government. Leaders should identify themselves by their involvement in this process. This method has been called the decisional method of identifying leaders.4 It is time consuming and expensive. For that reason researchers using it focus their attention on selected issues, rather than on the entire span of decision-making in the community. The problem, of course, is which decisions to focus on.5 The second problem of this method is its inability to distinguish between ministerial representatives of behind-the-scenes leaders and actual decision-makers.6 The second method is to identify leaders as the men holding certain offices, such as a mayor, a councilman, the president of a chamber of commerce, or a local political party official. Obviously, this method may or may not identify decision-makers, depending on whether there is an overlap between such institutional leaders and men who have a major impact on decisions. The third method is the so-called "reputational method" which is closely associated 7 with Floyd Hunter. The assumption underlying this method is that persons involved in civic organization in the community will discover, over a period 4Robert A. Dahl, Who Govems? (New Haven, Conn., I96I), pp. 332- 337. 5Peter Bachrach and Morton S. Baratz, "Two Faces of Power," '_T_h_e American Political Science Review, LVI, No. 4 (December, I962), p. 948. 6Robert Presthus, Men at the Top (New York, I964), p. 422. 7Floyd Hunter, CommunitLPower Structure (Chapel Hill, N.C., I953). 54 of time, who is most able to get their preferences adopted by the community government and will be willing to say who such persons are. One of the greatest virtues of this method is that it is cheap and simple. it also obviates the need to focus on a limited subset of the policy issues in the community. But this method has its disadvantages.8 The major criticisms of the reputational method are that it assumes a unitary power structure and finds it, and that it mixes social status with influence and, therefore, fails to identify persons of influence who have low social status, such as labor leaders. Ultimately, the inferiority or superiority of a method must be demonstrated by empirically comparing the results of two alternative methods and deciding which best identifies community leadership, as Agger et al . and Presthus have done. AlthOUgh he admits that he began his research with the belief that he would find the reputational method inadequate, Presthus found that the two methods he used, decisional and reputational, complemented each other.9 The reputational method helped to exclude from the ranks of leaders those who played only a ministerial role in expressing the preferences of another man who did not overtly participate. Agger et al. found that investigating the involve- ment of the reputational leaders in actual policy decisions caused them to 8Bachrach and Baratz, p. 947; and Presthus, p. 60. 9Presthus , p . 424 . 55 exclude only five of the l38 leaders.10 Also, in one community where their cross-sectional random sample included a large percentage of the community, the found that three of the four leaders they identified by the men's involve- ment in the political process, as judged by their responses in the survey questionnaire, were also identified by the reputational method.H AlthOUgh further research into the subiect is necessary and supplementary evaluations may add to the method, much of the criticism of the reputational method would seem unwarranted and non-empirical. The reputational method is used in this study to identify leaders in the two communities under study (see Appendix C for the procedures used). Potential Leaders in the Communfly. Another group of leaders identified for use in this study are the potential leaders. These men have bases of influence or backgrounds which have often been found among leaders in communities but lack the reputation of having influence. Six different leadership influence bases or backgrounds are found among the potential leaders. Economic dominants are the top two local officials in corporations employing 500 or more persons located in either of the two cities or their surrounding suburbs. Labor leaders are the business manager or the executive secretary or both for each of the local labor unions. Democratic and Republican campaign committeemen are persons from either community who belong to these local party committees . lORobert E. Agger, Daniel Goldrich, and Bert E. Swanson, The Rulers and the Ruled (New York, T964), p. 33] . “fluid” p. 717. 56 Candidates for office are any local residents in either of the cities or their suburbs who were candidates for public office in the fall elections of 1958, just before this survey. The offices to which the candidates aspired were county, state, and local. Finally, political contributors are local individuals who contributed $500 or more to one of the political parties. In my analysis, I will speak of these six samples as selected leadership background types. It is in the hypotheses dealing with these samples that it is impossible to maintain the distinction between the two communities. For example, if an economic dominant's corporation is located in Springfield and he lives in Eugene, as is frequently the case, for which community is he a potential leader? One of the two men who are leaders in both communities is an economic dominant living in Eugene and working in Springfield, thus demon- strating that a potential leader may become a leader in either or both com- munities. The some difficulty in assigning communities to the economic dominants exists for other potential leaders and leads to the combining of the cities for each of these hypotheses. As I have shown in the example above, actual leaders identified by the reputational method may also be identified in the leadership background samples. In such cases the leader is an actual leader and not a potential leader, thus he is excluded from the potential leaders sample. This overlapping between leadership background samples and actual leadership is vital to the evaluation of potential leader-actual leader differences. As the potential leader sample cannot be viewed as typical of the universe of potential leaders, the evaluation must be made by background types. The overlapping of actual leaders in such samples permits measurement of differences in each background type between those who are and those who are not actual leaders . Overlappinyf Leadership. Figure 3-l shows the overlapping among actual leaders on each of the samples of leadership background types. ln Springfield actual leaders were found on these background samples only six times. But in Eugene there are 44 cases where actual leaders were also found on these samples. The number of overlaps is not the same as the number of Eugene Leaders 9 4\ 9 Economic ‘__/5 / \ 4 \ Political Dominants / / , Contributors Democratic 2 , Republican Campaign ‘ 1 Campaign Committeemen Committeemen 2 Labor / Candidates Leaders 1 / 'for Office Springfield Leaders Figure 33-] Overlapping between actual leaders and the various leadership background types. members of each of these actual leadership groups which are also identified in the background type samples. For example, a Mr. Strus, now one of the 58 elected officials in the state of Oregon, was at the time of this study not only a member of the actual leadership in both communities but also a candidate for public office, a member of the Democratic campaign committee, and a political contributor. Thus, this one individual accounts for three of the overlaps for both the Springfield and the Eugene leaderships. In comparison with the actual leadership overlapping shown in Figure 3-l , only three of the potential leaders identified in these samples was in more than one of the samples. Two of them were both economic dominants and political contributors. The notable lack of overlapping of Springfield leaders as shown in Figure 3-l probably reflects the class characteristics of Springfield which will be discussed in Chapter IV. In turn, this causes very small n's in some of the tables dealing with these background characteristics when distinguishing between leaders and potential leaders. Available Leadership. One last distinction between leaders must be noted. This is the concept of available leadership. In a certain sense, all individuals in the community are available as possible leaders. We know from previous research in the community, however, that men of certain occupational backgrounds and certain bases of influence are more likely to be found among a community's leadership. Men who have these backgrounds and influence bases are the primary source of leadership available to a communitynthey are the available leadership. This empirical finding has already been used in identifying potential leaders, but an additional criterion was used in that case: 59 potential leaders were only those members of the selected background types who were [1th also identified as leaders in the community. No such additional criterion is used in defining available leadership. Both potential leaders and leaders in the community who have these selected backgrounds are identified as available leadership. In addition to the six selected background types used in identifying potential leaders, those men in any leadership sample who were found to hold public office were identified as. an occupationally vulnerable subcategory of available leadership. Hypotheses #4 and #5 concern patterns of policy preferences among available leadership. The Data The Choice of Issues. I am most concerned with the effect of the leaders' biases on their policy decisions as compared to the decisions that seem likely if the entire community were to make the decisions on policy. For this reason issues on which policy decisions seemed imminent were chosen for the analysis. The process of choosing these issues was simple but only roughly quantified. On the basis of the researcher's impression of the mass media coverage and other means of sensing the community's concern with a community issue, derived from living in the community, the issues of concern were identified first for one community and then for the other. Finally, the two sets of issues were combined, and attitudes on the combined set were asked . There was a great deal of overlapping between the issues of concern in Springfield and in Eugene. I will show in Chapter IV that on involvement in the issues 60 there are greater variations between issues than between communities. Once the issues were identified, the various samples were asked about their attitudes toward the issues, among other items of information gathered on a 24-page interview schedule. See Appendix A for the questions used in this study. Respondents were asked if they approved or disapproved of the issues and how strongly they felt either way. "Uncertain" and "Don't care" alter- natives were also given. In calculating the means and standard deviations of the resulting distributions of attitudes toward the issues, a score of zero (0) was assigned to "Strongly approve" responses, one (I) to "Approve ," two (2) to "Uncertain," three. (3) to "Disapprove," and four (4) to "Strongly disapprove . " "Don't care" responses were not included in the calculation of either statistic. The Level of the Data . The use of the above statistics necessitates the assumption that the attitudinal data on the issues are interval level data. This is the strongest assumption made about the level of the data used in this study. It is, of course, not uncommon for researchers to use Likert-response type data such as this as interval data, but I believe the assumption should be noted. 12 The data on such a scale are certainly better than ordinal level data, thoth perhaps not truly interval level data . The intervals between "Strongly approve" and "Approve" and between "Strongly disapprove " and "Disapprove" are ”Herbert McClosky, Paul J . Hoffman and Rosemary O'Hara, "Issue Conflict and Consensus Among Party Leaders and Followers ," The American Political Science Review, LIV (June, I960), p. 409; Warren E. Miller and Donald E. Stokes, "Constituency Influence in Congress," The American Political Science Review, LVII (March, I963), p. 49; William J . Goode and Paul K. Hatt, Methods in Social Research (New York, I952), pp. 273-275. 6] identical, as are the intervals between "Approve" and "Uncertain" and between "Disapprove" and "Uncertain." Thus, the level of the resulting scale could be called symmetrically equal interval. I have used the data only for comparisons of various leaders and followers on each issue and have abstained from averaging differences across issues because of sensitivity to the possible weakness of assuming the level of the data to be interval . To go to the other extreme and assume that the data is only ordinal would require the use of the median as a measure of central tendency and the range as a measure of spread . Both measures are much less sensitive than the interval level counterparts and would debilitate my analysis. The Meaningfulness of Attitudinal Responses. Two criticisms have been raised against using the data of opinion surveys as if they were synonomous with the public opinion of the universe from which the sample was selected . The first is a normative question arguing that the opinion of an individual must be an informed one before it can be included as public opinion to which leaders should respond . There is little doubt that many people will offer opinions on an opinion survey item while in possession of little or no information on that item. 13 I do not wish to get involved in this normative question of whether uninformed opinions should be given equal weight as informed Opinions. Sut to the degree that knowledge of the political process and the actors and issues in it is one of the variables associated with involvement in the political process, I" 0See Robert E. Lane and David O. Sears, Public Opinion (Englewood Cliffs, N .J ., I964), chapter vi, for a survey of the findings on His subject and its implications to democracy. 62 I will analyze whether the leaders selectively represent the more involved and probably better infomred members of the community. The second criticism is of great importance to survey analysis, that is, are the opinions expressed in a survey questionnaire meaningless because they are largely chosen at random from available alternatives by an uninterested respondent whose opinions are expressed to conform to the interviewer's values? It has been common for authors of survey questions to provide a "Don't know," "Uncertain," or "Don't care" response . Such responses allow the unopinionated to avoid expressing a meaningless opinion . But Philip Converse's work seems to indicate that this is not enough. '4 He specifically invited people [rat to express their opinions with the statements "Of course, different things are important to different people, so we don't expect everyone to have an opinion about all of these (issues). . . If you don't have an opinion, just tell me that." Nevertheless, when he analyzed these data for consistency across a three-time panel, he found in a sample issue that only I9 per cent had remained stable in . their opinions. Although he concedes that there can be meaningful changes in opinion, for the most part, he argues that the stability of an opinion is an indicator of the meaningfulness of the opinion . '5 Thus, he concludes that only a very small minority have meaningful opinions on this issue. He offers a model composed of a small group of strongly opinionated persons and a vast pool of 4Philip E. Converse, "The Nature of Belief Systems in Mass Publics," in Ideolggy and Discontent, ed. by David E. Apter (New York, I964), pp. 206-26I . 151%., p. 24I . 63 persons whose opinions are meaningless. '6 Though I might argue that requiring consistency over a four-year period may be a severely rigorous criterion for meaningfulness of opinion, his findings have great relevance to those who wish to make use of the Opinion survey. Another study attempting the same evaluation found opinions on specific concepts such as domestic issues were less stable than opinions on abstract concepts such as opinions on politicians. '7 Numerous studies, however, show consistencies in opinions which would not be expected in Converse's model was wholly correct. Key found that making succeedingly more rigorous requirements of necessary information for inclusion in the distribution of attitudes on a given issue had little effect on the distribution of opinions on that issue.18 When comparing activists and non-activists in both the Republican and Democratic parties, Eldersveld found that the non-activist members of each party differed from each other and, as groups, remained more stable in their preferences across districts in Detroit than party activists.19 According to Converse, with the opinionated activists removed, there should be no differences between followers of the two political parties, much less stability of these differences across districts. And finally, “1339., p.242. ”David M. Dobkin, "Political Cynicism and Liberalism-Conservatism: Stability and Instability," (Unpublished Master's Thesis, Department of Political Science, University of Oregon, I964), p. 77. '8V.O. Key, Jr., Public Opinion and American Democracy (New York, I96I), p. 86. ”Samuel J. Eldersveld, Political Parties: A Behavioral Analysis (Chicago, I964), p. 20] . 64 the discovery of attitude dimensions by means of scalogram analysis would seem to be contrary to the great randomness indicated by Converse. In my study I will be interested in showing how opinions between political strata are shared, rather than dealing with variables affecting individuals' opinions. The central tendency in a distribution of opinions for an aggregate of people would not be affected if a substantial proportion Of the contributors to that distribution chose their opinions randomly. The spread of the distribution, however, would increase greatly. If the responses of the persons who are not content sensitive were always in a particular direction, such as always being agreeable to the item, my findings could be affected . Such persons can be identified by asking them their opinions on two items which, because they are logically reversed, would solicit a reversal of Opinions from persons who are sensitive to the content of the items. The yea-sayers, however, would agree and the nay-sayers would disagree with both items. This phenomenon is dealt with in Appendix D. If a large percentage of the followers were to be yea-sayers or nay-sayers in their responses to the policy preference items used in this study, relationships between leaders and followers reported in the following analysis chapters would be affected . True yea-sayers and nay-sayers were found, however, to be a very small minority. Furthermore, nay-sayers were found not to have more disapproving attitudes on the issues than the entire sample, nor were yea-sayers more approving of the items. The phenomenon has little impact on these policy preferences . Note also that the 46.5 per cent Of the populace identified as being 65 content sensitive is nearly twice as great as the approximately 25 per cent of the papulace which would be so identified if Converse's model of random re5ponses were correct. On four items with agree-disagree response alternatives, sixteen patterns of response are possible. If answers were entirely random, each of the patterns would be equally chosen, and thus, one-sixteenth or 6.25 per cent of the sample would answer in each of the patterns. Four patterns indicate content sensitivity. Using "A" for agree and "D" for disagree, these patterns are: ADAD, DADA, ADDA, and DAAD . Thus, only 25 per cent of the populace should be identified by chance as content sensitive, not 46.5 per cent. Furthermore, 43 per cent of the sample answered in the first two of these patterns, ADAD and DADA, in contrast to the l2.5 per cent of the sample that would be expected to so answer the questions by chance. Converse's model is not correct for these attitude items. Opinions and Behavior. The reSponses of members of all of the samples to these issue items are viewed as their personal opinions on these items, Opinions which would largely determine their behavior if they were given the Opportunity or encouraged to act with respect to these issues. I am aware that situational and other attitudinal variables might greatly affect the acting out of these attitudes, especially among leaders. I have tried to indicate my realization that this is a limited study of behavior because of my excluding these variables from consideration, as I did in Chapter i. I feel that a multi- variate study of this question is premature at this time, in as much as we are ignorant of the effects of contributing variables. More elaborate and multivariate research designs will undoubtedly follow this study before we can Say we understand leader-follower relationships in the community. Other Concepts Representativeness. Perhaps the most important concept used in this study is representativeness. Representativeness is measured by the congruency of the distributions Of leaders' and followers' attitudes toward the same issues. Because no measure fully quantifies the congruence of two distributions, two measures Of the distributions were compared, the means and the standard deviations. Although the shape of a curve itself is important in its measurement and, with the mean and the standard deviation, define a curve, most of these distributions are unimodal . Thus, finding that two distributions have very nearly the same means and standard deviations is strong evidence that the curves are greatly congruent. Differences between the leaders' and the fol- lowers' means are the most used data in the following analysis. In some hypotheses, differences between standard deviations are used in the analysis Of the data . Also, in a very few instances a comparison Of the . percentages of persons answering "Strongly disapprove " and "Strongly approve" are used in the analysis. This is done because of this measure's sensitivity to increasing division or polarization Of attitudes toward the issues with increasing involvement. Inclusiveness. Another concept that is used in the comparison of leaders' and followers' attitudes is that Of inclusiveness. If the means of the distributions of attitudes of the various leadership groups under discussion are both greater and less than the mean of the public, the leadership groups are 67 Said to have attitudes which are inclusive of those of the public . To be more brief, I will frequently say in such cases that the leadership groups are inclusive of the public. Finding that a distinction among leaders does not result in inclusiveness is a roUgh indication that preserving or encouraging division among leaders on this distinction is not likely to be of much aid in improving the representativeness of leadership. Vulnerability. One of the more important distinctions among leaders is that made between those who are vulnerable to popular control and those who are less vulnerable to such control. AlthOUgh a finer classification could be made along this variable, only this dichotomy was used in this study to preserve adequately large n's. The vulnerable leaders are present holders of public office and present candidates for public office. Social Status and General Political Involvement. Each individual in the random sample of followers was assigned to two strata--his social status level and his general political involvement stratum. Social status is measured on the basis of education, income, and job status. See Appendix C for the procedures by which the index was constructed. General political involvement is an index of three dimensions of involvement in the community political process: discussion of local political matters with friends, civic or county leaders, and city or county officials; attendance at meetings where matters of city government were discussed; and taking an active part in a local issue. Again, see Appendix C for the procedures used . Involvement in an Issue. In addition to the general political involve- ment, the involvement of each individual in each issue was assessed. For 68 each issue he was asked to indicate his degree of involvement. The alter- natives were: "l'laven't heard about it"; "Does not matter tco much to me"; "Interested but haven't done anything about it"; "Have talked about it with friends or vauaintances"; and "Have taken an active part on one side or the other.