1-' u .5 4.. if .-, ... . ‘ in“ i». V1149» .Vnflu . haw". :a. ..¥.......tz..: '9 o -. IFnQ&.V|’-‘.! 3 n . ‘ 995715.- 2.5.1. J! .zr‘tI’ +‘ 2-». 2.2.} y 51: i.cv|. p .. .\.I$t!:.7 : a (£7.43? A; . . Gaul: . out: I. ‘ A ‘ ll. 1:.éitllhtoh n. or :1. 4:... 9:?- ' ‘Ith‘h-‘OAS‘Q‘ ; . 7... . Ail. .3 :II...’ IOJII OIIJ. .l, I..|I.v07a ’. Ill! Ol.< A. it". :0, 0!..‘5‘: s‘lf.‘ 1:01,. I} .vs‘t o~r1¢l ‘In'zfl‘z I\ Y: e. 3...“.IJI II . u ‘If.’ A “Ht-251$ dlll‘l‘ilii'l‘l’ll‘l'il WilliIiiiil'liflliiilliiliflliflllifil L 3 1293 00897 7591 This is to certify that the dissertation entitled THE IMPACT OF LABOR LEADER PRONOUNCEMENTS ON THE VOTING BEHAVIOR AND ISSUE OPINIONS OF UNION MEMBERS presented by Christine Lee Lepo has been accepted towards fulfillment of the requirements for Ph.D. degree in Social Science // Major profissor Date 8/ 3 la MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 * LIBRARY Michigan State J L— UnIverslty PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. ———__—_——_———_'—j DATE DUE DATE DUE DATE DUE . i r—T—fi __fi___i if MSU Is An Affirmative Action/Equal Opportunity Institution c:\cIrchm3-p.1 - THE IMPACT OF LABOR LEADER PRONOUNCEHENTS ON THE VOTING BEHAVIOR AND ISSUE OPINIONS OP UNION MEMBERS BY Christine L. Lepo A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY College of Social Science 1992 ABSTRACT THE IMPACT OF LABOR LEADER PRONOUNCEMENTS ON THE VOTING BEHAVIOR AND ISSUE OPINIONS OF UNION MEMBERS BY Christine L. Lepo Voting behavior and issue opinions of union members are significant components of strategy formation and evaluation regarding the political agenda of the American Federation of Labor (AFL-CIO) and the Committee on Political Education (COPE). This study examines the impact of labor leader pronouncements on the voting behavior and issue opinions of the rank. and file. .Appropriate :models regarding ‘these components must include criteria elemental to research focusing on decision making in national elections. Therefore, the relationship of these criteria (gender, age, race, religion, education, region, income, party identification and incumbency) to Presidential and Congressional voting behavior and issue opinion were examined. Four models were tested, first on a descriptive and then on an inferential basis. Each national election vote criterion was coded in a way which facilitates analysis of odds ratios, by dichotomizing it into "vote for (AFL-CIO/COPE) endorsed candidate" or "vote for non-endorsed candidate." Each issue opinion criterion.was coded as "favor AFL-CIO/COPE pronouncement" or "oppose AFL-CIO/COPE pronouncement." Results indicate that union status, although significant in bivariate analyses of Presidential and Congressional election voting behavior, is not significant in logistic regression analyses. Once control variables are added, the odds of union members voting for the endorsed Presidential and Congressional candidates are not significantly higher than are the odds of non-members voting for endorsed candidates. Implications of results garnered by the bivariate and logistic regression analyses are discussed both in terms of their implications for labor leaders and as regards the appropriateness of using odds ratios as the basis of research into the voting behavior of union members. Copyright by CHRISTINE L. LEPO 1992 This dissertation is dedicated to B. J., Dave, Bill, and Mark for their unconditional love and genuine interest in my success. And to Rocky and Jennyfer, that they may finally find peace. ACKNOWLEDGEMENTS I wish to express my profound appreciation to the faculty and staff of the School of Labor and Industrial Relations. Because of my positive educational experience in this department, I remain committed to Michigan State University and its academic mission. Professor Richard Block, Director of the School of Labor and Industrial Relations, and dissertation chair, has been a tremendous guide throughout my program of study. He has encouraged research, invigorated my interest in labor relations, added depth of insight and been a true teacher. I am also appreciative of the time and effort Professor Sandra Gleason has dedicated to my program. She is among the finest of instructors, and.is driven to assist all students in excelling to their highest potential. She was of immeasurable assistance in keeping me on track as each new academic hurdle presented itself. Professors Jack Stieber and Michael Moore have greatly increased my breadth of knowledge in their respective fields. Their standards and perspectives are reflected in the courses I currently teach. I! am grateful for the counselling and