‘' The percentage answering "Have talked about it with friends or acquaintances" was used as a measure of the involvement of the community in the issue. The Unrepresented Citizen. The final concept used in the study is the unrepresented citizen. These citizens are merely those members of the random sample of followers whose attitudes on any one of the issues are not shared by any leader in Springfield and by less than 5 per cent of the leaders in Eugene . Evaluatingthe Hypotheses Tests of statistical significance have increasingly come under attack. For the most part these attacks are criticisms of the blind application of these tests in instances where they are inapplicable or they are arguments against the necessity of finding statistically significant results rather than criticisms of the utility and meaningfulness of properly applied tests.21 The very nature of the samples used in this study precludes the use of such tests when part of the data 20 Agger, Goldrich, and Swanson, p. 688; Hanan C. Selvin, "A Critique of Tests of Significance in Survey Research," AmericrajnfiSociOIOgical Review, XXIII (AUgUSI’, I958), pp. 5l9-527; and Leslie Kish, "Confidence Intervals for Clustered Samples," American SocioIOgical Review, XXII (April, I957), pp. 154-165. 2IRobert McGinnis, "Randomization and Inference in Socio|Ogica| Research," American SocioloQicaI Review, XXII (October, I957), pp. 408- 414. ’ '69 is contributed by either of the leadership samples. Differences noted in the reputational leaders sample are a phenomenon of real differences and differences caused by the crudeness of the measuring instrument and coding-processing errors. As this sample may be considered a universe of leaders having a certain amount of influence, no sample error is present. Tests of statistical significance are inapplicable. Other than hoping that the instruments are adequate and exercising maximum caution to hold coding and processing errors to a minimum, no other techniques are available for distinguishing real differences from errors. Since there is no reason to expect either instrument error or coding-processing error to be systematic, this absence of techniques is not an obstacle . In addition to these errors, the potential leadership sample, as a whole, includes sampling error. And because of the non-random nature of this sample, no estimate of the amount of such error can be made. The use to which these data are put is limited accordingly. Fortunately, each of the background types among the potential leaders is a universe of leaders with that background, thus, no sampling error is present. The entire sample of potential leaders has sample error, but each of the background types included within it is a universe with no sample error. Differences between such background types are the primary use to which these data are put. The majority of the hypotheses deal with differences in means between the random samples of followers and these leadership samples. Each issue is used as a separate case for evaluating the hypothesis. Some confidence that 70 the discovered relationship is not entirely sampling error in the random sample or instrument measurement error and coding-processing error in both samples is gained from discovering consistency in the relationship across the nine issues and across the two communities.22 AlthOUgh the decision as to whether a relationship is strong enough not to be sampling error alone using such a con- sistency measure is somewhat arbitrary and subiective, it differs from trying to apply tests of statistical significance to this data only in that its weaknesses are more manifest. Because I am primarily concerned with the direction of the relationships and because there is no standard with which to compare the magnitudes of the differences except between the two communities, most of these comparisons are made using only the sign of the differences. 22Gabriel A. Almond and Sidney Verba, The Civic Culture (Princeton, I963), pp. 523-525. CHAPTER IV TWO COMMUNITIES AND NINE ISSUES This chapter is intended to give the reader some basic sociOIOgical information about the two communities investigated in this study as well as to analyze several relationships which must be understood to evaluate the hypotheses in Chapters V and VI. To cover such relationships at the time they are used in the analysis of the data relevant to the hypotheses would distract from the actual analysis of the hypotheses. For the moment the hypotheses are put aside . Two Communities Springfield and Eugene, Oregon are sister cities located at the southern end of the fertile Willamette River Valley. The river is usually thought of as the boundary between the cities, but Springfield is slightly upriver from Eugene and, shortly before this study, Eugene annexed a suburban area on the other side of the river, putting Eugene on both sides. A north-south Interstate high- way is the natural boundary between the two cities at this time. Althoth the cities are in close proximity, no community of interest exists to cause them to work more closely with each other than with any other medium-sized city in the state. In fact, relatively little interaction takes place between them at a governmental or other community-wide basis. Because of Eugene's greater size and its inexpensive, municipal services, EUgene has always served as a 7i 72 model to be emulated and a threat to Springfield autonomy. This is the primary impact that the two cities have had on each other. Eugene's popula- tion was about 50,000 persons at the time of this study and Springfield's about l2,000. The two cities share many of the problems associated with all north- western cities: an abundance of labor caused by the climatic and geOgraphic desirability of the region, and the difficulty of attracting new industry to an area relatively remote from market areas in the midwest, east, or southwest. The influx of people into the area, many of them unskilled or retired, has confronted both communities with the problem of providing increased and new municipal services without a greatly expanded revenue base. Despite these similarities, however, the contrasts between the communities at the socioIOg- ical level are most striking. A temperate climate, an abundance of evergreen trees, and rolling hills, all combine to make EUgene a pretty city. In the valley between these hills and the river are numerous businesses and offices as well as the University of Oregon. These are the primary sources of employment in Eugene. The main shopping area is in the older area of the city. This area, though clean and neat with many modern shops, suffers the common problems of all such pre- automobile areas, namely insufficient parking and a strangled traffic flow. Across the Willamette river on the flat river plains is the newly annexed area of Eugene. It is an example of middle class suburbia with its artificially curved streets, identical houses, absence of trees, and new schools. Similar 73 suburbs border Eugene on other sides. Though Eugene has blighted areas, it gives the appearance of a medium-sized, middle-class city. The large number of professional people and the white collar employees of the various shops and offices who spill onto the streets during the lunch hour, after work, and on the weekends help to fulfill this image of the middle-class city. EUgene's politics reflect this middle-classness. The elections in the city are formally non-partisan, but elected officials are primarily Republicans as is the registration in the community, 52 per cent Republican at the time of the study. The manager-council form of government is used in the city. In contrast, Springfield gives the impression of being a predominantly working-class community. There are many areas of poorly designed and poorly kept homes, sufficiently numerous to be characteristic rather than the exception . The shopping area is spread along the east-west highway with an overabundance of cheap cafes and used car lots. Springfield grew to its present size from a population of 2500 people in l940 primarily as a result of the rapid growth of the lumber industry which had developed because of the demands of World War II and the depletion of the forests of Washington. This industry is still the predominant source of employment in the city, and signs of it are everywhere and ever-present in the form of large lag trucks and the pungent odor of one of the wood processes. Springfield is an excellent example of the one-industry town with its best opportunity for employment available to the skilled and unskilled laborer. 74 Politics in Springfield reflect the predominance of the large numbers of working class persons in the city, as well as the continuance of party identification among those who migrated here from the South . Democrats outnumber Republicans in the registration. As in Eugene, elections in Spring- field are non-partisan. Unlike Eugene, however, Springfield's politics are characterized by continuing conflict over various issues such as whether to have a municipal power company or not. In summary, Eugene is the Republican, middle-class, white-collar city; and Springfield is the Democratic, working-class, blue-collar city. This class-based difference between the communities is evident in Table 4-1 . Table 4-l: Comparison of the percentage of persons in each social class. Social Class* Springfield Eugene N = 429” N = 504** 0) Low 15.2 “.5 1) 29.8 16.] 2) 25.2 23.6 3) High 14.7 28.2 Retired “.0 18.5 Single woman, 4.2 2.2 no occupation *See Appendix C for definition . MrStudents and persons giving insufficient information excluded . Attitudes on Nine lssues Considering the sociolOgical contrasts of the two cities, what is most striking in Table 4-2 is the similarity with which the publics of the two communities react to the issues investigated in this study. Whether one looks at the means on each issue, the standard deviations, or even the distributions of attitudes themselves, the differences between the communities are very small. The greatest difference is on the issue of urban renewal. in this case Springfield citizens are more divided but, on the whole, more in favor of the issue than the citizens of Eugene. The second tendency evident in the table is the overwhelming approval in both communities for adoption of the pragrams.I Assuming that these atti- tudes are predictive of voting behavior, only publicly supported kindergartens in Springfield and, perhaps, in EUgene seem to have the possibility of being defeated in an election . Two policy proposals, the desirability of attracting new industry to the city and the expansion of special education, are so overwhelmingly approved that they reach the level of approval normally classified as "consensus. "2 Opposition or disapproval is very limited except ]This concept of the approval given to the policy proposal around which an issue has developed is used extensively throughout the study. Each issue concerns a proposal that govemr..ent undertake a certain action, thus to take the approval side on an issue indicates the willingness to have government undertake the pragram. For the sake of brevity, I will speak of "favoring the issue," "having favorable attitudes on an issue," or iust being "favorable." In all cases this statement means that this group or category favors the adoption of the proposed policy or program. 2Herbert McClosky, "Consensus and Ideology in American Politics," The American Political Science Review, LVlll (June, 1964), p. 363. 76 Table 4-2: Comparison of policy preferences in Springfield and Eugene by percentages holding each attitude. Stand. SA* A U D SD DC DK NA Mean Dev. (0) (I) (2) (3) (4) Attracting Industry Springfield 52.7 40.0 2.9 .9 .2 2.2 .4 .7 .506 .25 Eugene 51.0 40.3 3.4 1.3 .8 I I 1.3 8 .559 .701 Annexation Springfield 12.3 44. 1.4 8.9 2.7 8.5 1.6 .71.392 .944 Eugene 15.6 41. 21.1 5.3 2.1 10.3 3-0 1.01.261 .909 Parking Lots Springfield 6.9 54.7 I 9.8 2.7 4.7 2.5 1.1 1.418 .884 EUgene 12.2 44.9 1 11.2 4.6 5.7 3.8 1.1 1.453 1.037 Special Education Springfield 17.4 56.3 1 8.3 2 7 1.3 2.2 .71.191 .9275 Eugene 24.0 48.3 10 8.6 3 8 1.7 1.7 1 1 1.161 1.027 Fluoridation . Springfield 17.0 3 1 17. 12.7 14.3 3.8 1.8 .71.735 1.319 Eugene 27.0 1 10.5 18.1 19.0 4.2 1.3 1.01.819 1.525 Public Housing Springfield 6.3 46.7 5 1 .9 3.6 2.9 3.1 .91.558 .919 EUgene 10.5 39.5 25 1 112.2 4.6 2.9 4.2 1 1 1.573 1.016 Urban Renewal Springfield 13.4 30.1 23.7 11.4 12.7 5.8 2.5 .41.779 1.244 Eugene 10.5 26.6 28.9 6 1 5.1 10.3 11.6 1.0 1.596 1.032 Metropolitan Park Springfield 12.1 56.5 .2 8.0 23.0 6.0 1.3 .91.253 .869 EUgene 16.5 47.0 15.2 6.7 3.4 7.0 3.0 1 1 1.250 .970 Public K indergartens Springfield 8.3 30.4 17.0 29.5 11.6 1.8 .9 .7 2.060 1.198 Eugene 12.0 29.8 12.7 24.9 14.8 2.9 2.3 .6 2.008 1.310 *The meanings of these abbreviations are: SA--Strongly Approve, A-Approve, U--Undecided, D--Disapprove, SD--Strongly Disapprove, DC--Don't Care, DK-Don't Know, NA--No Answer. on the fluoridation issue in both cities, urban renewal in Springfield, and public kindergartens in both cities. The citizens of Eugene and Springfield respond quite similarly in their approval of these issues. Involvement in the Issues Again, as in Table 4-2, the most conspicuous pattern shown in Table 4-3 is the similarity in responses in Springfield and Eugene. The cities vary greatly in their responses to only three issues. The citizens of Eugene are more involved than Springfield citizens on the issues of parking lots, fluoridation, and public kindergartens. Springfielders are more involved in the urban renewal issue. If the various alternative responses available to indicate involvement in the issue questions are ordered on a continuum, one would expect the more involved community to have fewer individuals indicating little or no involve- ment and more individuals indicating moderate or great involvement than in the less involved community. This is the case, thus supporting the ordering of the altematives on the question as well as giving some support for the unidirnen- sionaIity of the attitude measured by the question. But there is evidence of the interference of other variables in two of the response categories. if we look at the range of percentages of either of the cor-.imunities' responses to each of the reSponse categories (Table 4-s‘I-), the variations in the "Don't care" and "Have taken an active part" responses are small in comparison to the ranges of percentages of other categories . The stability of these two reSponse alter- natives would seem to indicate that they are measuring some other variable 78 Table 4-3: Comparison of the involvement in the issues in Springfield and Eugene by the percentages of respondents giving each answer. Haven't* Doesn't Interested, TaTked Taken Don't Issue Heard Matter Done With Active Know Nothing Friends Part Attracting Industry Springfield 18.5% 9.4 37.3 31.5 1.3 2 .0 Eugene 17.1 7.0 32.9 39.7 1.9 1.3 Annexation ’ Springfield 11.4 16.5 27.9 40.0 2.0 2.2 EUgene 11.2 20.9 24.7 38.6 2.9 1 7 Parking Lots Springfield 17.2 19.4 31 9 26.8 1 8 2.9 EUgene 4.0 9.3 26 6 57.2 1 5 1 3 Special Education Springfield 1 .6 8.0 40.4 28.3 3.3 2.2 Eugene .2 9.5 35.0 37.5 4.9 1 0 Fluoridation Springfield 8.7 9.8 29. 43.3 6.0 2.6 EUgene 4.4 8.6 15.0 60.5 9.9 1.7 Public Houfing Springfield 23.4 7.8 34.8 28.6 2.7 2.7 Eugene 24.0 9.7 35.7 27.8 1.5 1.3 Urban Renewal Springfield 10.3 10.3 21.7 48.9 6.5 2.5 Eugene 28.7 16.9 24.3 26.0 2.1 1.9 Metropolitan Park Springfield 21.9 12.7 36.2 25.4 1.3 2.4 EUgene 2 1 6 32.9 30.8 1.5 1.6 Public K inde rga rtens Springfield 30.6 14.3 27.9 22.8 2.0 2.4 EUgene 20.5 14.1 27.4 33.1 3 8 1.1 *The meanings of these mneumonic statements are: Haven't Heard-- Haven't heard about it, Doesn't Matter--Doesn't matter too much to me, Interested, Done Nothing--Interested but haven't done anything about it, Talked With Friends-- Have talked with friends or acquaintances about it, Taken Active Part--Have taken an active part on one side or the other. c . I ‘II III}! .111 I 1 II it 79 Table 4-4: Minimum and maximum percentage of either community on any issue using the various alternative reSponses for the issue involvement question. Alternative Minimum Percentage Maximum Percentage Responses Answering Answering Haven't heard. 4.0% 30.6 Don't care. 7.0 20.9 Interested but haven't done anything. 15.0 40.4 Have talked with friends. 22.8 60.5 Have taken an active part. 1 .3 9.9 besides the degree of individual involvement in the issue. Numerous hypotheses could be spun as to the variable or variables that interfere with involvement in the issue for the "Don't care" and "Have taken an active part" responses. But I shall be satisfied at this time to conclude that these responses should be excluded when comparing differences in community involvement in the issues. Using the "Have talked with friends" response as an indicator of involvement in the issue and 35 per cent as the cutting point between issues in which the community is involved and those in which it is not involved, the following issues are identified as not involving the community: attracting industry in Springfield, parking lots in Springfield, special education in Springfield, public housing in both communities, urban renewal in Eugene, metropolitan park in both communities, and public kindergartens in both communities. No pattern relating issues which most involve the communities is discernible. 80 Issue Involvement and Attitudes Toward the Issues The communities' striking similarity on the distribution of attitudes toward the various issues and their varying degrees of involvement with the same issues would seem to indicate no relationship between involvement in an issue and favorable attitudes toward it. Table 4-5 shows the means of the attitude responses for each involvement level. It is quite evident that with the exception of public housing, there is an increase in favorable attitudes toward the issue with each increase in involvement in the issue. The general pattern which is shown by ten of the eighteen issues reveals a slight decrease in favorable attitudes toward the issue from the "Haven't heard" response to the "Don't care" re5ponse. From there on, favorable attitudes increase with increased involvement. Issues showing this pattern have asterisks beside them in Table 4-5. In both communities the relationship in the public housing issue shows decreasing favorable attitudes with increasing involvement. This reversal of the pattern will be repeatedly discovered in the issue of public housing . Persons more involved in an issue also tend to favor the pragram involved yet, in a community showing lesser involvement, the same number of persons favor that issue as in the community which is more involved . Should not the community showing greater involvement of persons involved also show more favoritism toward the program? If in the community showing the lesser involvement, each category of involvement included more persons favoring the pragram than in the comparable category of the other community, the two 81 Table 4-5: Means of the distributions of attitudes toward the issues of each of the issue involvement categories. Issue Involvement Springfield Eugene Attraction of Industry Low .810 * .807 * .911 1 .000 .485 .596 .302 .382 High .166 .200 Annexation Low 1 .447 * I .594 1 .892 1 .554 1 .384 1 .245 1 .255 1 .041 High 1 .000 1 .800 Parking Lots Low 1 .529 1 .187 1 .493 1 .606 1 .345 1 .534 1 .421 1 .441 Iligh .750 .625 Special Education Law 1 .338 1 .232 1 .656 1 .688 1 .137 1 .252 1 .087 1 .035 High .866 .346 Fluoridation Low 1 .606 2.111 1 .870 1 .935 1 .804 1 .756 1 .709 1 .867 High 1.555 1.529 Public Housing Low 1 .510 1 .587 1 .689 2 .045 1 .506 1 .516 1 .527 1 .460 High 2 .250 2 .125 Table 4-5--Continued Issue Involvement Springfield EUgene Urban Renewal Low 1.645 * 1 .641 1 .965 1 .655 1 .785 1 .504 1 .796 1 .568 High 1 .586 2 .400 Metropolitan Park Low 1 .383 1 .326 * 1 .568 1 .391 1 .228 1 .240 1 .027 1 .194 High 1 .333 .625 Public Kindergartens Low 2.315 * 2.072 * 2 .383 2 .442 1 .920 1 .978 1 .718 1 .810 High 1 .444 2 .052 communities would, on the whole, have equal numbers of persons having favorable attitudes on the issue. Even if this were true only of the "Interested but haven't done anything" and "Have talked with friends" categories, this explanation of the contradictory findings would be true since these two categories are quite large. The explanation seems to be correct. In 13 out of 20 cases, the involve- ment categories of the community showing lesser involvement are more favorable than‘their counterparts in the community showing the greater involvement. It is especially true in two issues. In the issue of attracting new industry to the city, the community showing the lesser involvement, Springfield, has more persons favoring the pragram in four of the five involvement categories than does Eugene. The same 83 is true of EUgene for the urban-renewal issue. In both of these cases, the community showing the lesser involvement has more persons favoring the prOgram in each involvement category except one. In two other cases, where the pattern is not evident, the community showing the lesser involvement is also the com- munity showing the fewer number of persons favoring the issue. One would be hard-pressed to explain this phenomenon. It seems to be a random result caused by the predominant size of the "Interested but haven't done anything" and "Have talked with friends" categories, and by the relative weakness of the relationship between favorable attitudes and involvement for these particular categories. Polarization with Increased Involvement The relationship between increasing involvement and favorable attitudes would lead one to believe that, as one goes to the higher levels of involvement, the mode of distribution in an issue would generally move toward more favorable attitudes. Table 4-6 shows that many issues do meet this expectation. For example, the special education issue in Springfield shows that, with increasing involvement in the issue, there is a decrease in the percentage of persons answering "strongly disapprove" and an increase in the number answering "strongly approve. Other issues showing this pattern are: parking lots in both communities, special education in Eugene, and the metropolitan park issue in both communities. Many issues do not show this pattern. Instead, they show rectangular or even bimodal distributions among the more involved members of the public, Table 4-6: Polarization of attitudes with increasing involvement in the issue measured by percentages of sample in strongly approve and strongly disapprove categories. Involvement Springfield Eugene Issues SA SD SA SD Attraction of Industry Low 31 .3 0 33.3 0 19.0 0 16.2 0 53.5 0 45.7 6 72.5 7 68.4 1 4 High 83.3 0 80.0 0 Annexation Low 3.9 .0 3.4 .0 .0 1.4 6.4 .9 13.4 2.4 10.0 .8 17.7 4.4 28.1 3.0 High 44.4 11.1 20.0 20.0 Parking Lots Low 5.1 .0 9.5 .0 1.1 2.3 4.1 2.0 7.6 1.4 11.4 7.1 9.8 5.7 12.6 .3 High 25.0 .0 62.5 .0 Special Education Low 10.1 2.5 20.3 3.1 13.9 8.3 8.0 6.0 16.8 2.7 18.5 4.9 24.2 1.6 28.9 3.0 High 20.0 .0 69.2 .0 Fluoridation Low 7.7 .0 4.3 8.7 .0 .0 2.2 4.4 11.3 15.8 24.1 11.4 23.4 17.3 28.6 22.6 High 44.4 25.9 53.8 28.8 Public Housing Low 4.8 .0 5.6 2.4 2.9 .0 2.0 3.9 3.1 3.8 9.6 2.7 12.4 3.9 19.2 8.9 High 8.3 33.3 12.5 12.5 Table 4-6--Continued Involvement Springfield Eugene Issues SA SD SA SD Urban Renewal Low 2.2 .0 2.6 .7 .0 2.2 2.2 .0 9.1 9.1 14.1 4.7 18.1 17.6 20.4 10.9 High 34.5 27.6 27.3 45.5 Metropolitan Park Low 8.1 2.0 14.0 3.5 5.3 3.5 8.2 1.6 11.6 1.8 15.0 2.9 20.9 .9 22.2 4.9 High .0 .0 50.0 .0 Public Kindergartens Low 2.9 13.9 5.6 0.2 3.1 9.4 .0 8 7.9 11.8 9.0 18.1 17.3 10.6 23.0 .9 High 33.3 .0 20.0 0 in contrast to the nearly normal distributions of the less involved. I have called this bimodality or division of attitudes in an issue, polarization. Increasing polarization with increased involvement is strongly evident in the fluoridation issue in both communities. The strength of this relationship, however, varies greatly between issues. There is a strong indication that the issues which show large increases in polarization are those which most involve the publics in the communities. Seven of the ten issues which were classified earlier as not involving the particular city are less polarized than the least polarized of the more involving issues. When speaking of the polarization of an issue, I am speaking of the number of involvement categories polarized on the issue. If all five categories showed polarization of attitudes on a given issue, that issue 86 would be highly polarized . Figure 4-1 presents an explanatory model of issue polarization . This model is consistent with the theories of Schattschneider and Degree of Involvement in the issue Low Medium High (A) B) (C) Activists Polarized on the issue X X Non-activists Polarized on the X Issue Figure 4-1: The possible relationship between involvement in an issue and issue polarization . Dahl. They believe that conflict develops within the leadership and because of the institutions of democracy, spreads or is socialized to include the public. Excluding the apparently consensual issue of attracting industry, for which there is strong approval in both communities, the following issues appear to be of the type "A," as indicated in the model: parking lots in both communities, special education in both communities, and metropolitan parks in both communities. All of these issues show stable or decreasing percentages of "strongly disapprove" answers and varying slight increases of the percentage answering "strongly approve . " 3E.E. Schattschneider, The Semisovereign_People (New York, 1960), p. 138; and Robert A. Dahl, Who Governs? (New Haven, Connecticut, 196]), p. 322. 87 Il')" The issues which seem to be of the type a are: annexation in both communities, public housing in both communities, and urban renewal in both communities. These issues are characterized by the high polarization of the more involved and by the great decrease in polarization between these categories and that of the less involved . Fluoridation and public kindergartens in both communities approach being type "C" issues. Public kindergartens would seem to be a dormant type "C" issue because this issue is almost equally divisive at all levels of involvement. Yet, it does not appear to be an issue which involves many people in either community. There is no issue of the opposite type in which there is great involvement without polarization of attitudes. This fact would tend to offer some evidence that the casual relationship is not that, as implied thrOUghout the previous discussion, increased involvement leads to poiarization of attitudes. But rather, issues on which attitudes are polarized are likely to involve the community to a great extent. Undoubtedly, the dynamics of the development of a divisive issue are cyclical. Given a certain divisiveness or polarization on an issue, involvement is likely to increase and, thereby, further polarize opinion on the issue. The Generalig of Political Involvement Table 4-7 gives the relationship between general political activity and the number of issues in which a person takes an active part. Twenty-four of the 36 persons who are in the highest general political involvement stratum in 88 Table 4-7: Degree of general political involvement for persons active in one or more issues, by percentages. General Political Involvement Number of Issues in which Low High Total an individual is involved (0) (1) (2) (3) % N 1 18.3% 56.7 17.3 7.7 100.0 104 2 9.4 50.0 21.9 18.8 100.1 32 3 27.3 36.4 36.4 100.1 11 4 75.0 25.0 100.0 4 5 50.0 50.0 100.0 2 6 33.3 66.7 100.0 3 7 50.0 50.0 100.0 4 Percentage of Each 13.8 50.6 20.6 15.0 100.0 160 Category Among Persons ‘ Involved in One or More Issues Percentage of Each 33.1 53.1 10.0 3.8 100.0 954 Category in Combined Communities both communities have taken an active part in one or more of these issues. Papulating the “high" general political involvement column in Table 4-7, they constitute only 15 per cent of those taking part in one or more issues. The more frequently a person takes an active part on issues, the more likely he is to be generally politically active. But on any given issue persons of less general political activity are in the majority. This is especially true of fluoridation in which 43 of the 82 persons taking an active part are not active in any other issue. Although there is a strong relationship between involvement in various issues and general political involvement, many members of any stratum of general political involvement, even the most active, will not be actively involved in a given issue. 89 Given the strong relationship between involvement in an issue and polarization of attitudes toward it, one might expect that the less issue-involved persons in each general political involvement stratum would decrease the polarization differences between the highest and the lowest strata. In part this does seem to be the case. None of the issues in Table 4-8 show the marked increase in polarization that is shown in he same issues in Table 4-6. But with the exception of the consensually approved issue of attracting industry, only one issue, the metropolitan park issue in Springfield, does not show evidence of some issue polarization. In addition, parking lots, special education, and the metropolitan park issue in Eugene show a low level of strong disapproval which remains stable with increasing general political involvement. But in general, polarization with increasing involvement is more evident in Table 4-8 than in Table 4-6, though not as strongly in some of the issues. The best explanation of this effect is a baseline of polarization among those who are general political activists. On most issues of policy, such as those dealt with in this study, a minority of persons strongly disapproving of the issue will exist. Few, if any, of these persons take an active part in the issue they oppose; thus, if one looks at those who are active in a type "A," low- involvement issue, one will not find persons who strongly disapprove of the issue. It is indeed interesting that none of this strongly disapproving minority take an active part in such issues. Whether these persons are consistently the some persons or if there is some other variable that keeps activists opposed to such issues from getting actively involved in an issue, is a question deserving further research . 90 Table 4-8: Polarization of attitudes with increasing general political involve- ment by percentages of sample in strongly approve and strongly disapprove categories. Springfield , Eugene Issues involvement SA* SD SA SD Attraction of Industry Low 37.1 .6 43.0 .0 61.9 .0 53.3 1.1 _ 62.5 .0 68.3 1.6 High 60.0 .0 47. .0 Annexation Low 7.5 2 .5 10.1 .0 14.0 3.4 15.3 1.5 18.8 .0 25.4 6.3 High 26.7 6.7 33.3 9.5 Parking Lots Low 6.9 2.5 7.0 3.8 5.5 2.5 13.1 4.4 6.3 .0 17.5 4.8 High 26.7 13.3 28.6 4.8 Special Education Low 13.8 4.4 17.1 4.4 20.8 .4 24.8 2.2 9.4 6.3 33.3 7.9 High 26.7 13.3 42.9 4.8 Fluoridation Low 13.8 12.6 15. 17.4 15.3 32.1 19.0 25.0 15.6 39. 23.8 High 33.3 13.3 19 0 33.3 Public Housing Low 6.9 3.1 12.7 3.8 5.9 2.1 8.4 5.1 3.1 9.4 15.9 3.2 High 13.3 20.0 9.5 9.5 Urban Renewal Low 7.5 7.5 6.3 1.9 14.8 14.0 12.0 4.7 21.9 25.0 15.9 11.1 High 40.0 20.0 9.5 19.0 91 Table 4-8--Continued Springfield Eugene Issues Involverent SA* SD SA SD Metropolitan Park Low 10.7 3.1 10.1 3.2 12.3 1.7 17.5 3.6 15.6 .0 23.8 1.6 High 20.0 .0 38.1 4.8 Public Kindergartens Low 8.2 11.3 7.6 16.5 8.1 12.3 14.6 13.5 6.3 6.3 12.7 9.5 High 13.3 13.3 14.3 33.3 *The meanings of these abbreviations are: SA—Strongly Approve and SD--Strong|y Disapprove. The baseline of polarization provided by this minority and the dampening effect of the general political activists not involved in a given issue explain the differences noted between Table 4-6 and 4-8. Polarization and favorable attitudes toward issues are both relationships which are evident with increased involvement, whether involvement is measured in a particular issue or generally. The percentages of the "strongly approve" and "strongly disapprove" responses have thus far been used to show the increasing divisiveness of attitudes toward the issues among more issue involved or generally involved individuals in the community political process. AlthOUgh this is not a commonly used measure of distribution spread, it did clearly portray this polarizing of attitudes. Standard deviations are more frequently used as measures of distribution spread . Table 4-9 shows the standard deviations for each of the general political Table 4-9: Standard deviations of the distributions of attitudes toward issues for each category of general political involvement. Springfield EUgene Issues Involvement High - Low High - Low Involvement Involvement Attraction of Industry Low .689 .650 .572 .745 .550 .675 High .490 -.199 .499 -.151 Annexation Low . 942 . 773 .960 .874 .819 1 .042 High 1 .186 .244 1 .325 .552 Parking Lots Low .923 .918 .799 1 .055 .910 1 .067 High 1 .356 .433 1 .243 .325 Special Education Low 1 .019 .980 .820 .988 .921 . 1 .233 High 1 .298 .279 .980 .000 Fluoridation Low 1 .278 1 .360 1 .329 1 .556 1 .364 1 .652 High 1.288 .010 1 .499 .139 Public Housing Low .909 .939 .846 1 .014 1 .063 1 .036 High 1 .298 .389 1 .179 .240 Urban Renewal Low 1 .067 .835 1 .274 1 .045 1 .487 1 .200 High 1 .543 .476 1 .328 .493 Metropolitan Park Low .963 .915 .837 .997 .711 .856 High .771 -.192 I .020 .105 93 Table 4-9--Continued Springfield Eugene lssu es Involvement High -Low Involvement Public Kindergartens Low 1 .209 1 .261 1 .195 1 .326 1 .092 1 .262 High 1 .288 .079 1 .388 .127 involvement categories and the increase in the magnitude of the standard deviation from the least involved stratum to the most involved stratum . To get a comparable measure from Table 4-8, the percentages of the strongly approving and strongly disapproving were added for each stratum, and the difference between the lowest and the highest political strata were measured . Again excluding the consensual attraction of industry issue, the Spearrnan r for the rankings of the magnitudes of the increased divisiveness measured by both measures is .424. The standard deviation uses more of the data than the sum of the "strongly disapproves" and the "strongly approves" and shows the same pattern of greater division among the more involved. But it is not as sensitive to the continuity of this pattern shown in the distributions as the measure of polarization. The measure of polarization is, however, of much more limited use in assessing the spread of opinion. The standard deviation is sensitive to all differences. Except in those hypotheses where the polarization with increasing involvement is of concern, the standard deviation will be used . It should be noted that the relationships shown in Table 4-9 support my 94 classification of issues into the dimensions of involvement and polarization. To be consistent with Figure 4-1, issues which do not involve or divide either the community or its more involved members would be expected to have generally low standard deviations and little increase in standard deviations with increasing involvement. The metropolitan park issue in both communities is of this type. The issues in which the community is most involved would also show little change with increasing involvement, but all strata of involvement would be greatly divided . Public kindergartens and fluoridation are of this type. Finally, the issues which moderately involve the community should show a very large increase in the standard deviations with increasing involvement. The urban renewal issue in both communities is of this type. Table 4-10 shows the average standard Table 4-10: Comparison of issue divisiveness types on average standard deviation and average increase in standard deviations with increasing involve- ment. Divisiveness Average Average Increase in of the Issue Standard Standard Deviations Type Deviation with Increased Involvement Type "A" .982 .158 Type "B" 1 .083 .399 Type "C" 1 .334 .089 deviation and the average increase in standard deviations with increased involve- ment for each of these types of issues. The relationships are as expected . 95 General Political Involvement and Attitudes Toward the Issues The relationship between increasing involvement and more favorable attitudes toward an issue is also obscured when general political involvement is controlled rather than involvement in each issue. This is shown in Table 4-11 . The procedures in the following analysis are again reversed for the consistently reversed public housing issue. Five of the cases, all in Eugene, do not show increased favorable attitudes toward the issue when general political activity is used . Only two cases failed to show the relationship in Table 4-5 when involvement in the issue itself was controlled. For the issues that do support the conclusion, the average decrease in the mean or increase in favorable attitudes, when controlling for general political involvement, is .354 between the lowest and the highest strata, compared to .501 when issue involvement is controlled . The relationship between increasingly favorable attitudes and involvement remains, but it is not as strong when general political involvement, rather than issue involvement, is controlled. It is the variable of general political involvement that will be of greatest use in the succeeding analysis chapters. I will be concerned with the existence of various biasing processes affecting the selection of community leadership and the direction of this bias. These processes are rather long-term phenomena, thus it is unlikely that only men taking certain positions on a given issue would become leaders. The issues are, for the most part, too current for such phenomena. Such issues as the racial issue in the South might be approached in this manner because such an issue is of continuing concern to the communities of the South . The Table 4-11: Means of the distributions of attitudes toward issues for each category of general political involvement. Issue Involvement Springfield Eugene Attraction of Industry Low .684 .677 .417 .562 .406 .380 High .400 .523 Annexation Low 1 .578 1 .316 1 .336 1 .252 1 .193 1 .1 18 High 1 .230 1 .380 Parking Lots Low 1 .510 1 .411 1 .339 1 .465 1 .516 1 .416 High 1 .400 1 .450 Special Education Law 1 .340 1 .206 1 .073 1 .144 1 .290 I .209 High 1 .333 .800 Fluoridation Low 1 .853 2 .020 1 .721 1 .698 1 .516 1 .606 High 1 .266 2 .550 Public Housing Low 1 .471 1 .338 I .517 1 .652 1 .843 1 .583 High 2 .333 2 .100 Urban Renewal Low 1 .777 1 .485 1 .808 1 .584 1 .687 1 .724 High 1 .466 1 .888 Metropolitan Park Low 1 .375 1 .330 1 .213 1 .272 1 .103 1 .035 High 1 .066 .894 97 Table 4-1 1--Continued Issue Involvement Springfield Eugene Public Kindergartens Low 2 .087 2 .040 2 .064 1 .950 1 .843 1 .881 High 2 .066 2 .650 relationships between involvement and favorable attitudes and involvement and polarization of attitudes are more evident when controlling for issue involvement, though they persist when controlling for general political involvement. Thus, investigating them gives us assurance that these relationships do exist. The marginals of issue involvement for the various issues also are good indicators of the community's concem with the issue . The Relationship Between Involvement and Social Status Although there is a strong relationship between general political involve- ment and social status as shown in Table 4-12, many members of medium and lower status categories are also members of the most involved political stratum. More than half of the top political stratum are not from the highest social status category. The relationship between attitudes on the issues and both social status level and involvement level must be evaluated to understand the bias, if any, of community leadership. Table 4-13 shows the above relationship also to be true if issue involve- ment is used as the measure of involvement. The highest social status level is the most involved level in eight of the nine issues, but variations are notable. 98 Table 4-12: stratum, in percentages. Social status composition of each general political activity General Political Social Status Level" Activity Total Stratum Low (0) (I) (2) (3) High % N Low (0) 29.0% 28.6 26.6 15.9 100.1 290 (I) 11.9 25.1 32.4 30.7 100.1 522 (2) 5.5 20.9 22.7 50.9 100.0 110 High (3) 10.0 13.0 32.6 43.5 100.0 46 _ *See Appendix C for definition. Table 4-13: Social status composition for the "Have taken an active part" response of issue involvement by percentages. Issue Social Status Level* Total Low (0) (1) (2) (3) High % N Attraction of Industry 9.1 .0 18.2 72 .7 100.0 22 Annexation 9.6 7. 7 32 .7 50 .0 100 .0 52 Parking Lots 11.8 29.4 29.4 29.4 100.0 17 Special Education 58.2 8.2 36.7 46.9 100.0 49 Fluoridation 7.3 13.4 1 .5 59.8 100.0 82 Public Housing 15.0 15.0 .0 55.0 100.0 20 Urban Renewal 8.1 18.9 .0 45.9 99.9 37 Metropolitan Park 5.9 17.6 11.8 64.7 100.0 17 Public Kindergartens 2.9 17.6 26.5 52 .9 99.9 34 *See Appendix C for definition. 99 It is apparent that many issues, such as attraction of industry and the metropolitan park issue, are primarily concerns of the highest social status level, while the parking lot issue attracts activists from all social status levels. Social Status and Attitudes Toward the Issues. Social and economic status has, as I have previously commented, often been used as an easily obtainable indicator of attitudes. A consistent, if not strong, relationship between increasing social status and increasingly favorable attitudes toward the issues is shown in Table 4-14. Again, public housing seems to be the reversed issue because both communities show decreasingly favorable attitudes toward the issue with increasing social status. Two issues, parking lots and the metropolitan park, show little or no relationship between these two variables. Other than these two issues, there is a class position on the issues which is generally favorable. Except for the attraction of industry issue, the communities show a remarkably similar relationship between increasingly favorable attitudes toward the issues and increasing social status level. ThOUgh the changes in favorable attitudes are not as'neatly continuous as for some of the previously presented relationships, this similarity of change to favorable attitudes with increasing social status does not mean that a given social status level in both communities equally favors the issue. For example, in EUgene each social status level is more in favor of public housing and urban renewal and less in favor of special education than its counterpart in Springfield . This is true in each case for three of the four levels. 100 Table 4-14: Means of the distributions of attitudes toward issues for each level of social status. Social Issue Status Springfield Eugene Low - High Low - High Attraction of Industry Law .575 .574 .504 .506 .490 .508 High .274 .301 .535 .039 Annexation Low 1 .508 1 .541 1 .504 1 .230 1 .339 1 .314 High 1.135 .373 1.093 .448 Parking Lots Low 1 .532 1 .437 1 .452 1 .486 1 .273 1 .473 High 1 .457 .075 1 .431 .006 Special Education Low 1 .253 1 .327 1 .338 1 .118 1 .037 1 .137 High .968 .285 1.021 .306 Fluoridation Low 2 .015 1 .872 1 .779 2 .120 1 .379 1 .805 High 1.327 .688 1.144 .728 Public Housing Low 1 .365 1 .333 1 .533 1 .277 1 .518 1 .639 High 1.950 -.585 1 .671 -.338 Urban Renewal Low 1.811 1.710 1 .924 1 .580 1 .523 1 .742 "Iigh 1 .466 .345 1 .325 .385 101 Table 4-14--Continued Social Issue Status Springfield EUgene ._ Low - High Low - High Metropolitan Park Low 1 .258 .938 1 .194 1 .297 I 1 .171 1 .339 High 1.216 .042 1.055 -.117 Public Kindergartens Low 2 . 075 2 . 1 15 2.080 1 .883 2 .037 1 .758 High 1.758 .317 1.978 .137 Summary of the Findings and Conclusions Table 4-15 presents a summary of the analytical findings in this chapter that will be utilized in later chapters. It should be noted that, for some variables, I felt it necessary to make relatively fine distinctions, such as for the community's attitude toward the issues where I distinguished degrees of favorable attitudes. But in others, such as the relationship between increased involvement and favorable attitudes, little distinction as to the strength of the relationship was made. There are two reasons for this variability: the crudeness of some of the instruments and the analytical utilization of the data. The measurement of community involvement was imprecise, thus the "Not involved- Involved" dichotomy. Favorable attitudes change with differences in involve- ment and social status are to be used in evaluating the direction of leadership bias. Thus, the direction of the relationship and not its strength is necessary. This chapter has been presented to familiarize the reader with the two communities and the importance of these issues to those communities. It was Table 4-15: Summary of discovered findings and relationships. Issue Attraction of Industry Springfield Eugene Annexation Springfield Eugene Parking Lots Springfield EUgene Special Education Springfield Eugene Fluoridation Springfield Eugene Public Housing Springfield Eugene Urban Renewal Springfield Eugene Metropolitan Park Springfield Eugene Public Kindergartens Springfield Eugene Community's Attitude Towa rd, Issue Most Favorable Most Favorable Favorable Favorable Favorable Favorable Very Favorable Very Favorable Less Favorable Less Favorable Favorable Favorable Less Favorable Less Favorable Favorable Favorable Opposed Opposed Community's Involvement in Issue Not Involved Involved Invo lved Involved Not Involved Involved Not Involved Invo lved Invo lved Invo lved Not involved Not Involved Involved Not Involved Not Involved Not Involved Not Involved Not Involved 103 Table 4-15--Continued Relationships with Increasing Involvement. Relationship between Increasing Social Status. Issue Favorable Attitudes Favorability Polarization Attraction of Industry Springfield Increased Consensual Increased EUgene Increased Consensual Increased Annexation Springfield Increased Moderate Increased EUgene Increased Low Increased Parking Lots Springfield Increased Low No Relationship EUgene Curvilinear Moderate No Relationship Special Education Springfield Increased ‘ Low Increased Eugene Increased Moderate Increased Fluoridation Springfield Increased High Increased Eugene Increased High Increased Public Housing Springfield Decreased Low Decreased Eugene Decreased Low Decreased Urban Renewal Springfield Increased Moderate Increased Eugene Curvilinear Moderate Increased Metropolitan Park Springfield No Relation- Low No Relationship ship Eugene Increased Low No Relationship Public Kindergartens Springfield Increased I ligh Increased Eu gene Increased High Increased 104 also presented to deal with various relationships which will be of use later in this study . CHAPTER V LEADERS AND FOLLOWERS The variable of leadership policy preference representativeness is of primary concern in this study. Hypotheses #2 through #11 explore conditions under which leadership might be expected to be more representative. The final hypothesis deals with leadership representativeness as an independent variable which has certain effects on the community. All but this last hypothesis will be evaluated in this chapter. The chapter is ordered by evaluating each hypothesis in sequence. Leaders and Followers Certain discrepancies between the policy preferences of community leaders and of the public should be expected because of measurement errors and chance. What is of concern here, however, are systematic biases within community leadership. Table 5-1- shows that there is unrepresentativeness within the two communities, varying from very slight in EUgene on the parking lot issue to very substantial in Springfield on the fluoridation issue. Furthermore, the bias is systematic because it is consistently in favor of the issues. Fifteen of the eighteen instances where the direction of the bias can be evaluated, nine issues in two communities, show leaders to be more in favor of the issues in question than the followers. Two of the three instances in which leaders are less favorable in relation to the public are the public housing issue in the two 105 106 Table 5-1: Comparison of leaders and followers in the two communities based on their means. Springfield EUgene Issues Leaders Followers Difference* Leaders Followers Difference* Attraction of .181 .506 .325 .287 .55)‘ .272 Industry Annexation .833 l .392 .559 .424 1 .261 .837 Parking Lots .863 1 .418 .555 1 .450 l .453 .003 Special Educa- 1.068 1.191 .123 1.139 1.161 .022 tion Fluoridation .761 1.735 .974 1.138 1.819 .681 Public Housing 2 .022 1 .558 -.464 2 .028 1.573 -.455 Urban Renewal .977 1 .779 .802 1 .068 1 .596 .528 Metropolitan .651 1 .253 .602 .611 1 .250 .639 Park Public 1.818 2.060 .242 2.197 2.008 -.189 Kindergartens *Difference is equal to the followers' mean less that of the leaders. Note the sign of the difference. communities. I have previously noted the consistently reversed patterns for this issue. Thus, finding leaders less in favor of this issue supports the conclusion of a systematic bias. In contrast to the striking similarity of attitudes toward these issues in the two communities leaders of the two communities do vary in the distributions of their attitudes. In seven of the nine issues, Eugene leaders are less favorable than their Springfield counterparts. Given the consistent bias of the leaders, this relatively unfavorable attitude of the Eugene leaders might be expected to make them the more representative of the two community leadership groups. And 107 in Seven of the nine issues, Eugene leaders are more representative. This is indicated by the smaller differences between their means and those of the Eugene public, as compared with the differences between the means of the Springfield leaders and followers. The two issues in which the Eugene leaders are not the most representative, annexation and the metropolitan park issue, are also the issues which they favor more than Springfield leaders do . Thus, in all cases the more unfavorable of the two community leadership groups is also the more representative of its community. This strong relationship between favorable attitudes and representativeness among leadership groups will be found consistently in the later tables . It is a phenomenon of the overall bias towards favorable attitudes of leadership in these communities. Table 5-2 shows leaders in the two communities vary similarly in their Table 5-2: Rank order ofdifferences in the means of leaders and followers in the two communities from smallest to largest. Issues Springfield EUgene Special Education Public Kindergartens Attraction of Industry Public Housing Parking Lots Annexation Metropolitan Park Urban Renewal Fluoridation ‘OCDNOthN—J CDO~\I~O-'UI-§OON Spearman r = +.7l7 representativeness in the various issues. There are two exceptions to this similarity: the parking lot issue in which Eugene leaders are by far the more 108 representative and the annexation issue in which Springfield leaders are the more representative. But in general, whatever are the dynamics that explain variations in the representativeness of leaders in different issues, they do not differ greatly between these two communities. I have previously stated that the spread of opinions within the leadership stratum must also be considered when evaluating its representativeness. If a leadership stratum were greatly divided so that even I'I'tOUgI't the maiority of these leaders was not in agreement with the public on a particular issue, it is possible to conceive of means by which the more representative minority might achieve a policy decision closely in line with the public's attitudes. Such a decision might not be expected if the bias of leadership were evaluated entirely by the use of distribution means. The minority could achieve this policy decision by means of the instruments of democracy or by appealing to the conscience of the maiority in the name of the public . Unrepresentativeness without such division is more likely to result in decisions less in line with the public's will. Table 5-3 shows that leaders are in greater agreement than the communities on fourteen of the eighteen instances available for'evaluation. This is especially true of Springfield leaders, who are more divided than the community only on the public housing issue. It is apparent that Hypothesis # 1 must be accepted. Leaders in both communities are not only more consistently in favor of the various issues, but are also in greater agreement in their attitudes than are the communities. This is as true of one community as it is of the other. A possible distinction might be 109 Table 5-3: Comparison of leaders and followers in the two communities by standard deviations of their attitude distributions. Springfield Eugene Issues. Leaders Followers Difference* Leaders Followers Difference* Attraction of . 385 .625 .240 .559 .701 .142 Industry Annexation . 720 . 944 . 224 .791 . 909 .1 18 Parking Lots .725 .884 .159 1 .275 1 .037 -.238 Special Educa- .862 .925 .063 .821 1 .027 .206 tion Fluoridation .923 1 .319 .396 1 .251 1 .525 .274 Public Housing 1.322 .919 -.403 1.194 1.016 -.178 Urban Renewal 1 .195 1 .244 .049 1 .076 1 .032 -.044 Metropolitan .773 .869 .096 .736 .970 .234 Park Public 1.092 1.198 .106 1.095 1.310 .215 Kindergartens *Difference is equal to the followers' means less that of the leaders. Note the sign of the difference. that Springfield leaders are slightly more in agreement. Although differing from each other in many ways, leaders in the two communities are more alike than their communities. They are also alike in the degree of sharing of attitudes between themselves and those communities. Vulnerable and Less Vulnerable Leaders Would the leadership of the communities reflect the policy preferences of the community better if a larger percentage of the leaders were chosen by the publics of these communities, assuming that the public's choice were not 110 sor.2.-..-how subverted? In my first chapter, I critiqued the various theories and research dealing with the way in which the public supposedly controlled public officials. The material would lead many theorists to say that the answer to the above question is yes. Given the questionable effectiveness of the means of permitting public control of leaders, I have questioned the public officials' capability of better serving the public's will. Capability is measured by the policy preference representativeness of leaders. 1f public officials and other leaders in the community who are vulnerable to p0pular control can be shown as not more representative of the public than the less vulnerable leaders in the community, it would mean that they would be greatly limited in their capability of better serving the public's will. Table 5-4 shows the means of the vulnerable members of the communities' leadership and those of the less vulnerable leaders. If the various methods of popular control are viable, the vulnerable leaders should be more representative . The two communities vary greatly in the percentage of vulnerable leaders. Only 7 per cent of Springfield's leaders are vulnerable as compared with 29 per cent of Eugene's. l have previously noted that Eugene leaders are more representative than Springfield leaders in seven of the nine issues. This might be taken as evidence that a large percentage of vulnerable leaders among a community's leadership will make leadership more capable of serving the public's will. But an issue by issue comparison of vulnerable and less vulnerable leaders in both communities does not support this conclusion. In both communities vulnerable leaders are less representative than less vulnerable leaders in six of the nine issues; 111 Table 5-4: Comparison of vulnerable and less vulnerable leaders in the two communities on the basis of the differences between their means and those of the communities. Springfield Differences from the Public" Less Less Vulnerable Vulnerable Vulnerable Vulnerable Issues Leaders Leaders Community Leaders Leaders N = 3 N = 41 N = 448 Attraction of Industry .000 .195 .506 .506 .311 Annexation .333 .871 1 .392 1 .059 .521 Parking Lots 1 .000 .853 1 .418 .418 .565 Special Education 1.000 1.073 1 .191 .191 .118 Fluoridation .667 .769 1 .735 1 .068 .966 Public Housing 2.000 . 2.024 1.558 - .442 - .466 Urban Renewal .333 1 .024 1 .779 1 .446 .755 Metropolitan Park .333 .675 1 .253 .920 .578 Public Kindergartens 2 .333 1 .780 2 .060 - .273 .280 Eugene Differences from the Public" Less Less Vulnerable Vulnerable Vulnerable Vulnerable Issues Leaders Leaders Community Leaders Leaders N = 21 N = 52 N = 526 Attraction of Industry .238 .307 .559 .321 .252 Annexation .238 .500 1 .261 1 .023 .761 Parking Lots 1 .500 1 .431 1 .453 - .047 .022 Special Education 1.000 1.192 1.161 .161 - .031 Fluoridation .937 1 .204 1 .819 .882 .615 Public Housing 1 .736 2.137 1 .573 - .163 - .564 Urban Renewal .571 1 .269 1 .596 1 .025 .327 Metropolitan Park .667 .588 1 .250 .583 .662 Public Kindergartens 2.100 2 .235 2 .008 - .092 - .227 *Difference is equal to the community's. mean less that of the type of leader, 112 Although differentiating between vulnerable and less vulnerable leaders does yield variations in policy preferences between them, the differentiation is sufficiently great in only three issues to make the two leadership types inclusive of the public . There are differences between the policy preferences of vulnerable and less vulnerable leaders, but the strength of this vulnerable-less-vulnerable variable is not nearly as great as the leader-follower variable. Both vulnerable and less vulnerable leaders show the favorable bias of leadership. There is no reason to accept Hypothesis #2 . Vulnerable leaders are not more representative of the public than are the less vulnerable leaders. Nor is it true that less vulnerable leaders are more representative for this is also not strongly supported by the data. The distinction does not help us understand what leaders represent the public most adequately. Potential and Actual Leaders At several points in this analysis, I have commented on the strong bias of leadership toward favorable attitudes in the policy issues investigated in this study. Leadership in both communities is a biased selection of the policy preferences of the communities; But is the bias of leadership a reflection of the occupation and political backgrounds of the men chosen as leaders in the community? Or is the selection process one which only allows men holding certain policy preferences, regardless of their occupation, to become leaders? To evaluate this question an additional stratum of activists called "potential leaders" was identified and interviewed. As 1 have said, these men have occupational and political backgrounds commonly found among community 113 leaders, but lack the reputation for being influential. Community leaders will be called "actual leaders" during the analysis of this hypothesis in order to distinguish them from the less influential, "potential leaders." Table 5-5 shows that the actual leaders re5ponded more favorably to the nine issues than potential leaders in 38 of the 54 responses measured . A biasing process is operating in these two communities to assure that men who respond favorably to the issues are leaders. Furthermore, the process seems prominent among labor leaders (eight of the nine issues) and Democratic party activists (six of the nine issues and one tie). The type of leadership background least affected by the process is candidates for office. In this case the actual leaders are more in favor of only five of the nine issues. The process is most conspicuous in the annexation and the metrpolitan park issues where all six of the selected leadership background types are affected. But in the special education issue only two of the leadership types are so affected . There seems to be little distinction between these issues. The metropolitan park issue and special education issue are both non-involving, esoteric, middle-class valued, prOQrams. The weakness of this biasing process among candidates is caused by the exceptionally favorable attitudes toward the issues of those candidates who are potential leaders. Candidates who are actual leaders are very close to being typical of all actual leaders in their favorable attitudes toward the issues. But candidates who are potential leaders are not typical of all potential leaders, thus the biasing process is not as strong. Theoretically, a subgroup of leaders would be typical of all leaders if equal numbers of leaders were more and less favorable toward the issues than was this subgroup. Of the 45 comparisons 114 Java! _oBuo..i< vac coco! _a:coo0n_l.n_ 8.6 30233530 9.92 mo mEcaoE 9t... \bo. _ mmm. _ #3.. :7 auZ .< oum.m mmm. _ n00. _ o_N.~ «.1.— oom._ So.— ooo._ mmm. EHZ . o a BoSfltcoU 30:20; mafim Km. _ 35. now.— omm. m3. _ ooo. _ ooo.m 03; n3. moo.— coo.— mnm; Km. _ mmm. mmm. vow. N3. DnZ nuZ O< Co. 350 co”. moeovmvcou ooo.m oom. ooo. ooo. _ ooo. _ oom. oom. ooo. ooo. «HZ .< . EoU www.— 08. moo.— 000.. 02.— 0;. mom.— noc . mmm . NW2 0 a \ . dErOU 0:969:00 ooo.N mmm. muZ .< mmv. N mum.— mmfim 30. N 03. .n. .EoU.QEoU cou:naaom moo . mmm . moo . mmm. mmm.m mmm . mmm . ooo. mmm. _ .mwv. N com; QB.— No... ono.~ vow; emu.— nom . _ onm . KN" Z .m cacao; coma.— oom.m obim acotamcomfix 0:33; $00. was.— xcom co.:_on_oto<< 8: on: 3323. :35 Sea .5.“ 3301233. 08.. o8.“ 8:02.03". R: 8a... 82833202; 8o.~ EX 20.. 9:10.. m3. 8: 5:285. 8.. me. $3220 SEER :uz NNuz .< to moan. ficafiEon— umEocoum . 8:588 0.6 £252 @503 pca 3335ch $533235 opazto :9: mo «:02: .3 «our: pcaohoxuao $5.532 333mm :23; £252 3:233 pea .0200 m0 camcodEoU "min mick. O 115 possible in Table 5-5, candidates who are actual leaders respond less favorably than actual leaders of other backgrounds in 20 of the comparisons. They are quite close to being typical of all actual leaders. In contrast, candidates who are potential leaders are less favorable than other potential leaders in only ll of the 45 comparisons. They are more favorable than most potential leaders. One might hypothesize that the potential leader sees conformity to the policy preferences of the existing leadership as desirable if he also sees candidacy for public office as a means of achieving leadership. This. biasing process shows potential leaders as being more representative in 37 of the 54 comparisons measured (see Table 5-6). Labor leaders and members of the Democratic campaign committee are most affected by the process. In 8 of the 9 issues, the members of the community leadership with these backgrounds are less representa- tive than those who are not actual. leaders. In contrast, actual leader candidates are more representative than potential lemler candidates in 6 of the 9 issues. In regard to the signs of the differences between means in Table 5-6, a negative sign means that the leaders of this type are less favorable than the public on the particular issue. The difference is the mean of the public less the mean of the leaders. The lower values indicate more favorable attitudes. Potential leaders are less favorable than the public in 24 of the 54 measurements, While actual leaders are less favorable in only 15 of the 54 measurements. Potential leaders are not a systematically biased selection of the public, but actual leaders are. The biasing process appears to be a single step process with potential leaders not being a systematically biased selection of the public, at least in terms of their policy preferences. But actual leaders are a biased llé 3... I AVon. I woo. $0. I ovm. mo.. 0R. I v.0. I mm... mNo. on... I mom. I omo. .mo. I 2.0.. do... 0.... mm... ouZ o.uZ .< .n. noSntEoU _ao..._on_ mo.. I 03.. vow. to. I ova. omo. voo. I #0... I .oo. oo... om.. .0... 53. .3. I #3.. am... Rum. omma DaZ nnZ .< .n. oath—O can. ma.ap.p:aU owe. I mm... won. «me. NR. owe. com. new. men. .90. .00. wk. «we. moo. 03... mom. .3. N8. I ng N.nZ .4. .m .EoU .oEaU 0:0: 00:60 omo. wow. I mmo. no.. I no... mmm. I #0... I 08.? non. km... N... I 3.. 53.7.. omh. I mom. .mN. I mm.. .mm. muZ muZ .< J: .EoU .oEaU :aa._n.oaom j 0 .U n. moo. no... mom. com. I wow. 08.. m3. mm.. muZ .4. moo. cm... 3... e .,..... I ma.. I 0N7 N2. .2. mmuZ .m mom—ace.— Boa.— own. I em... I own. 3.. I 3.. I mmo. m.m..I omm. I new. mmm. I mom. I 30. I ooo.... 5.0.7 3.0. mum. o3. mom. ..uZ Nng .< .n. 2:05:50 u.Eo:oom acetaocavc. v. a. 33m {an :a: 319.02 730:3. :045 @5301 0.33m 52%...an cotauapm _a.uaom :3 9.25.. :o..axe::< Eton:— .o 5209.24. 833. £0.32. 3:5an or: .o amor: 60:. ”came to... .0 30:93.5 9+. .3 moo... vcaocoxuan mEfiavafl 3.8.9. 5.3 Eavafl 32:29. pca 328 .o :omtadEoU "elm £32. 117 selection from among the potential leaders. Other evidence supporting this conclusion will be given later. It should be noted that this conclusion is limited only to these six types of leadership background . The universe of potential leaders is not measured, and the six types used are not necessarily typical of all potential leaders, although they are the common types of leadership background in most communities. There is a tendency for potential leaders to be more divided in their policy preferences than the community; In Table 5-7, looking only at the sign of their differences from the combined communities, potential leaders are in Table 5-7: Comparison of actual and potential leaders with the combined communities based on standard deviations. Potential Actual leferences from LeadeIs Leaders Communities Communities Potential Actual N=92* N=37* N=1226 Leaders Leaders Attraction of Industry .619 .587 .667 - .048 - .070 Annexation 1.047 .679 .932 .115 - .253 Parking Lots 1 .174 1 .279 .967 .207 .312 Special Education 1.076 .861 .980 .096 - .119 Fluoridation 1 .418 1 .145 1 .433 - .015 - .288 Public Housing 1 .298 1 .294 .971 .327 .323 Urban Renewal 1 .257 1 . 149 1 .149 .108 .000 Metropolitan Park .998 .765 .925 .073 - .160 Public Kindergartens 1 .174 1 .196 1 .258 - .084 - .062 *Because the selected leadership background types used ifiables 5-5 and 5-6 are not mutually exclusive, the combined N for the background types will exceed the N's given in this table. 118 greater agreement than the public in only three of the nine issues. This is indicated by the negative sign on differences between standard deviations. On the other hand, actual leaders are in greater agreement than is the public on six of the nine issues and more consensual than the potential leaders on seven of the nine issues. ,Looking at the columns of differences in standard deviations between the combined communities and the two classes of leadersgit is evident that potential leaders are generally more divided than the public, while actual leaders are less divided than the public. Potential leaders are more representative of the public in terms of standard deviations on six of the nine issues. This is additional evidence that the bias of community leadership is achieved in the selection of actual leaders from among potential leaders, and not in the inherent bias of the men likely to be community leaders. Hypothesis #3 is accepted. Potential leaders are more representative of the public . The Alternative Policy Preferences of Community Leadership l have previously defined available leadership as the combination of actual and potential leadership. One must be very careful in analyzing this data . It is neither a universe nor necessarily a typical sample of available leadership in the community. Thus, its characteristics as a whole cannot be generalized to the universe of available leadership in the community. The data can be used, however, to note which of the various leadership background characteristics included define the extremes of policy preferences in the community and what differences there 119 are between the included types of leaders. I noted in the discussion of the previous hypothesis that potential leadership is more divided than actual leader- ship. Which of the leadership types studied defines the extremes ofpolicy preferences within available community leadership? Certain of the leadership backgrounds studied are consistent in the relationship of their mean on a given issue with those of the other leadership types in Tables 5-8 and 5-9. Economic dominants, with the exception of the attraction of new industry issue, are consistently not in favor of the issues. Candidates for office and political contributors also show consistency, with the candidates for office being consistently more moderate in their approval of the issues than the other leadership background types. Political contributors are consistently moderately opposed to the issues. Leaders having other backgrounds vary greatly from issue to issue . No group can be clearly identified as the favorable counterpart of the consistently opposing economic dominants. Economic dominants seem to fulfill at the local level the conservative role so frequently attributed to them at the state and national level. Yet, four of the leadership background types show more approval on the issues than labor leaders. Labor leaders and economic dominants, on the average, are inclusive of the public, but among available community leadership, only economic dominants are generally at one extreme in their attitudes on the issues. Hypothesis #4 is rejected . Candidates usually have the most favorable attitudes towards the issues, but the members of the Democratic campaign committee are very nearly as approving. The members of the Republican campaign committee, however, are 120 SAN no... Nov. mo... mo.r.N mNN. 3m. omuZ 9.0.000.— coma; 30.. m3. mNa. ooo.. mm... 8... v.N.. Km. 3n. IHZ .EoU .aEaU 0:209:00 mmmN mmm. mmm. mmmN mmm. mmm. 03.. V3. mmm. ouZ 0_n.20:_o> 3:022...qu ooo.. 0.0. mon. RN... ooo.. 96.. XX... omN. omN . vNuZ 8E0 .8 25228 oom. N m.... m.m.. onNN om... mmn. mam. .Nm. amN. mNnZ _aa...._on_ mom . ..uZ :o.on..._..:oU . EoU .aEaU :au._n_aa0~_ moo. w.N. mmuZ .ca:.Eoo 0.Eo:oum m:0tao._0p:.v. 0.33. {an— :a.._oao._.0.2 _a>>0:0~. cant: @530: 0.33. 8.32.02". :o..aoopm .a.u0¢m 30.. 926.. :o..axa::< .33 a:_ .o cotuactd. «03m. 6:023:32”. 03.23 :05. ...o m:a0E 0r: :o .0033 “02.5....an foo 20:. «0:3200 mom...— vcooaxoan 3:30.80. .0320... .o :Om.:aanU “elm 030.. ~Ol\ V 121 [\C') mN n o “0:225:00 9.0.000.— soon; 0") m . EoU . oEaU 0:050:00 n 0_oa._0:_:> _0:o...0ooouO (0‘1” ('0 CON m 850 3. 202.250 lO‘O l0“) ION v _OU_.:_On_ II— [\C') 0000 0000 N 3.3..zcoU .EoU .oEaU .:0:.Eoo :00. 32.0”. 0.Eo:oum 0.:3. 0oa..0>< :0t090a:.v_ 0.33.. x:0m :0: _ooo._.0<< 330:0”. :35 m:.30: 0.33m 5:02:03”. cotaooam _0.00om as 9.36.. :o..aX0::< Etna:— .o :o..ua._:< 8:3. . m:.>o._ao_a :00— o. m:.>o.&o0 .38 E0... 6:00.: :0... .0 .303 4:0: 0.... .3 81.... 0:30.6qu a.e_m._0pa0_ 00.020”... .0 :omtamsou "oIm 030.. 122 not the leaders least approving of the issues. If approval and disapproval of these issues are the local counterpart in political attitudes to the liberal- conservative dimension used in the analysis of nationwide political attitudes, the two parties do take the expected positions with respect to each other. The Democrats are more approving than the Republicans in six of the nine issues. The three issues which the Republicans favor more, attraction of industry, annexation, and fluoridation, show no particular pattem relative to the other issues. The issue on which there is the greatest discrepancy is the issue of attracting industry to the community. The Republicans are most in favor of it and the Democrats least in favor. This fact is consistent with the philosophy normally attributed to the parties. Thus, this issue should probably be considered a reversed issue, one which conservatives would be expected to support. The reversal on this issue, therefore, supports the discovered relationship. AlthOUgh the two parties orient themselves as one would predict from knowledge of the national studies, they do not define the extremes of available leadership opinion in the community. What is more, they both approve of the issues more than the public. Hypothesis #5, as stated, must be rejected . It is apparent from Table 5-9 that, although actual leadership in the community is consistently more in favor of the issues than the public and seldom inclusive of that public in its policy preferences, available leadership in the community does give a sufficient span of opinion on the issues to allow a representative leadership to be chosen. Attempts at overcoming or altering the recruitment of leaders from among available leadership do have a chance of producing a community leadership more capable of serving the public's will than 1’23 does the present community leadership. Table 5-10 shows the ranking of each of the leadership background types on each of the issues by the magnitude of the difference between their mean and that of the public. None of the categories of leaders show any consistent pattern of being more or less representative of the public . On the average, however, it is not candidates for office or the present holders of public office who are most representative, but the labor leaders. It is notable that labor leaders were previously discovered to be the leadership type most susceptible to the biasing process in the selection of actual leaders from among potential leaders. Neither one of the political parties seems, to offer the public a better chance for a more representative leadership than does the other. The best conclusion that can be derived from Table5-10 is that none of the leadership groups provide a set of attitudes toward these issues such that a community leadership composed of that leadership type would greatly improve the capability of community leadership to serve the public's will. Also, no leadership type, including the economic dominants, seems much less capable of serving the public's will. Just in terms of this static comparison of attitudes toward the present issues facing these communities, a discovery of an economic and business elite in a community need not mean that the public's will is not served. It is not the background of the community leaders that makes leader- ship unrepresentative, but it is a bias in the attitudes of men as leaders. Whom do the Leaders Represent? Thus far, I have considered two distinctions among leaders and the 124 N.N m... m... N... a... e... m... V.53. «952.. o. m m . o N N 2.590.050. 0.3.... m e N m N e . vE... 5.3335. 0 m n b m . N 330:0”. :35 N m m _ o. o N 9.30.. 02...... N v o m m N . 8.8.2.8.. N N e m m . o 5:833 .0003 m m e . N e N to. 9.0.3.. . 4 m N m e N 8:225 _ N m n v N o 93.0... .o 8.02.... 8:3. c0000.. .EoU .oEaU 030.0:_:> 00...O ._o. .0.:N.....:0U .EoU .oEaU 205800 .630... 0:000:00 .0:0..0a:000 0.00.0:0U .00....0. :00._..:n.0~. 0.50:00m .8...::EE00 00:3E00 0;. .0 300.: 0... 80.. «00:90.30 .30.... .0 .020 v_:a._ 0.... .3 med... 05.89.02. 3...»...0000. 0000.0» .0 contaoEoU 6.1m 030... l25 effects of these distinctions on leadership representativeness. The general question of interest in making these distinctions is what variations exist within community leadership which affect their policy preferences and which, in turn, affect their representativeness. In this section of the chapter, I continue this inquiry by considering whether the unrepresentativeness of leadership is a consistent bias towards the policy preferences of those with higher social status in the community or those most active in the political process in the community. My purpose is twofold . First, I wish to question whether there is a class or political stratum position on the issues, and if the leaders with their higher social status and greater political activity reflect this position? If so, presumably the representativeness of community leadership might be improved by making community leaders more like the community in social status and political activity. The latter correction might be difficult because we rely, as a culture, on voluntary candidacy for public office. The second purpose of this analysis is merely to give a fuller understanding of the dynamics of leadership selection in the community. Bias Towards the Policy Preferences of Those with Higher Social Status I have shown in Chapter IV that there is a strong and continuous relation- ship between increasing social status and approval of the issues in six of the nine issues. I have also shown that a reversed relationship exists in the public housing issue. If leadership or any of its subgroups were biased toward the policy preferences of those with higher social status, the leaders' means on the dis- tributions of attitudes should be closer to those of the high social status levels 126 and further from the means of lower social status levels. Community Leaders. Springfield leaders clearly show the existence of such a bias in Table 5-ll . Leaders in Springfield most accurately reflect the policy preferences of the highest social status level (3) and least accurately reflect those of the lowest social status level (0). Leaders in Eugene show a relationship in the same direction. They accurately reflect the policy preferences of the higher social status levels and inaccurately reflect those of a lower level in six of the nine issues. But the relationship is not so perfect as in Springfield. Hypothesis lll6 is strongly supported in Springfield and supported in Eugene. Vulnerable and Less Vulnerable Leaders. What is most conspicuous in Table 5-12 is the lack of differences between vulnerable and less vulnerable leaders as to which social status level is most represented . In each community both leadership types best represent and least represent the same social status levels in seven of the nine issues. Second, in the four issues where there are differences, two issues show that vulnerable leaders better reflect the policy preferences of a lower social status level than the less vulnerable. leaders; and two issues show the reverse relationship. Vulnerable leaders are no less affected by the bias towards the policy preferences of the higher social status level than are less vulnerable leaders. Hypothesis #7 is rejected. Actual and Potential Leaders. Table 5-13 gives only the level of social status best represented by the leaders of various backgrounds. In 23 of the 54 comparisons between actual and potential leaders given in the table, the two types of leaders differ as to which level of social status they best represent. In I27 Table 5-I I: Comparison of the direction and degree of social status bias among Springfield and Eugene leaders. Springfield Social Status Social Status Level* Most Level* Least Leaders' Represented Mean Represented Mean Issue Mean Attraction of . IBI 3 .274 0 .575 Industry Annexation .833 3 I . I35 0 I .508 Parking Lots .863 2 I .273 0 I .532 Special Education I .068 2 l .037 I l .338 Fluoridation .76I 3 I .327 O 2 . OI 5 Public Housing 2.022 3 1.950 0 I .365 Urban Renewal .977 3 I .466 I I . 924 Metr0politan Park .65I 2 I . WI 0 I .258 Public Kindergartens I .8I8 3 I .758 I 2.080 Eugene . Attraction of .287 l .506 0 ' .574 Industry Annexation .424 3 I .093 0 I .54I Parking Lots I .450 0 I .437 I I .486 Special Education I. I39 2 I . I37 0 I .327 Fluoridation I. I38 3 I. I44 I 2 . I20 Public Housing 2.028 3 I .671 I I .277 Urban Renewal I .068 3 I .325 2 I .742 Metropolitan Park .6” 0 .938 2 I .339 Public Kindergartens 2.I97 0 2.I15 2 I .758 *Social Status is measured on a scale varying from a low value of 0 to a high value of 3. See Appendix C. I) omo. N mnN. 0N0.— mom.— m.o.N mom . . mum . c0000.. .:_:> 80.. 820:5 3.2m .0.3m 02:08.50... .30.. .0 :002 0:0 «_0>0.. ._OO 00 EN. . mmN.. 0N0.— mom.— m.o.N www.— Nmm. . mom . — mum . £0000. .:_:> .— OO 00 wmn. . K... 000.. one.— NNmJ nmo. . mNN. . mm... FmN. £0000. . :_:> 30.. 522:0. 3.2m .28.". 03:03:03. .022 .0 :00<< 0:0 s..0>0.. NN mm ”N (000 owoN K... 00.... one. NNm. mom . mNN . . mm... EN. 00000.. . :_ a> 0.0...m:..=.m COCO N NCO C") own... mno. 0N0. . .VNo . N mom . mno . . mmw . Kw. m2. .vuZ :002 .B0000.. .C_D> 3.. .005 «2.20 .200» :0... 00 00.500 0:0 car—U0... . 0 0 .. :0 E0 .. 000. 030.059 30. 0:0 0380: 2.0 “N.-. .3... :00.2 10000.. .:.:> m:0t0900:. v. 0. .03. v...0n. :0.._on.o:.0<< 330:3. :00... 0£30... 0.33. 5:00.30. ... :o..00:0m .0.00Qm .20.. 8.0.3.. :o..0x0::< .0050:— .0 :o...00:.:< 039.. .0 50:09.0. 00m .m .0 0:.0> gm... 0 o. o .0 00.9. 30. 0 So... m:.\C0> 0.000 a :0 00:00.00E m. 3.2m .2009. www.— 0mm.— N.R.— RN. 0N..N .No.. 000.. I29 Sun. in. 00000.. .:.:> 90.. 09m.— 00m. N.R.— nmN. 0N..N NNm. . .mw. . .vm. . Km. 00000.. . :.:> 820:... 3.2m .0.0om 002000.03. .80.. .0 :002 0:0 «.90.. N 000 CO m...N wma. mNm. K0.— 3.... mm... 50... moo.— eon. E0000. . :.:> 000.. CON mm (‘00 '— m...N mma. mNm. . .50.. .4..— .No.. 03%. moo. _ com . c0000.. . :_ 0> 520:3 3.2m .0.00m 003000.00”. 00.2 .o :005. 0:0 0.0.6.. 0:0m0m .-- COCO COO _ mmNN mmm . ooN . . mm..N VON. No..— .9... 000 . mom . NmuZ :002 .833 . :_:> 0.00.. oo_.N N00. Km. 2K . . mma . ooo.. com.— me. mmNe .NuZ :002 000000.. .:_:> 20:09.05. u.32.. v00. cat—00003.2 .030:0~. cant: 0.5030... 0.3:. 5:00.002". :020000m. 3.00am 20.. 9.0.5.. :0..0X0::< b.0505 .0 :0...00.f< 020. 82:80:30 .30... £0000. .0:.00rr< 0:0 .0000. .0..:0.0..in. "0:0 202030.500 000.... .0 m:.:00:. 0...... COCO COCO COCO '— 00 (O0 m m m m . < . ... 0.0.00.5:0U .00... .on. COCO (ON m m .< COCO '- COCO m .m .n. 8.50 a. 382.88 00.58.00. .00m .0>0.. 3.0... .0.00m COO COCO 000 CO m .4. COO COCO OCO (O o .x. . EoU . dEau 02200800 COO CO COCO (O m .< COCO 0— COCO o m .. . EoU . QEau :00._n.0n.0~. OO COCO OCO m m .< P'— m .. 0.0000. .0004 COCO OCO (‘00 m m .< COO m m e... 0.:05500 0.Eo:oom 0:0...090 05 v. 0 . 3:. 000.. :0...0..0..0<< 330:0”. :00: @530... 0.3:. :0..00..0:_... :0..00:0m .0.000m 20.. 9.0.5.. :0..0X0::< 0.3.2. .o 5.02.2 02...». .3... 3.0.0 .500... 50... 0. 5.3.0.. 5 80000. .0..:0.0... 0:0 .0200 .0 0:055an0 6...... 030... l3] 15 of the 23 instances where there are differences the potential leaders are more representative of lower social status levels than the actual leaders. Thus, there is good evidence that Hypothesis #8 is true; potential leaders are less representa- tive of those of higher social status than actual leaders. Furthermore, in 3 of the 8 comparisons in which actual leaders represent a lower social status level than potential leaders, the difference between the two social status levels best represented is one level. Only 3 of l5 comparisons showing potential leaders as representative of the lower level of the two leadership types are such small differences. The relationship is not as strong as I would like in order to accept the hypothesis, but the relationship is there. Hypothesis #8 is accepted . Conclusion . Although social class differences on the issues are apparent on 7 of the 9 issues (see Table 4-14), and leaders generally reflect the policy preferences of the higher social status levels, the distinction between vulnerable and less vulnerable leaders has little impact on this relationship. The bias of leadership toward the policy preferences of the higher social status levels cannot be greatly altered by increasing the percentage of vulnerable leaders among the leaders in the community. It is not possible to say that, by overcoming the biasing process in the selection from potential leaders, the over-representation of the higher social status preferences could be alleviated . Potential leaders are not sufficiently free of the bias towards higher social status preferences to allow such a conclusion. They are most representative of the two higher social status levels in 31 of the 54 measurements. But when contrasted with the actual leaders most representa- tive of the two higher social status levels in 38 of the 54 measurements, it is {I ll!‘ l..‘ I“: apparent that the class bias could be partially removed by making actual leaders more typical of potential leaders, that is, by overcoming the biasing process in the selection of leaders. Leadership Bias Toward the Policy Preferences of the Politically Active The relationship between increasing political activity and increasing approval of the various issues is not as consistently continuous as is the relation- ship between increasing social status and increasing approval of the issues. Thus, if one leadership type is found to be most representative of the most active political stratum (3) and another is most representative of the next lower political stratum (2), it need not mean that the two leadership types differ slightly in their policy preferences because stratum 3 may be most favorable on the issue and stratum 2 least favorable of the strata into which the public is divided. To the degree that such inconsistencies exist within the public, it is less relevant to speak of the leaders' bias'toward the policy preferences of the more active . In part, the hypotheses dealing with this bias are refuted by the lack of a truly political activist's position in some of the issues. But many of the issues do weakly show such a relationship. Community Leaders. Again Springfield leaders more clearly show a bias in the hypothesized direction than do Eugene leaders, but it is not as perfect as the relationship shown in Table 5-l l . In eight of the nine issues, Springfield leaders are most representative of a higher political stratum than that which they least represent. They are most representative of the most active stratum (3), however, on only four issues. I IllalTT Illilllll‘l It’ll T33 The relationship is even weaker for the Eugene leaders because they are more likely to 1915‘! represent the higher political stratum. Five of the nine issues. show this unexpected relationship. They most represent the highest political stratum on three issues and least represent it on four issues. Hypothesis #9 is accepted for Springfield and rejected for Eugene. Vulnerable and Less Vulnerable Leaders. The distinction between vulnerable and less vulnerable leaders in Table 5-l5 is not very different from the relationship in Table 5-l4. Vulnerable and less vulnerable leaders differ in only four of the comparisons as to the political stratum they best represent. Vulnerable leaders are more representative of a lower political stratum in three of these four comparisons. Hypothesis #10 is rejected because there is no indication that the variable is a meaningful distinction. Actual and Potential Leaders. l have found that the process of leader- ship selection results in a bias toward the policy preferences of those with higher social status. Table 5-16 does not show the existence of such a biasing process toward the preferences of the more politically active. in the nineteen comparisons where actual and potential leaders differ in the political stratum they best represent, ten show potential ieaders best representing a higher political stratum than the actual leaders. Hypothesis #li is rejected because no relationship is evident. Comments on the Issues of Public Housing and the Attraction of Industry in nearly every relationship considered in this chapter and in Chapter N, the public housing issue has shown a pattern opposite to that shown by the 134 Table 5-14: Comparison of the direction and degree of leadership bias toward the preferences of the political activists. Issue Mean Attraction of .181 industry Annexation .833 Parking Lots .863 Special Education 1.068 Fluoridation .761 Public Housing 2 .022 Urban Renewal .977 Metropolitan Park .651 Public Kindergartens 1.818 Attraction of .287 Industry Annexation . 424 Parking Lots 1 .450 Special Education l .139 Fluoridation l .138 Public Housing 2 .028 Urban Renewal 1 .068 Metropolitan Park .61 1 Public Kindergartens 2 . 197 Springfield Political Stratum* Most Represented Mean 3 2 (AN (:30 .400 . 193 . 339 . 073 .266 .843 . 466 . 066 .843 .380 .118 .450 .144 .606 .100 . 485 . 894 .040 Political Stratum* Least Represented O 0 ON CO 000 (do 000 Mean . 684 . 578 .516 .340 .853 .471 . 808 . 375 . 087 . 676 . 380 .411 .800 .550 . 338 . 888 1.330 .881 *Political Stratum is. measured on a scale varying from the lowest stratum of political activity with a value of O to the highest with a value of 3. See Appendix C . 1 35 nmo.m 0 new.. me.. o mnm.. mom.. . mom.. Kw. . o Kw. . 0mm. — 0 0mm. — ovm.. o ovm.. e.m.. N o.m.. mum.. o mun.. ewe. o ewe. E0000. c0000.. . c.:> 30.. .:_:> Ear—0.5% _UU—:_Om 8.58.53. .80.. .0 000.2 .000 e.0>o.. NO 00 ,__ O ”5.0— N NwoaN O Dako— 080 —. m 08. —. m WNOo 030—. m 030 — m V80 —. mmo.. . mko.. . mko.. 8... m 8.... m 8.. :flZ 9.0.000. 20000.. .33.... see; .8..._on_ .:_=> mop—congamm to: $0 33 000.2 000 10>?— 23.8.3 mmmd mmm. mmm. ooo.~ ooo.. ooo.. mmm . ooo . muZ C002 .38... .025 20590.85. 0.3.... .6. 5:30.232 .0305”. 02...: 9.330... 0.33. 5:01.33". 520033 3.023 .8. 9.33.. 00:98:06. 98.8. .o 8.02.2 029.. op: “Pogo. 35 :0... .0 0960.0 .vco 5:00....0 0.... co co 5.3.2.00 .00..—..On. mo moucmweou—OLQ .000. 030.0015 30. 000 030.55.; .0 comradesou 6.10 mice .U x.0:0an.< 00m .m .0 0:.0> 0 {.3 89...... of 0. o .0 0:.0> 0 r...) 222.00 .00....0a .0 20:... .830. 0....— Eo... mc.b0> £00... 0 co 00.300E m. E283 .00....0..e oneN m oneN m SON o SON o mmNN 8..N 8:88.80. 0.3.... 08.. o 08.. o 3... N a... m we... .8. .8. 85088.2 mam. m 8.... m me... o 3.... o $N.. .3. .288. 8e: mm... o 8.... o 8..N N N8. . Nm..N 8N. m8.28... 2.8.. on... N m om... N m 8%. . N 88. . N 3N. . N8. 8:88... n. m 8... m 8N. N SN. 0 8... . Ne... ooo.. 8:88.. .28.... m3. . E... o 2.... N we... . .9... 8... 8.. 85...... 0mm. . m 8... . m m. .. . N m. .. . N 8.... 8N. 8.888... 8%.. o .5. o mom. N 8... N B... 8N. 8.3.2. .o 8.82.2 Nng .NHZ e3em. ago—ooo..— EOVOGA 30300.— 9.0300..— .c_3> mum..— .c_3> .c_3> mum..— .c_:> c002 2002 .883 .88.... 53.—0....m .00220m EP—Ofim .00..._Om .:_D> .c_3> 12:08.13. .80.. .0 200.2 .80 e_e>0.. 02:03.03. .32 .0 30.. c002 .000 £30.. ocomom 82:80... 28 ...3... 137 ..0000. .0:.00l< 0:0 .0000. .0..:0.oaln. ”0.0 32.030300 000... .0 m:.:00:. 0...? o o m o o o m m N N o o . N N N m m . < . a. 80.30....cou .00... .0. o N m _ o o . m N N _ m . N N N m m . < . .. 000...O .0. A..0.00.0:0U 02:08.03. .m0m E383 .:0E0>_0>:_ .00... .0. .0..0:00 CO N m .< N OCO '— N o .n. . E0U . anU 02200800 CO NO N m .4. CO COO NCO o m .n. . EoU .QE0U :00. 3:00. C“) COO OCO N m COO -- F- m . < . n. 0.0000... .000. CO COCO NO N m .< COO COCO NO N O C e... 0.:0:.Eoo 0. Eocoom 0:0...0m..00:. v. 0. 3:. 000. :0.. _ 008.02 .0305. cant: @530: 0.30.. :0..00..0:_ “. :0..00...0m .0.00am no. 8.46.. :0..0x0::< 2:30:— .0 :0..00...< 00......— .«.0.n. .».>..00 .00..._0.....0.._. .0 :0..00._.0 0.... :0 E0000. .0..:0.0n. 0:0 .0200 .0 :00...0..EoU .076 0.00... l 2.8 other issues. Also, I have found that compared with Democratic committeemen Republican campaign committeemen are less in favor of all issues but the issue of attracting industry. Many of these issues, though local in terms of decision- making, are originated in national politics and have been found to be consistent with the liberal-conservative distinction at the national politics level. Thus, the political party officials are taking stands on these local issues consistent with the national positions of their parties. The issue of attracting industry is rather distinctly local and, as l have argued, is consistent with one's expectations of the Republican party and its association with American business. I have discussed these two issues at length because I think they indicate a possible expansion of the research design used in this study and give some insight into the decision-making process in the community. I think the issues used in this study are an over-representation of elite originated, middle-class valued, nationally defined issues. Future replications and elaborations of this study should include more issues, if possible, which are indigenous to the com- munity under study, valued by the lower social classes and opposed by the higher social classes, and identified as important by the public itself. The inclusion of such issues would allow greater insights into the dynamics of the leader-follower relationships investigated in this study. Another aspect one might want to include is that type of issues which would amount to, or demand, a change in leadership in the community or which would mean a basic change in the values of the American public. Issues which normally are not considered issues, such as appointments of officials or other "administrative" decisions, might also be included as part of this distinction between issues. l3$> I am not criticizing the decision to include only issues which were of concern in the two communities, but I am squesting additional criteria for the selection of issues to investigate . To include issues of these other types would greatly broaden the subject of interest. Choosing only issues which are of con- cern in the communities fails to include issues of these other asPects. This is evidence of the control of issue selection by community leadership as suggested by Bachrack and Baratz. Summary of Findings and Conclusions Hypothesis # l: Hypothesis #2: Hypothesis #3: Hypothesis #4: Hypothesis #5: Hypothesis #6: Hypothesis #7: Leaders are not representative of the community. Vulnerable leaders are more representa- tative than are less vulnerable leaders. Leaders are less representative than potential leaders. Labor union leaders and economic dominants define the extremes of leadership policy preferences. Democratic and Republican party leaders define the extremes of leadership policy preferences . Leaders are more representative of persons of higher social status. Less vulnerable leaders are more repre- sentative of persons of higher social status than are vulnerable leaders. ACCEPTED . REJECTED . ACCEPTED . REJECTED . REJECTED . ACCEPTED . REJECTED . lPeter Backrach and Morton S . Baratz, "Two Faces of Power, " The American Political Science Review, LVI, N0. 4 (December, 1962), pp. 947- 952; and Peter Bachrach and Morton S . Baratz, "Decisions and Nondecisions," The American PoliticalLScience Review, LVll, N0. 3 (September, 1963), pp. 632-642 . I40 Hypothesis #8: Leaders are more representative of ACCEPTED . persons of higher social status than are potential leaders. Hypothesis #9: Leaders are more representative of ACCEPTED the more politically active members FOR SPRINGFIELD . of the community . REJECTED FOR EUGENE . Hypothesis #10: Vulnerable leaders are more repre- REJECTED . sentative of the politically active ' members of the community than are the less vulnerable leaders. Hypothesis #1 l: Leaders are less representative of the REJECTED . more politically active members of the community than are the potential leaders. Four major findings are evident in the preceding analysis. Community leadership shows a consistent bias of policy preferences relative to those of the community. The direction of this bias is toward approval of the issues. The strength of this relationship is such that it is difficult to find subgroups within leadership that hold policy preferences less favorable or even as favorable as the public . The second major finding is that it is irrelevant to make a distinction between leaders who are subject to popular control, such as holders of public office or candidates for such offices, and leaders who are not as subject to such control. The three hypotheses that deal with this distinction are rejected, for the most part, because of the lack of demonstrated differences between the types of leaders so differentiated . A rewarding distinction among leaders is that made between actual leaders in a community political process and men who have backgrounds commonly found among such leaders but lack or have yet to achieve the influence of actual leaders. Elllrlll'Illlllllll-ITII‘EV I4] This latter group of men, called "potential leaders," does not show the consensual approval of the issues characteristic of the actual leaders. Rather, potential leaders are consistently less in favor of the issues than actual leaders and are more divided on the issues than are actual leaders. When leadership background is controlled, 0 strong biasing process is evident in the policy preferences of those who succeed in becoming leaders. Future research will have to consider the nature of this biasing process. I will discuss this in the last chapter. When background is controlled for the combined sample of actual and potential leaders, very little consistency is noted in the policy preferences of the different leadership backgrounds. It is difficult to characterize any of the leader background types as being the most or the least favorable. Thus, it seems apparent that the bias in the policy preferences of leadership is a bias in the attitudes toward the various policies rather than a bias in the leaders' back- grounds. This is but another confirmation of the fruitfulness of research on this question. The fourth major finding showed the weakness of the bias toward the policy preferences of the political activists. AlthOUgh there is increasing approval of the issues and a polarization of opinion with increasing political activity, leadership in Springfield shows only a weak bias toward the preferences of the more active. And Eugene leaders show some tendency to run counter to 2Edmond Costantini, "lntraparty Attitude Conflict: Democratic Party Leadership in California," The Western Political Quarterly, XVI (December, I963), p. 97l; Samuel J. Eldersveld, Political Parties: A Behavioral Analysis (Chicago, I964), pp. l90-l96. I42 the preferences of the more active. There are many inconsistencies in the relationship between policy preferences and the level of political activity. In general, there are patterns of policy preferences in the community. And frequently, these patterns are not consistent with the background variables often used as indicators of them. It is also apparent that many of the distinctions among leaders are of little importance in understanding these patterns of policy preferences . CHAPTER VI AWARENESS ON THE PART OF THE UNREPRESENTED CITIZEN My concern in this chapter is with those persons who are most poorly represented by community leadership in each of the issues. Several questions concerning these persons will be investigated, but my greatest interest will be in their awareness of their situation and the potential for community conflict should they grow more numerous and more aware of their situation. The Unrepresented Citizen Persons whose attitudes on any of the issues are held by none of the leaders in that community are identified. as unrepresented citizens. Because of the few persons so identified in Eugene, a less rigorous criterion was used there . If the attitude of an individual in EUgene was held by lessthan 5 per cent of the leaders in that community, he was identified as an unrepresented citizen. Table 6-I shows the number of persons so identified in each community on each issue and the attitudes they held . It is quite apparent that the unrepresented citizens in these communities are those who do not approve of adopting any of the pragrams involved in these issues. There is very little overlapping among those who are unrepresented on the various issues. Table 6-2 shows that only 24 per cent of the unrepresented citizens in both communities are unrepresented in more than one issue. The citizen who I43 I44 Table 6-1: The attitudes on the issues and the number of persons identified as unrepresented in each of the issues in both communities. Springfield The Number of Persons Issue Unrepresented Their Attitudes on the issue Attraction of Industry I8 Uncertain, Disapprove, Strongly Disapprove Annexation I2 Strongly Disapprove Parking Lots I2 Strongly Disapprove Special Education I2 Strongly Disapprove Fluoridation 64 Strongly Disapprove Public Housing 0 Urban Renewal 0 Metropolitan Park 45 Disapprove, Strongly Disapprove Public Kindergartens 0 Eugene Attraction of Industry 29 _ Uncertain, Disapprove, Strongly Disapprove Annexation 39 Disapprove, Strongly Disapprove Parking Lots 0 Special Education 20 Strongly Disapprove Fluoridation O . Public Housing 0 Urban Renewal 27 Strongly Disapprove Metropolitan Park 53 Disapprove, Strongly Disapprove Public Kindergartens 0 Table 6-2: The number of issues on which the unrepresented citizens are unrepresented . Community Number of Issues on Which Unrepresented One Two Three Four Springfield 95 2 I 8 I Eugene 99 27 5 I45 Is unrepresented in only one issue would be expected to be less conscious of his situation than a citizen who is unrepresented in many issues. And even if he were aware of it, he would be expected to be less troubled. The number of issues on which an individual is unrepresented or the generality of his unrepresentativeness will be shown to be an important dimension in the following analysis. Althoth the unrepresented citizens in any given issue are not likely to be unrepresented in any other issue, they are, as a class, not in favor of adopting any of the pmgrams. On every issue in both communities, the unrepresented citizens reflect less favorable attitudes’ than other citizens. This attitude might well be expected from the favorable bias of community leadership. Tables 6-3 and 6-4 show that the unrepresented citizens are not atypical of the more'represented persons in the communities in terms of general political Table 6-3: Comparison of the general political involvement of the unrepresented citizens and the more represented citizens. General Political Involvement (0) Law (I) (2) (3) High Total % N Springfield Represented 34.8% 53.6 7.8 3 .8 100.0 319 Unrepresented 39.0 52.8 5.7 2.5 99.9 . I23 EUgene Represented 32.] 51 .8 3.9 100.0 386 Unrepresented 26.2 56.9 I2.3 4.6 I00.0 I30 146 ’ Table 6-4: Comparison of the social status of the unrepresented citizens and the more represented citizens. Social Status (0) Low (1) (2) (3) High Total % N Springfield Represented 17.2 34.3 29.9 18.6 100.0 274 Unrepresented 21 .3 37.2 28.7 12.8 100.0 94 Eugene Represented 13.4 20.2 31 3 35.2 100.1 307 Unrepresented 18.3 20 4 36.6 100.0 93 involvement or social status. The unrepresented citizens of Springfield show a very slight tendency, not statistically significant, to be less politically involved and of lower social status than the represented; but little can be said about the unrepresented citizens of Eugene.l They closely reflect the general political involvement and social status characteristics of their represented counterparts. Despite the relationships between the approval of the pragrarm and both higher social status and greater general political involvement and the fact that the unrepresented citizens have generally less favorable attitudes on the issues, being unrepresented is not a social class or political stratum characteristic . Again the dangers of using social class as an indicator of attitudes are apparent. lChi Square was used to evaluate the statistical significance of the relationships presented in this chapter. As the data in this chapter are entirely derived from the random sample of followers in the two communities, a relatively accurate measure of sampling error can be made. 147 Awareness of Being Unrepresented .The unrepresented citizens in both communities show evidence of being aware of their situation, althOUgh they seem to perceive their situations as typical of all voters or citizens. They do not see themselves as an unjustly treated minority. In both communities the unrepresented citizens show a statistically significant pattern, seeing city officials as not acting consistently with the public's will. Table 6-5 shows this relationship. Furthermore, this belief that I Table 6-5: Comparison of how the unrepresented and the represented citizens view the activities of city officials. ' City officials do: Pretty much What some of the What they what the more influential themselves citizens want. people Want . think best. N Springfield ~ Represented 52 .4% 26. O 21 .6 273 Unrepresented 35 .5 34 . 5 30 . 0 1 10 Eugene Represented 41.1 34.2 24.8 319 Unrepresented 26 . 3 45 . 8 28 . 0 1 1 8 the city officials do not do what the citizens want grows even stronger when the individual is unrepresented on more than one issue, as shown in Table 6-6. But another relationship is also evident in Table 6-6. Not only do the unrepresented manifest their awareness of their situation, but they also have differing perceptions in the two communities as to whose policy preferences are represented. In Springfield the pattern is for those who are more 148 Table 6-6: Comparison of the unrepresented and the represented citizens' view of the activities of city officials, controlling for the number of issUes on which the unrepresented citizen is not represented . City officials do: Pretty much What some of the What they what the more influential themselves citizens want. people want. think best. N Springfield Represented 52.4% 26 .0 21 .6 273 Unrepresented (1) 39.0 30.5 30.5 82 (2) 31.6 42.1 26.3 19 (3) 12.5 62.5 25.0 8 (4) 0.0 100.0 0.0 1 Eugene Represented 41.1 34.2 24.8 319 Unrepresented (1) 31 .5 46.1 22.5 8"}- (2) 12 .0 44.0 44.0 25 (3) 0.0 50.0 50.0 4 frequently unrepresented to see city officials responding to the preferences of the more influential. The unrepresented citizens of Springfield see an elite con- trolling the community political process. In contrast, the unrepresented citizens of EUgene are more divided than their counterparts in Springfield as to whose policy preferences influence the behavior of city officials. The belief that city officials do what they themselves think best is common among the move unrepre- sented in Eugene. But does this response mean that they see city officials as members of the elite, city officials and the elite being coterrninous, or that they see city officials as being motivated by technical or other opinions apart from what the influentials desire. The fact that an alternative to the question indicating that the respondent 149 sees elite control in the community was available and that such a large per- centage of the unrepresented did choose to answer in that way would seem to belie the former interpretation, because persons believing city officials were the community's elite would be expected to answer using that alternative response. But better evidence is available for evaluating which interpretation is more correct. Tables 6-7 and 6-8 do offer some evidence in support of the latter inter- pretation of the data in Table 6-6. The unrepresented citizens of Eugene tend Table 6-7: Comparison of the unrepresented and the represented citizens' view of the importance of voters in the making of key decisions on major policies in the community. Voters Very Important Not so Important N Springfield Represented 87 . 5% 12 . 5 304 Unrepresented 85.3 14.7 1 16 Eugene Represented 86.6 13 4 365 Unrepresented 70.7 1 16 Table 6-8: Comparison of the unrepresented and the represented citizens' view of the importance of businessmen in the making of key decisions on major policies in the community. Businessmen Very Important Not so Important N Springfield Represented 88 .6% 1 1 .4 299 Unrepresented 94 . 0 6 . 0 1 14 Eugene Represented 92 .1 7. 9 366 Unrepresented 92 .7 7 . 3 1 10 G. o to view the voters in the community as not being very important, but the unrepresented citizens of Springfield show only a very weak and statistically insignificant trend in this direction. When the citizens are asked to evaluate the importance of businessmen, there is no difference between the represented and the unrepresented in EUgene; but the unrepresented of Springfield see business- men as being more important than do the represented citizens of Springfield . The two patterns would seem to support the argument that the unrepresented citizens of Eugene see city officials as not acting consistently with the desires of the citizens, nor as the businessmen might want. If the unrepresented citizens of EUgene saw city officials and the community elite as. being coterminous, they would show the same relationship as the unrepresented citizens of Springfield, namely, they would define businessmen as being "very important." Many of the unrepresented citizens of Eugene see city officials as acting on preferences other than those of the people of the community, but not necessarily acting on the preferences of businessmen or other members of the influential elite . The community political process is seen as a conflict of wills between the public and the city officials. The patterns also support the conclusion that Springfielders see an elite controlling the city. If an elite controlled city officials, as the unrepresented citizens of Springfield show evidence of believing, some, but not all, voters would be without influence in the community. But for the unrepresented citizens of EUgene, who see the political process in the community as a cmflict between the officials and the citizens, voters as a whole would be seen as "not so important. " The question of the importance of voters would draw more clearly 151 formed attitudes from the unrepresented citiztms of EUgene because of their perception of the political process. This evaluation of voter significance grows more negative with increasing generality of unrepresentativeness in Eugene, as seen in Table 6-9. The pattern for the Springfield unrepresented citizens is even less apparent when the generality of unrepresentativeness is controlled . Table 6-9: Comparison of the unrepresented and the represented citizens' view of the importance of voters in the making of key decisions on major policies in the community, controlling for the number of issues on which the unrepresented citizen is not represented . Voters Very important Not so Important N Springfield Represented 87 . 5% 12 . 5 304 Unrepresented (1) 87.9 ' 12.1 91 (2) 64.7 35.3 17 (3) 100.0 0.0 7 (4) 100.0 0.0 1 Eugene Re presented 86 . 6 1 3 . 4 365 Unrepresented (1) 77.0 23.0 87 (2) 56.0 44.0 25 (3) 25.0 75.0 4 The Accuracy of the Perceptions of the Unrepresented Citizens There is some cause to doubt the accuracy of the unrepresented citizens' perceptions of the community political processes. It is true that not one of the city officials of Springfield was identified as a community leader. This saggests that they, at best, have no role in decision-making in Springfield and perhaps, 152 that they are only ministers of non-official leaders in that community. Further- more, this non-city-official leadership group in Springfield has many of the characteristics of an elite. It is relatively monolithic in comparison with the community, and it reflects the policy preferences of the higher social-status level. There would seem to be some truth, then, to the perceptions of the Springfield unrepresented citizens. But the differences between the cities are only a matter of degree . Eugene shows all of the same relationships shown in Springfield but more weakly. Only two of the city officials of Eugene, the mayor and a councilwoman, are among the community leaders. The majority of the occupationally vulnerable leaders are county officials. And Eugene leaders also better reflect the policy preferences of the higher social status persons in the community and are more consensual in their policy preferences than is the public. There is very little evidence to show the political process in Eugene is a conflict between city officials and the public. If it were, more city officials would be expected among the leadership, and vulnerable leaders in that community would be expected to differentiate themselves from less vulnerable leaders. Neither relationship is different from that found in Springfield. In conclusion, it appears that the unrepresented citizens' perceptions of the nature of the community political process are accurate in the sense that they see community leadership as less than accurately reflecting the policy preferences of the public . But these perceptions are inaccurate in the sense that city officials in one community are seen as acting at the behest of the influential 153 and in the other community, are seen by many as acting on their own personal preferences. No such distinction is apparent in the data . The Irfinct of the Realization of One's Unrepresentativeness What is strongly apparent in the preceding analysis and in Table 6-10 is that the unrepresented citizens do not view themselves as atypical. Whatever Table 6-10: Comparison of the unrepresented and the represented citizens' view of their personal influence. Personal Influence Very More Average Less Influential Influential in Influential than Most Influence than Most N People Springfield - Represented 0.0 4 .7 54 .1 41 .3 320 Unrepresented 0 .0 2 .6 54 .7 42 .7 1 17 Eugene Represented .5 5.3 49.2 44.9 376 Unrepresented . 8 5 .6 46 .8 46. 8 126 difficulties they have in getting the community political process to be responsive to their policy preferences, it is seen as a difficulty shared with all citizens. There is no reason to feel personally disenfranchised if everyone shares your situation. Table 6-11 shows only a slight and not statistically significant tendency for the unrepresented to see themselves as less efficacious. The other measure of the alienated voter syndrome saggested in Chapter II is political cynicism, and it is also affected by the distinction between an 154 Table 6-11: Comparison of the unrepresented and the represented citizens' sense of efficacy. Sense of Efficacy (4) Low (3) (2) (I) (0) High N Springfield Represented 12 .O% 23.5 28.9 12.0 23.5 166* Unrepresented 12 . 9 35 .7 21 .4 12 9 17.1 70 EUgene ‘ Represented 12.1 26.8 15.8 16.2 29.0 272 Unrepresented 13.8 34.5 18.4 9.2 24.1 87 *Data derived from mail-back questionnaires. The low N's are a result of the failure of some respondents to return the questionnaires and of the non- scale types which were excluded from the table . elite-controlled political process and a political process in which public officials do what they think best. The persons who see an elite controlling the public officials or local politicians might well be more cynical of these persons than a person believing all is well. But the person who views the political process as a conflict between the citizens and the officials, with the citizens being unable to get a response from the officials, would be expected to be even more cynical of these officials and politicians. This is the case as shown in Table 6-12. The unrepresented citizens in both communities are more likely to be cynical of local politicians than are the represented citizens, and the relationship grows stronger with increasing generality of unrepresentativeness. The relationship is stronger in Eugene, as one would expect. Although the unrepresented citizens do not feel that they are less influential or show evidence of decreased efficacy, they are more cynical. 155 Table 6-12: Comparison of the unrepresented and the represented citizens' cynicism toward local politicians, controlling for the number of issues on which the unrepresented citizen is not represented . Cynicism" High Medium Low (0,1) (2,3,4) (5,6) N Springfield Represented 22 . 7% 52 . 2 35 .0 154 Unrepresented (1) 29. 2 35 . 5 35 .4 48 (2) 12.5 37.5 50.0 8** (3) 66.6 33.3 0.0 6 EUgene Represented 17 . 0 42 . 4 40 .7 248 Unrepresented (I) 28.2 40.7 31 .2 64 (2) 50.0 33.3 16.6 12** (3) 100.0 0.0 0.0 I *See Appendix C. **The small n's in this table would suggest the collapsing of categories. This was not done because of the importance of showing the consistency of the relationship. Another Hypothesis, The Disgruntled Man It is possible that unrepresentativeness and an awareness of this state are components of the same syndrome and not independent and related concepts as suggested in the above analysis. This hypothesis, which I will call the disgruntled man hypothesis, would suggest that some persons in any society are, by their very nature, disgruntled. This is expressed by them in dissatisfaction with political leaders and in being against innovative government pragrams in the community. Thus, my method would identify such a person as unrepresented because of his .1 felt I Ill-lull I. 1| ll'.‘ 1 I Ilrll l ‘1 . In'l {lull 156 negative attitudes on new prOgrams. He would also be identified as being aware of his unrepresented situation because'of his dissatisfaction with political leaders, etc . Unrepresentativeness, therefore, would not explain his diSpleasure with the politics of his community as such, but would be another aspect of his general dngruntlement. Furthermore, it might be argued that the failure of leadership to represent the disQruntled man is desirable in democracy because his attitudes are not rationally formed and, if considered, would obstruct the functioning of democracy. Again this is a normative question in which I do not wish to get involved . The question of how large the disgruntled minority must be before democratic leaders must respond to it, however, must be answered by persons taking this position. Table 4-2 would indicate that the diSQruntled man is not always the very small minority which is expected to have unpopular attitudes of any given issue. Many of the distributions are much flatter than the normal curve and often bimodal .1 Are the opinions of 15 per cent of the community to be disregarded as "dngruntled"? Aside from this normative implication of the hypothesis, there is some evidence available in the study to suggest that the hypothesis is incorrect. First, the fact that so few individuals were unrepresented on more than one issue would suggest that their attitudes are dependent on the issue rather than the general negativism squested by the hypothesis. Second, the growth of intensity of awareness with increasing generality of unrepresentativeness shown in Tables 6-6, 6-9, and 6-12 would saggest a relationship between independent variables of unrepresentativeness and awareness of unrepresentativeness rather than their being part of the same syndrome . 157 Several research designs can be suggested for evaluating the correctness of the disgruntled man hypothesis. First, the research design used in this study can be repeated in the hope of discovering an unrepresented minority which supports the adoption of a new governmental program. Such a group would not be expected under the hypothesis. Second, a longitudinal study could be under- taken to ascertain whether those who are negative to local politics and politicians adopt unfavorable attitudes to suggested new pragrams or whether those who have unfavorable attitudes and are unrepresented in those attitudes grow more negative about their perceptions of local politics and politicians as the issue moves from introduction to decision. The former would support the hypothesis. Third, a longitudinal study could be undertaken to show whether there is growing satisfaction among those aware of their unrepresented situation when government responds to their wants . The second design is best as it is not dependent on events or circumstances in the community. Conclusions My emphasis throughout this study has been the response of the public to the actions of leaders in the political process. This concern has been focused on one of many possible leader-follower relationships, namely the attitudes of leaders on issues of immediate concern to the community, the representativeness of these attitudes with respect to those held by the public, and the response, if any, of the public to conditions of unrepresentative leadership. If some linkage between leaders and followers is desirable in democracy or important to the stability of a political system, I have attempted to demonstrate the ineffectiveness of other 158 pr0posed linkages such as cor..peting groups or competitive political parties at the community level. Thereby, l have attempted to emphasize the importance of exploring the individual's perceptions and expectations of political leaders and government, the leaders' role perceptions and knowledge of the public's attitude, and other social-psychological linkages between leaders and followers . The concept of the alienated voter plays an unusual role in this analysis. In my reading of the literature of the alienated voter, I was struck by the lack of empirical evaluation as to whether this means of response to the political system was explainable in terms of personality factors or whether it was a response which would be understandable in terms of the social-psychoIOgical position of persons showing alienated responses. My personal research bias is to exhaust the latter type of explanations before seeking personality variable explanations. My concern with the alienated voter behavior then was with discovering if this behavior is that of individuals who believe that leadership is not responsiveto their wants, and in addition, if this belief is founded on their somehow perceiving the unrepresentativeness of that leadership. I have not sought to explain all alienated voter behavior. Thus rather than broadening and improving my investigation of the alienated voter, I would suggest a more extensive exploration of the social-psychological role of the individual in community politics and the objective basis of their perceptions. I am inclined to expect the disgruntled manhypothesis will be proven incorrect. Thus apart from investigating this fundamental issue, several improve- ments and expansions of this study might be suggested for inclusion in future research. The following discussion therefore assumes the incorrectness of the 159 disgruntled ..1an hypothesis. A longitudinal study of one or more communities is called for infuture research in this area. This study would focus on the development of issues in the community, the evolution of leadership's attitudes on those issues, and in turn the reSponse of the public to those leadership attitudes both at the conception of the issue, during its development as a concern to the community, at the time of decision, and for some time after the decision. Such a study would allow an understanding of the causal relationship involved in the public's response to leadership attitudes. Attention should be directed to the communication process by which the public becomes aware of the leaders' attitudes and whether the knowing of these attitudes is a long-term process derived from limited but consistent information received by the public. Is there consistency in the response of the leadership to new issues and does the public perceive that consistency? In this study I found evidence that the unrepresented were aware of their situation but did not show the negativism attributed to the alienated voter. I suggest that this is a result of the low generality of unrepresentativeness and the viewing of their situation by the unrepresented as not being unusual or atypical. Future research should include a more extensive evaluation of the expectations of the public with respect to the leaders, their satisfactions with the performance of those leaders, and their acceptance of beliefs which are integrative to the political system. Some theoretical work along the lines of Almond and Verba's work in the Civic Culture and Lipset's in The Political Man would be needed before these integrative beliefs could be operationalized. I have previously noted that the issues of concern in the communities investigated were very similar 160 in a number of ways. These issues were middle-class valued, elite originated, and largely nationally defined. Are there issues which originate from the public, issues which are lower-class valued, issues which are locally defined? Preparatory research into the communities to be studied would seemingly be necessary to define such issues, and the longitudinal design would facilitate detecting them as they are introduced in the community and to identify the originator or originators. If the generality of unrepresentativeness is important to the nature of the response to it, preparatory research might also allow the identification of com- munities in which the leadership appeared to reflect very inaccurately the desires of the community. Such communities would presumably include persons who are extensively unrepresented by leadership and who might more clearly show the results of perceiving this unrepresented situation . I do not wish to argue that the public in American communities is issue oriented and will rise up against those leaders who fail to do the public's bidding . I merely argue that the public over a period of time can perceive certain biases of its leadership if the biases are sufficiently great and out of ignorance as to what to do about it and frustration with their inability to change the situation may grow dissatisfied with the existing leadership in the community. This anger may vent itself in any instrument which is perceived by the unrepre- sented as capable of showing that anger to the leaders. The public's expectations of leadership, the areas and degree to which the public perceives unrepresenta- tiveness, their perceptions and misperceptions of how to best vent this anger, 161 and their estimations of their success are, of course, all questions which should be investigated . In general, I suggest that expectations and perceptions of the public with respect to community leadership as well as the responses of that public to the behavior of leadership needs to be thorothly explored. This study might well be combined with the study of other hypotheses wggested in the concluding chapter. CHAPTER V“ CONCLUSIONS The relationship between cleavages in beliefs, values, and attitudes among the members of a society and conflict in that society has been of interest to social scientists for some time .1 In one of the most extensive discussions of the subject, Coleman sees a very common pattern in the development of conflict once it is started.2 But he notes the need for basic cleavages in values or interests in order to kindle the initial spark of conflict as well as to perpetuate its development once started .3 But what types of cleavages are there, and which are most apt to spark conflict? l have been most concerned with the existence of cleavages in policy preferences between leaders and followers in two American communities. Previous theory and research was used to suggest possible meaningful distinctions among leaders which might affect the disparity between their policy preferences and those of the followers. Although leadership has long been of interest to researchers and much is known of the characteristics of leadership as compared with the public, IGabriel A. Almond and Sidney Verba, The Civic Culture (Princeton, 1963), p. 492; and Seymour Martin Lipset, Political Man (Garden City, N.Y., I959), p. 78. 2James S . Coleman, Community Conflict (Glencoe, Illinois, I957), p. 9. 3|bid., p. 10. I62 I63 4 little is known of leaders' beliefs and values. Nor has much research been completed on the importance of beliefs and values in integrating the community. My purpose has been to explore these beliefs and values and to note the effects of cleavage between leaders and followers on the conflict potential in these communities . The Bias of Community Leadership As I have previously noted, leadership in these two communities has shown itself to be a relatively monolithic group holding policy preferences not entirely shared by the publics of the two communities. Leadership in these two communities is unrepresentative of the public . For the most part, their bias is in the direction of more favorable attitudes on the issues than those held by the public . The Distinction Between Vulnerable and Less Vulnerable Leaders Another fact noted in the preceding chapters was the absence of differences in policy preferences between leaders who are under the threat of public disfavor through elections and leaders not subject to such threats. There is little evidence 4Wendell Bell et al . give an excellent summary of the known character- istics of public leadership. Wendell Bell et al., Public Leaderslip (San Fran- cisco, Calif. , l96l). Chapter II includes a summary of the limited research completed on various beliefs and attitudes of leaders. 5Phi|ip E. Jacob and Henry Teune, "The Integrative Process: Guidelines for Analysis of the Bases of Political Community," in The Integration of Political Communities ed. Karl W. Deutsch et al. (Philadelphia, I964), p. 22. 164 that the public in these two communities has succeeded in electing men whose personal policy preferences on these issues reflect the policy preferences of the public . Coercion of leaders via elections does little to improve the representa- tiveness of leadership. Being subject to elections was the objective criterion I used in deter- mining whether a leader was vulnerable or not. Leaders may have personal assessments of their own vulnerability, and those perceiving themselves as more vulnerable may strive more than others to express what they perceive to be the desires of the public. Thus, it may be that a subjective rather than an objective assessment of vulnerability would better differentiate between leaders. Serving the public's will through one's awareness of personal vulnerability to the public is schematically shown in the lower left hand cell of the schema in Figure I-I . The entire lower row of this schema was not included in this research . In addition to the research design used in this study, future research should undoubtedly be directed to measuring the degree of subjective vulnerability among leaders and the degree of accuracy of leaders' perceptions of the followers' policy preferences. Once such measurements are made, the assessment can be made as to which of the four cells in Figure l-l best explains the degree to which leaders serve the public's will. The Distinction Between Actual and Potential Leaders Six leadership types were sampled in this study. Within each of these types, a noticeable difference existed in the policy preferences of those who were also identified as leaders and those who were not so identified . The process 165 or processes that affect the policy preferences of the leaders work to make these leaders a biased selection of potential leaders. This biasing of leaders' policy preferences was not as neatly related to increasing involvement in the political process as Eldersveld noted in his study of the Detroit area.6 But there was no evidence that men in the highest echelons of leadership were less biased, as sUggested by Costantini's research .7 Rather, the relationship was very nearly a step function with the entire difference found in the distinction between actual leaders and all other persons in the community . Further research on this discovery should attempt to discover how this bias in the policy preferences of actual leaders develops. Four processes can be conceived: (I) selective recruitment of men holding attitudes consistent with those of the existing community leadership; (2) cooptation of leaders holding opposing preferences by means of status satisfaction, etc .; (3) self-socialization on the part of new leaders who have learned of important facts or conditions previously unperceived; or (4) atypical attitudes among the most successful men within each of the leadership background types, with the most successful men in each background type being most likely to become leaders. A longitudinal study of community leadership would be necessary to demonstrate which of the alter- native explanations best accounts for the bias, or whether all are significantly 6Samuel J. Eldersveld, Political Parties: A Behavioral Analysis (Chicago, I964), p. 193. 7Edmond Costantini, "lntraparty Attitude Conflict: Democratic Party Leadership in California," The Western Political Quarterly, XVI (December, I963), p. 97l . I66 important. A large pool of potential leaders would also have to be interviewed so that changes in their preferences, if any, on entering leadership could be noted. I have previously commented on the improbability that men are selected for leadership because of their attitudes on specific issues such as those dealt with in this study. The short-term nature of these issues would seem to preclude this process. But a set of more basic beliefs or an ideolon may underlie the attitudes of leaders on these specific issues, and the selection may be made to conform to the ideolon of existing leadership. For this reason attempts to extend the work done with ideologies by Agger et al . and by Williams and Adrian would also seem necessary in order to identify which one or more of the processes listed above best explains the bias of leadership.8 IdeOIOgies are but one of the different types of beliefs that could have been asked instead of, or in addition to, the palicy preferences an imminent issues studied here. In the last part of Chapter V, I noted the lack of variability in the aspects or facets of the items included in this study. For the most part, the imminent policy issues are originated by political activists, nationally defined, and middle-c lass valued. Issues exemplifying other facets should be included in future studies. If the longitudinal study could be extended over a long period of time, efforts might be made to capture attitudes on an issue or issues before and after they became topical in a given community. At the same time, the develop- ment of consensus, if any, could be noted among existing leaders along with 8Robert E. Agger, Daniel Goldrich, and Bert E. Swanson, The Rulers and the Ruled (New York, I964); and Oliver P. Williams and Charles R. Adrian, Four Cities: A Study in Comparative Policy Making (Philadelphia, I963). I67 other responses of leadership to new issues. Awareness of Being Unrepresented The more extensive the cleavage between leaders and followers on policy preferences, the larger would be the number of followers who do not share the policy preferences of the leaders. Looking at this unrepresented minority in Chapter VI, three observations were made. First, 'these unrepresented citizens differed from issue to issue. There were a few persons who were unrepresented on more than one issue, but for the most part, no group of citizens was chronically unrepresented . Second, these unrepresented citizens did show an awareness of their situation. Third, rather than seeing themselves as atypical, the unrepre- sented citizens believed a! voters shared the same difficulty. This perception may account for the lack of negativism in their attitudes. The lack of negativism should not obscure the fact that there is evidence of a breakdown of adherence to beliefs which are integrative of the polity among the unrepresented. They see voters as having less impact of the decisions made in the community and they have a more cynical opinion of local politicians than do the better represented citizens. Leader-Follower Cleavage as a Source of Community Conflict Apparently some degree of unrepresentativeness or cleavage exists between leaders and followers, and those members of the public most affected by this unrepresentativeness show an awareness of their situation . Thus, leader-follower cleavage does have an impact on the followers. Cleavage leads to decreased adherence to integrative beliefs and thus, to a greater potential for conflict. Given some initial spark, this potential could lead to conflict in the community. I63 What factors would seem to affect this potential for conflict? The generality of unrepresentativeness is important in developing aware- ness of unrepresentativeness, as shown in Chapter VI . Thus, factors affecting this variable will affect the potential for conflict derived from unrepresentative- ness. The greater the skewness and the more monolithic leadership's preferences are relative to those of the public, the more numerous would be the unrepresented citizens in that particular issue. Also, the unrepresentativeness of leadership affects the number of unrepresented citizens. This, in turn, affects their chances for being unrepresented in more issues. Apart from the size of the class of unrepresented citizens, the more interdependent the preferences of the public are in the various issues, the more likely it is that an unrepresented citizen will be unrepresented in more than one issue. Converse uses a concept of constraint to understand the c0gnitive system of an individual. Two attitudes have constraint on each other if a change in one demands a change in the other. He argues that there is little constraint or interdependence between the attitudes of less educated and less involved persons, and that constraint increases with education and involvement.9 The absence of this constraint between attitudes decreases the probability of having a general class of chronically unrepresented citizens .. Such constraint between beliefs might be derived from a common interest among a subgroup of citizens such as retirees or lower-class home owners; an 9Philip E. Converse, "The Nature of Belief Systems in Mass Publics," in IdeoIOgy and Discontented.David E. Apter (New York, 1964), p. 24] . I69 ideolOgy, independent of interests; the need for loQical consistency in one's attitudes; or an in-group out-group distinction extended to our beliefs and their beliefs. All of these sources of constraint can be tatht. Thus, a person interested in achieving influence in the community could mobilize the cleavages existent in the community by educating constraints between certain attitudes. llis h0pe would be to make the unrepresented in each of the beliefs more aware of the unrepresentativeness of community leadership and more willing to support him . The opportunity to see one's situation would seem necessary to develop an awareness of it. As I have noted, the frequency of being unrepresented is important to awareness, but other factors have their effect. The size of the class of unrepresented citizens seems to be a factor which would encourage inter- action among the unrepresented. This could result in an unrepresented citizens' group capable of concerted action and of educating and proselytizing others. The very size of the unrepresented citizens' group may encourage alternative leadership, distinct from existing leadership, to utilize the unrepresented citizens as a political stepping stone. By pointing out their situation, this alternative leadership could stimulate awareness and act as a catalyst to develop the potential for conflict. If numerous issues demand decision in the community, unrepresented citizens would be given more opportunity to note the unrepresentativeness of community leadership. The actual activity in the community political process may also make such matters more exciting and more salient to the public, and the result would be a greater awareness on the part of the unrepresented citizens. I70 Finally, developing the potential for conflict from unrepresentativeness necessitates that the unrepresentativeness of leadership be contrary to the expectations of the followers. If they did not expect leaders to be responsive to their desires and wants, or if they card little about the policy output of govern- ment, or did not believe government could help them with matters important to themselves, there would be little discontentment when they became aware that they were not represented . External phenomena, as I have noted, have their impact both on the dynamics of unrepresentativeness which lead to a potential for conflict and on the dynamics of potential conflict sparking into actual conflict. A politically adept alternative leadership could cultivate even minor unrepresentativeness. Through the use of propaganda and secrecy, existing leadership can obscure even gross unrepresentativeness; or they can coopt all alternative leadership. Existing leadership may be able to disregard discontent among the followers through the use of violence or more subtle sanctions to maintain their position . External or internal demands for change may unsettle the lethargy of the public and make leadership's actions more manifest. Limitations of the Study In many ways the findings of this study are limited in their applicability to other communities and to all issues or beliefs. First, the two communities studied, although contrasting sharply in a sociOIOgical comparison, are both in the same region of the United States and are also in close proximity. There is no reason to assume they are atypical of all American communities, but there is 17] also no reason to assume they are typical. l have already commented on and outlined a more extensive set of preferences and beliefs which might be included in future studies. But the issues included in this study represent a universe of issues of great concern and imminent decision in these communities at the time of this study. Thus, the findings can be generalized to the level that one can say leadership in these two communities is unrepresentative in the issues of concern in these communities and such unrepresentativeness is likely to have immediate policy results. Another limitation is the incompleteness of the inventory of integrative beliefs used in the analysis of awareness. Only Lipset and Almond and Verba have made any attempt to offer a more or less complete set of such integrative beliefs.lo Awareness of unrepresentativeness, however, is manifested even on this limited set of integrative beliefs, but little can be said about the nature of this awareness ;. Finally, only a crude method of identifying the unrepresented citizens was used in this study. What is the smallest percentage of leaders that can give adequate expression to an attitude on an issue? If ten per cent of the public strongly disapproved of an issue, can one per cent of the leadership sharing this attitude give this minority adequate expression? There is, however, no reason to believe the method used in the study explains the results. In this study, I have explored the concept of the representativeness of community leadership and the impact of unrepresentativeness of leaders on the loLipset, p. 8]; and Almond and Verba, p. 16. I72 potential for conflict in the community. Conceding the above limitations, two processes were discovered which deserve further research -- the biasing processes of leadership selection and the reaction-to-being-unrepresented process among the followers. Both processes were evident in these two communities; and working tOgether, they lead to increase potential for conflict in the two communities. Our understanding of these processes is too limited, and we have no knowledge as to whether these processes were common in the past; but the increased importance of local govemment in the average American's life and the rapid changes and growth being experienced by local governments would seem likely to accelerate the processes leading to community conflict. Certainly, the defeats of school bond issues and the conflict over fluoridation, if taken as examples of community conflict, would indicate that conflict has become more frequent in recent years. A thorough understanding of the processes leading to conflict and the resulting potential for corrective action is a necessary goal for social scientists. BIBLIOGRAPHY Books Agger, Robert E., Daniel Goldrich, and Bert E. Swanson, The Rulers and The Ruled, New York, John Wiley 8. Sons, Inc ., I964. Almond, Gabriel A. and Sidney Verba, The Civic Culture, Princeton, Princeton University Press, I963. Apter, David E., IdeOIOgy and Discontent, New York, The Free Press of Glen- coe, I964. Bell, Wendell, Richard J. Hill and Charles R. Wright, Public Leadership, San Francisco, Chandler Publishing Company, I96I . Berelson, Bernard R., Paul F. Lazarsfeld, and William N. McPhee, Voting, Chicago, The University of Chicago Press, I954. Bollens, John C., Exploring the Metropolitan Community, Berkeley, California, University of California Press, I96I . Burdick, Eugene and Arthur J . Brodbeck (eds.) American Voting Behavior, Glencoe, Illinois, The Free Press, I959. Burke, Edmund, The Works of the Right Honorable Edmund Burke, 2, Boston, Wells and Lilly, I826. Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stakes, The American Voter, New York, John Wiley 8. Sons, Inc. , I960. Campbell, Angus, Gerald Gurin, and Warren E. Miller, The Voter Decides, Evanston, Illinois, Row, Peterson, and Company, I954. Cantril, Hadley, The Politics of Despair, New York, Collier Books, I958. Centers, Richard, "Attitude and Belief in Relation to Occupational Stratification," in Katz, Cartwright, Eldersveld, and Lee eds. Public Opinion and Progganda, New York, Henry Holt and Company, I954, pp. l32-I50. Coleman, James S ., Community Conflict, Glencoe, Illinois, The Free Press, I957. I73 I74 "- I Converse, Philip E., he Nature of ‘3elief Systems in Mass Publics," in David E. Apter, ed., Ideolon and Discontent, New York, The Free Press of Glencoe, I964, pp. 206-26l . Dahl, Robert A., Modern Political Analysis, Englewood Cliffs, New Jersey, Prentice-Hall, Inc., I963, pp. 55-60. Dahl, Robert A., A Preface to Democratic Theory, Chicago, The University of Chicago Press, I956. Dahl, Robert A., Who Govems?, New Haven, Yale University Press, I96I . Dahl, Robert A. and Charles E. Lindblom, Politics, Economics, and Welfare, New York, Harper 8 Row Publishers, I963. Downs, Anthony, An Economic Theory of Democracy, New York, Harper 8. Brothers, I957. Eldersveld, Samuel J., Political Parties: A Behavioral Analysis, Chicago, Rand McNally 8. Company, I964. Goldstein, Marshall N ., "Absentee Ownership and Monolithic Power Structures: Two Questions for Community Studies," in Bert E. Swanson, ed . , Current Trends in Comparative Community Studies, Kansas City, Com- munity Studies, Inc., I962, pp. 49-59. Goode, William J . and Paul K . Hatt, Methods in Social Research, New York, McGraw-Hill Book Company, Inc., I952. Greenstein, Fred I. , The American Party System and the American People (Foundations of Modern Political Science Series), EngTewood Cfiffs, New Jersey, Prentice-Hall, Inc., I963. Hunter, Floyd, Community Power Structure, Chapel Hill, The University of North Carolina Press, I953. Jacob, Philip E. and Henry Teune, "The Integrative Process: Guidelines for Analysis of the Bases of Political Community," in Karl W. Deutsch et al . , ed., The Integration of Political Communities, Philadelphia, J .B. Lippincott Company, I964, pp. I-45. Janowitz, Morris, Deil Wright, and William Delany, Public Administration and the Public - Perspectives Toward 'Govemment in a Metropolitan Com- munity, Michigan Governmental Studies No.436, Ann Arbor, Mihigan, Institute of Public Administration, University of Michigan, I958. {Ill '4' ll I75 Key, V.O . Jr., American State Politics, New York. Alfred A . Knopf, I956. A" Key, V.O. Jr., Politics, Parties, and Pressure Groups, New York, Thomas Y. Crowell Company, 4th ed. , I958. Key, V.O. Jr., Public Opinion and American Democracy, New York, Alfred A. Knopf, I96I . Lane, Robert E., Political Ideol0gy, New York, The Free Press of Glencoe, I962. Lane, Robert E., Political Life, Glencoe, The Free Press, I959. Levin, Murray 8., The Alienated Voter, New York, Holt, Rinehart and Winston, Inc. , I960. Lippman, Walter, The Public Philosophy, Boston, Little, Brown and Company, I955. Lynd, 'Robert S. and Helen Merrell Lynd, Middletown in Transition, New York, Harcourt, Brace and Co., I937. Mayo, H .8., An Introduction to Democratic Theory, New York, Oxford Uni- versity Press, I960. Michels, Robert, Political Parties, New York, Collier Books, I962. Milbrath, LesterW., The Washington Lobbyists, Chicago, Rand McNally and Company, I963. Miller, Warren E. and Donald E . Stokes, Representation in the American Congress, forthcoming, I965. Press, Charles, ed . Main Street Politics, East Lansing, Institute for Community Development, Michigan State University, I962. Presthus, Robert, Men at the Top, New York, Oxford University Press, I964. Schattschneider, E.E., Party Government, New York, Rinehart 8 Co. , I942. Schattschneider, E.E. , The Semi-Sovereigp People, New York, Holt, Rinehart, and Winston, I960. Simon, Herbert A., Donald W. Smithburg, and Victor A. Thompson, Public Administration, New York, Alfred A. Knopf, I959. I76 Stouffer, Samuel A., Communism, Conformity, and Civil Liberties, New York, Doubleday 8 Company, Inc., I955. Truman, David 3., The Governmental Process, New York, Alfred A. Knapf, I960. Wahlke, John C., Heinz Eulau, William Buchanan, and LeRoy C. Ferguson, The Legislative System, New York, John Wiley and Sons, Inc. , I962. Williams, Oliver P. and Charles R. Adrian, Four Cities: A Study in Comparative Policy Making, Philadelphia, University of Pennsylvania Press, I963. Zeigler, Harmon, Interest Groups in American Society, Englewood Cliffs, New Jersey, Prentice-Hall, Inc. ., I964. Articles and Periodicals Agger, Robert E., Marshall N. Goldstein, and Stanley A. Pearl, "Political Cynicism: Measurement and Meaning," The Journal of Politics, August I96I , pp. 477-506. Bachrach , Peter and Morton S. Baratz, "Decisions and Nondecisions: An Analytical Framework," The American Political Science Review, LVIl , September I963, pp. 632-642. Bachrach, Peter and Morton S. Baratz, "Two Faces of Power, " The American Political Science Review, LVI, December I962, pp. 947-952. Chapman, Loren J . and Donald T. Campbell, "The Effect of ACquiescence ReSpons-Set Upon Relationships Among the F scale, Ethnocentrism, and Intelligence," ociometry, 22, June I959, pp. I53-I6I . Converse, Philip E. , "Information Flow and the Stability of Partisan Attitudes," Public Opinion Quarterly, XXVI, Winter I962, pp. 578-599. Converse, Philip E . , "New Dimensions of Meaning for Cross-section Sample Surveys in Politics," International Social Science Journal, XVI, I964, pp. I9-34. Converse, Philip E . and George Dupeux, "Politicization of the Electorate in France and the United States," Public Opinion Quarterly, Spring I962, . I-23. Costantini, Edmond, "lntraparty Attitude Conflict: Democratic Party Leadership in California," The Western Political Quarterll, I6, December I963, pp. 956-72. I77 Dean, Dwight G. ., "Alienation and Political Apathy," Social Forces, 38, March I960, pp. I85-I89. Dean, Dwight G ., "Alienation: Its Meaning and Measurement," American SocioIOgical Review, 26, October I96I , pp. 753-758. Kish, Leslie, "Confidence Intervals for Clustered Samples," American Sociological Review, 22, April I957, pp. I54-I65. Mack, Raymond W . and Richard C. Snyder, "The Analysis of Social Conflict - Toward an Overview and Synthesis," Conflict Resolution, I, June I957, pp. 2I2-248. McClosky, Herbert, "Consensus and IdeOIOgy in American Politics," The American Political Science Review, LVIII, June I964, pp. 36l-382- McClosky, Herbert, Paul J . Hoffman, and Rosemary O'Hara, "Issue Conflict and Consensus Among Party Leaders and Followers," The American Political Science Review, LIV, June I960, pp. 406-427. McGill, Edward L. and Jeanne Clare Ridley, "Status, Anomia, Political Alienation, and Political Participation," The American Journal of SocioIOgy, LXVIII, September I962, pp. 205-2I3. McGinnis, Robert, "Randomization and Inference in SociOIOgical Research," American Sociological Review, 22, No. 5, October I957, pp. 408-4I4. Minar, David W . , "IdeOIOgy and Political Behavior," Midwest Journal of Political Science, V, November I96I, pp. 3I7-33T. Press, Charles, "Concepts of Access and Representation in Metropolitan Area Studies," paper delivered at I96I American Political Science Association meeting, reported in Press eds. Main Street Politics, East Lansing, Michi- gan State University, I962, pp. 84-85. Prothro, James W. and Charles M. Grigg, "Fundamental Principles of Democracy: Bases of Agreement and Disagreement," The Journal of Politics, May I960, pp. 276-294. Rossi, Peter H . , "Community Decision-Making," Administrative Science Quarterly, I, March I957, pp. 4I5-443. Rothman, Stanley, "Systematic Political Theory: Observations on the Group Approach," The American Political Science Review, LIV, March I960, pp. I5-33. I78 Schulze, R.O. and L.U. Blumberg, "The Determination of Local Power Elite," American Journal of Sociol0gy, 63, November I957, pp. 290-296. Scott, William A. , "Empirical Assessment of Values and Ideol09ies," American Sociological Review, 24, June I959, pp. 299-3I0. Selvin, Hanan C ., "A Critique of Tests of Significance in Survey Research," American Sociological Review, 23, No. 4, AUgust, I958, pp. 5I9-527. Shils, Edward, "Primordial, Personal, Sacred and Civil Ties," British Journal of Sociol0gy, June I959, pp. I30-I46. Simmel, Arnold, "A Signpost for Research on Fluoridation Conflicts: The Concept of Relative Deprivation," The Journal of Social Issues, XVII, No. 4, I96I, pp. 26-36. Stokes, Donald E. and Warren E . Miller, "Party Government and the Saliency of Congress," Public Opinion Quarterly, XXVI, Winter I962, pp. 53I - 546. Thompson, Wayne E. and John E. Horton, "Political Alienation as a Force in Political Action," Social Forces, March I960, pp. I90-I95. Woodward, Julian L. and Elmo Roper, "Political Activity of American Citizens," The Amefican Political Science Review, XLIV, December I950, pp. 872-875. Unpublished Material Dobkin, David M., "Political Cynicism and Liberalism-Conservatism: Stability and Instability," Unpublished Master's Thesis, Department of Political Science, University of Oregon, I964. Goldrich, Daniel, On the Concept of Politicization, Unpublished paper prepared for the Research Seminar on "Processes of Community Decision-making and Change and Their Influence on Education," University of Oregon, August I963. Sigel, Roberta S . and H. Paul Friesema, "Urban Community Leaders' Knowledge of Public Opinion," Unpublished paper prepared for the American Political Science Association meetings in New York, September I963. ill T‘ IT-.. Tl ‘IIr’ I .I APPENDIX A THE QUESTIONS The data used in this study was only a small part of an extensive 24 page schedule administered by The Institute for Community Studies of the University of Oregon in I959. The following questions are only those used in this study. I3. I5. I6. 6I. How many grades of school have you completed? Grades: I23456789IOIII2 Trade School: I 2 3 4 College: I 2 3 4 Business School: I 2 3 4 Graduate Work: years Within which of the following income categories was your total family income (before taxes for I958)? 0) Under$l,000 4) s4,ooo-s4,999 8) $ro,ooo-$r4,999 r) sumo-$1,999 5) $5,000-55,999 9) 515,ooo-$24,999 2) $2,ooo-$2,999 6) $6,000-$6,999 10) $25,000-and over 3) $3,ooo-$3,999 7) snow-$9,999 II) NA ‘ What is your regular occupation or job? (Please be specific, such as insurance salesman, machinist, housewife, etc.) How often have you seriously discussed local government or community matters during the past year with: Civic or City or Community County Friends Leaders Officials Often Once in a while Not at all I79 66. 79. 83. 85. 86. I80 Has anyone come to you within the past year for advice on what can or should be done in regard to local government or community welfare? 0) Yes I) No Have you taken an active part on any local government or community issue during the past two or three years? 0) Yes 1 ) No Have you attended any meetings or gatherings during the past two or three years in which city government matters were a major subject of consideration? 0) Yes I) No Would you say that: 0) You are very influential I) You are more influential than most people 2) You are about average as far as influence is concerned 3) You have less influence than most people Would you like to be able to have more influence in community affairs than you now have or are you pretty much satisfied with what you have? 0) Would like more influence I) Satisfied with present influence Next we would like to get some information on your relation to a few Specific matters that have come up - in Eugene (Springfield). 89. 90. 9I. 94 . 95 . 96. 97. 98. I8I The first matter is the attempt by some people to attract new industry to Eugene (Springfield). Which statement best applies to you? 89909I92939495969798 0) Haven't heard about it I) Does not matter too much to me 2) Interested but haven't done anything about it 3) Have talked about it with friends or acquaintances 4) Have taken an active part on one side or the other The next matter is annexation of suburban areas. (Check above). The next matter is the traffic and parking problem in Eugene (Springfield). (Check above). The next matter is spending more money on special education pragrams in the public schools (gifted and retarded children, etc.). (Check above). The next matter is whether the water supply should or should not be fluoridated . (Check above). The next matter is starting a local public housing pragram for the aged and the poor. (Check above). The next matter is the city's urban renewal program. (Check above). The next matter is expanding the park system in the EUgene-Springfield metropolitan area. (Check above). The next matter is adding kindergartens to the public school system. (Check above). 99. 0) I) 3) 4) 8) 9) 10) II) 12) IOI. I82 Would you tell us what you feel about the following things or people, whether you strongly approve, approve, are undecided, disapprove, or strongly disapprove, or don't care about it? Strongly Strongly Un- Dis- Dis- Don't Approve Approve decided approve approve Care Urban renewal Bringing new industry to the city Annexation to the city of suburban areas Creating a metro- politan park along the Willamette River Public Housing Fluoridation of the community's water SUPPl)’ Spending more money on special education City-owned parking lots Increasing taxes to provide public kindergartens Which of the following statements do you think best applies to these policy makers? City Officials 0) Do pretty much what the citizens want I) Do what some of the more influential people want 2) Do what they themselves think best I83 I03. Generally speaking, how important are the following groups in making the key decisions on major policies in Eugene (Springfield)? (Check below). Very Not so Important Important 0) the businessmen I) the voters 2) the labor leaders 3) the political parties The following items were used to construct the efficacy, cynicism, and content sensitivity scales. They were on a questionnaire which was left with the respondent on the completion of the interview schedule, and he was asked to fill it out and retum it in the stamped envelope which was also given to him. INSTITUTE FOR COMMUNITY STUDIES The following statements have been given to a large number of people thrOUghout the country. These are all matters of OPINION; there are No right or wrong answers. We simply want to compare the replies made to them by people in this community with replies from people elsewhere . Once again, we would like to stress that your reply to this as well as to the other part of the questionnaire will be completely confidential. PLEASE DO NOT SIGN YOUR NAME . We have a code number which will identify it sufficiently for our purposes. Would you simply put your completed form in the return envelope and mail it back to us as soon as you have completed it. Please read each statement carefully. Then indicate the extent to which you agree or disagree by checking in the appropriate column. Please give your opinion on every statement. Do not worry over individual items. It is your first impression, the immediate "feeling" about each statement, that we want. Thank you very much for your cooperation. We do not like to ask you for so much of your time, but we sincerely believe that our understanding of people and how they live in the cities of Twentieth Century America will be increased by your taking the time to fill out this form. I0. II. 22. I84 CHECK IN ONE OR ANOTHER BOX FOR EACH STATEMENT All in all, it is better to be humble and honest than to be important and dishonest. I don’t blame anyone for trying to grab all he can get in this world . The most important things to me are my duties to my job and to my fel lawman . A person does not need to worry about other peeple if only he looks after himself. Voting is the only way that people like me can have any say about how the government runs things. Agree Agree Agree Disagree Disagree Disagree Strongly Somewhat Slightly Slightly Somewhat Strongly 23. 24. 25. 4I. 42. Sometimes politics and government seem so comp- licated that persons like me can't really under- stand what's going on. People like me don't have any say about what the govern- ment does . I don't think public offic- ials care much what peeple like me think. In order to get nominated , most candidates for political office have to make basic compromises and undesirable commitments. Politicians spend most of their time getting re- elected or re- appointed . I85 Agree Agree Agree Disagree Disagree Disagree Strongly Somewhat Slightly Slightly Somewhat Strongly 43. 45. 46. Money is the most important factor in- fluencing public policies. A large number of city and county poli- ticians are political hacks. People are very frequently manipulated by politicians. Politicians represent the general inter- est more fre- quently than they represent special interests of groups. I86 Agree Agree Strongly Somewhat Agree Disagree Disagree Disagree Slightly Slightly Somewhat Strongly APPENDIX (5 TI 2E SAMPLE AND SAMPLE DENSITIES Selecting the Sample Several sampling techniques were used in sampling followers in the cities and suburbs studied. In Springfield and in all of the surrounding suburbs, a systematic sample of every Kth household was taken as a sample. The listings of households were obtained from utility companies and the Lane County Planning Commission. In Springfield every seventh household was sampled, in the River Road area every 33rd household, in Willakenze every 22nd household, in East Springfield every I0th household, and in Glenwood every 7th household . Once the household was identified, the interviewers were sent out with specific addresses and told to alternate between males and females in interviewing a person in these households who was over 2I years old. If the required gender was unavailable, the interviewer was told to interview the available household member, and then in the next three households he was told to interview persons of the opposite sex . Thus the male-female ratio was maintained. In EUgene some effort was made to stratify the sample on the basis of social and economic class. Area boundaries of region roughly corresponding to neighborhoods were obtained from the Lane County Planning Commission . The regions or census enumeration districts as they were called, were thOUght to be relatively homogeneous as to their social and economic characteristics and had I87 I88 been prepared with the hepe that they would be used by the Bureau of Census for the I960 census. Several small districts were combined with adjacent districts of similar social and economic characteristics. A sample of approximately I3°/o of the blocks within each district were then sampled. Between one and four blocks in each district were thus included in the sample. Once the blocks were chosen, the interviewers were told to start in the Northeast corner of the first black they were given to interview in and take a systematic sampling of the households on that block working clockwise . The interviewers continued this systematic sampling on each succeeding block, counting from the Northeast comer. Approximately one in every 28 households were sampled in Eugene. The individuals interviewed within the households were chosen as before . An extensive analysis of these samples by members of the Institute for Community Studies comparing the samples with known population parameters from the Census of I960 and local party registration roles has shown the Eugene sample to most overrepresent higher educated and older persons. But even for this mmple the differences are slight. Deriving Statistics for the Combined Communities Because the sample densities of the two cities and their suburbs differed, to get a mean of a distribution of attitudes on an issue for the combined com- munities, it is necessary to weight the various means by sample size and sample density. Fortunately roUgh adjustments were made within each community for I89 greater household density in some areas of the communities. In the heart of EUgene, for example, there are fewer individuals living in each household as compared with suburban areas in EUgene . A larger percentage of households were interviewed in low household density areas. Thus the household sampling density is approximately equal to the individual sampling density. The house- hold sampling densities were used in weighting the means for the combined sample . APPENDIX C IDENTIFYING LEADERS AND ASSIGNING STRATA Identifying Actual Leaders Two randomly and independently selected samples of officers of formally organized voluntary associations and elective or appointed officials of local government were used as informants in the process of identifying leaders in the community political process. In each community twenty such persons were selected for the first panel of informants and seventeen for the second panel. At least two people representing each of the following areas were included on each panel: education, municipal government, business and the professions, and civic or service organizations. The remaining nine and twelve members respectively were selected from among the heads of social-welfare, fraternal, special- services, veterans, social, country-club, and religious associations. The first panel of twenty "informants" were asked the following questions: Suppose a major project were before the community that required decisions by a group of leaders that nearly everyone would accept. Which people would you choose, regardless of whether or not you know them personally? In most cities, certain persons are said to be influential "behind the scenes" and to have a lot to say about pragrams that are planned, and projects and issues that come up around town . What persons in Eugene (Springfield) are influential in this way. or are influential in being able to stop particular community policies? Are there any other people with whom these leaders work that have not been named so far and should be included in a list of community leaders? Persons mentioned by one or more of these panel members were listed in I90 I9I alphabetical order and shown to members of the second panel who were asked to do the following: We have talked to a number of people in Eugene (Springfield) who have given us a list of peOple whom they consider to be important in community policy-making. We would like you to look at this list and indicate which of these people you would consider to be among the twenty or twenty-five most important peOple in this regard . By most important people we mean peopTe—who can get a major policy or project adopted in Eugene (Springfield). You may feel free to add anyone whom you think is important in community policy-making who is not on this list. Persons who received two or more votes in this second panel were designated leaders. The Index of Social Status The index of social status ranges from a lowest social status category of zero (0) to the highest of three (3). Each of the contributing items of information, education, income, and job status, were dichotomized into a high category assigned the value one (I) and a low category assigned the value zero (0). The cutting points for each of these dichotomies was the median of the distribution of the entire random sample of followers on that particular item. NA and DK responses were, of course, excluded from these computations. The resulting cutting points were: Education -- between eleven and twelve years of school completed Income--between $5,000 and $5,000 and above Job Status--between codes I, 2, and 3 (white collar) and codes 0, 4, 5, 6, 7, and 8 (blue collar). The individual's social status score then was the sum of his status scores on each of these items of information. If a respondent's occupation was either housewife, or his response was DK or NA and information of the spouse was given, this I92 information was used to assign social status. Adjustments upward were made then when the respondent and his or her spouse differed in education or job status. The Index of General Political Involvement The index of general political involvement also ranges from a low participation category of zero (0) to a high participation category of three (3) . Three items of information were used in assigning the individual to a category: I) discussion of local government with friends, leaders in the community, or city and county officials; 2) attendance at meetings concerned with city government issues; and 3) taking an active part in a local government or com- munity issue. If a respondent indicated that he "once in a while" or "often" discussed local government matters with either friends, leaders in the community, or city or county officials, he was assigned a score of one (I) on this variable. If not, he was assigned a score of zero (0). On the other two variables, an answer of "yes" was assigned a score of one (I) and "no" was assigned a score of zero (0) . The individual's general political involvement stratum‘was the sum of his involve- ment scores on these three items of information . The Index of Political Cynicism. The index of political cynicism varies from a highly cynical score of zero (0) to a trusting score of six (6) . The attitudes expressed by the respondents to questions 4I through 46 in the mail-back questionnaire were used in assigning this overall score. For each of these items the responses of the respondents were I93 dichotomized to assign a score of zero or one. The overall score then is the sum of these individual item scores. Scores on each item of zero, indicating cynicism, were assigned to respondents who gave the following answers to these individual items: item 4i -- agree strongly item 42 - agree strongly and agree somewhat item 43 - agree strongly and agree somewhat item 44 -- agree strongly, agree somewhat, and agree slightly item 45 -- agree strongly, agree somewhat, and agree slightly item 46 -- disagree strongly, disagree somewhat, disagree slightly, agree slightly, and agree somewhat. For an extensive discussion of this variable see Robert E. Agger, Marshall Gold- stein, and Stanley Pearl, "Political Cynicism: Measurement and Meaning" in The Journal of Politics, August I96I . The Index of Political Efficacy The index of political efficacy varies between an inefficacious score of four (4) and an efficacious score of zero (0) . As with all of the preceding indices, the overall score is the sum of the dichotomized individual item scores. For each item in this index the respondent was assigned a score of one (I) for an item if he disagreed to the item, no matter whether he disagreed only slightly or more strongly. See Angus Campbell, Gerald Gurin, and Warren E. Miller, The Voter Decides, page I82 for an extensive discussion of this variable . APPENDIX D CONTENT SENSITIVITY The use of the survey questionnaire in the study of human behavior necessitates the assumption that persons answering the questionnaire are responding to the content of the items on that questionnaire. But a disturbing phenomenon of persons responding to questionnaires by other than the content of the items has frequently been noted .' The most frequently discovered "response set" is that of acquiescence or yea-saying . Persons responding in this way agree to all items on which they are asked their opinion. I am not concerned herewith the questions of what types of individuals are vauiescent or why they are vauiescent. I am concerned with whether non-content sensitivity is common to my sample and whether its existence affects the relationships noted in my analysis. Two discovered relationships could be affected by acquiescence or yea- saying. I prefer to call the phenomenon, yea-saying, thereby avoiding the implication of some psychol09ical purpose being served by always answering agreeably. The first relationship possibly affected is the favorable attitude bias on the issues noted among community leaders. If yea-saying were characteristic 'See Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E . Stakes, The American Voter, (New York, I960), pp. 5I2-5I5; and Loren J . Chapman and DonaTd T. Campbell, "The Effect of Acquiescence Response-Set Upon Relationships Among the F Scale, Ethnocentrism, and Intelligence," Sociometry, 22 (June I959), pp. I53-I6I . I94 I95 of a large segment of the populace, the relationship would be stronger than noted as yea-saying would make the public appear more favorable than is actually the case. Similarly, the relationship between increasing social status and increasing favorability would be stronger than noted if yea-sayers were common and primarily from lower social status levels as research has shown to be the case. Notably, the existence of the phenomenon would obscure the strength of these relation- ships. Thus controlling for non-content sensitivity if it were found would improve the relationships. The following four items were used to identify non-content sensitive persons: I . All in all, it is better to be humble and honest than to be important and dishonest. 2. I don't blame on one for t in to rob all he can et in this world. Y '7 9 9 9 3. The most important things to me are my duties to my job and to my fellowman. 4. A person does not need to worry about other people if only he looks after himself. . These items were taken from an article by Robert E. Agger et al . who adopted them from Robert Christie.2 Item I is thought to contradict item 2 while item 3 contradicts item 4. Within the limitations of measuring instrument error, a person who was sensitive to the content of the items would be expected to not be able to agree or disagree with both. items in these two sets of items. 2Robert E. Agger, Marshall N . Goldstein, and Stanley A. Pearl, "Political Cynicism: Measurement and Meaning," Journal of Politics, 23 (AUgust I96I), pp. 503-506. I96 With dichotomized responses of agree-disagree, sixteen patterns of response are possible for these four items. Using "A" for agree and "D" for disagree, the following four patterns are consistent with the contradictions in logic of the two sets of items and thus indicate content sensitivity: A D A D D A D A D A A D A D D A . All other patterns indicate some degree of non-content sensitivity. The response of yea-sayers is A A A A while that of nay-sayers is D D D D. ' Table D-I shows the number of persons identified in each category. Table D-I: Number and percentage of sample falling into each category of content sensitivity . ’ Number Percent of Sample Yeo-sayers (4 Agrees) 46 6.2 (3 Agrees) I38 I8.6 Content Sensitive 344 46.5 (3 Disagrees) I66 , 22.4 Nay-sayers (4 Disagrees) 8 LI Non-content sensitive 39 5. 3 (AADD a. DDAA) Table D-2 shows the means of the nine issues of the "yea-sayers, the nay-sayers, and the combined publics. Yea-sayers are more approving than the public on I97 Table D-2: Comparison of yea-sayers' and nay-sayers' responses to the nine issues with those of the entire public. Issues Yea-sayers Public Nay-sayers Attracting Industry .638 .543 .7I4 Annexation I .428 I .389 I .833 Parking Lots 1 .382 1.460 . 1.500 Special Education I .I08 I . I69 I. I42 Fluoridation I .972 I .77I I .428 Public Housing I .352 I .6I4 I .428 Urban Renewal I .392 I .652 I .666 Metropolitan Park I .205 I .246 I .500 Public Kindergartens I .972 2.000 I .857 six of the nine issues, and nay-sayers are more disapproving than the public on five of the nine issues. The relationships are in the predicted direction but are very weak . This response set is of little importance in this study. The relation- ships among classes of content sensitivity are very weak; and to the degree that they do exist, the discovered relationships among leaders and followers are strengthened, not weakened .