direction given to me by Professor Moore during the comprehensive and dissertation phases of my program. vi Professor Claude WYnn has earned my respect and profound appreciation for his patience, interest, caring and accommodation as I completed requirements toward my Ph.D. degree. He continues to be a mentor in the true sense of the word. I wish to extend thanks to Jill and Jack Connolly. They were never too busy to type or re-type this manuscript, lending helpful suggestions regarding its format whenever appropriate. I wish to thank Paul and Pat Mahoney for their friendship and financial help during the writing of this dissertation. I am forever indebted to my parents, Betty and Bill Horn and to my brother David Mahoney. Their genuine caring and encouragement have been invaluable. They have been the brightest of lights in my life. Their love, humor and companionship have guided my steps along a variety of paths. Finally, I wish to express my appreciation to Mark, for entering my life, making it profoundly richer and teaching me what compassion for mankind is all about. vii TABLE OF CONTENTS CHAPTER I . . . . . . . . . . . . . . INTRODUCTION . . . . . . . LIMITATIONS OF PRIOR RESEARCH . . . JUSTIFICATION FOR FURTHER RESEARCH UNION THEORY AND AMERICAN UNION POLI RESEARCH OBJECTIVE . . . . . . . . DESCRIPTION OF CHAPTERS . . . . . . I S 0080000 0000000 CHAPTER II . . . . . . . . . . . . LITERATURE REVIEW . . . . . . . AFL-CIO and COPE ENDORSEMENTS . . AFL-CIO and COPE ISSUES OPINION PR ISSUES IDENTIFIED . . . . . . . . Economic Policy Issues . Foreign Policy Issues . . Social Policy Issues . Labor Issues . . . . BIOGRAPHICAL CHARACTERISTICS Union Status . . . . . . . Relationship of union member political ON :......§..: behavior and leadership pronouncements Gender . . Age . . . . Race . . . Religion . Education . Region . . Income . . Party Identi Incumbency ati 16C Ii CHAPTER III . . . . . . . . . . . . . METHODOLOGY . . . . . . . . . . DATA . . . . . . . . . . . . . . The Sample and Sample Design STUDY POPULATION . . . MULTI- STAGE AREA PROBABILITY SAMPL DESIGN . . . . . . . . . THEORETICAL EXPECTATIONS Dependent Variables . Independent Variables Union Status . . ooeeetljeoeeo viii EMENTS P‘H bwmeI-‘H 15 15 15 23 26 26 31 33 38 43 44 51 53 55 57 59 61 62 63 64 69 72 72 73 75 75 76 76 77 78 78 Gender . 78 Age . . . . . . . . . . . . . . . . . . . . 78 Race . . . . . . . . . . . . . . . . . . . 79 Religion . . . . . . . . . . . . . . . . . 79 Education . . . . . . . . . . . . . . . . . 79 Region . . . . . . . . . . . . . . . . . . 80 Income . . . . . . . . . . . . . . . . . 80 Party Identific ti 1on . . . . . . . . . . . 8O Incumbency . . . . . . . . . . . . . . . . 80 Research Hypotheses . . . . . . . . . . . . . . 81 ANALYSIS . . . . . . . . . . . . . . . . . . . . . 82 SUMMARY . . . . . . . . . . . . . . . . . . . . . . 85 CHAPTER IV . . . . . . . . . . . . . . . . . . . . . . 87 Actual Vote . . . . . . . . . . . . . . . . . . . 87 Actual Vote Plus Intended Vote . . . . . . . . . . 92 Endorsed House Candidate . . . . . . . . . . . . . 96 Three Voting Models . . . . . . . . . . . . . . . . 100 Issues . . . . . . . . . . . . . . . . . . . . . . 106 CHAPTER V . . . . . . . . . . . . . . . . . . . . . . . 123 SUMMARY, CONCLUSIONS, AND IMPLICATIONS . . . . . . 123 Summary . . . . . . . . . . . . . . . . . . . . 124 Presidential vote . . . . . . . . . . . . . . . 124 Congressional vote . . . . . . . . . . . . . . 125 Issues . . . . . . . . . . . . . . . . . . . 126 CONCLUSIONS AND IMPLICATIONS . . . . . . . . . . 127 Suggestions for Further Research . . . . . . . 129 Suggestions for the Practitioner . . . . . . . 131 APPENDIX O O O O O O O O O O O O O O O O O O O O O O O 136 LIST OF REFERENCES O O O O O O O O O O O O O O O O O O 144 ix Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table LIST OF TABLES Actual Vote in 1988 Presidential Elections . . . . . . . . . . . . . . Logistic Regression of Actual Support for the Union Endorsed Presidential Candidate . . . . . . . . . . . Intended Vote in 1988 Presidential Elections . . . . Logistic Regression of Actual plus Intended Support for the Union Endorsed Presidential Candidate . . . Actual Vote in 1988 House Elections . Logistic Regression of Support for the Union Endorsed Candidate for the House of Representatives . . . Combined Logit Results of Individual Voting Behavior . . . . . . . . . Central America . Child Care . . . Civil Rights . . Defense Spending Drugs . Education . Environment . . Food Stamps . . . . Healthcare Insurance Homelessness . . . . Import Quotas . . . . Preferential Hiring . Research Funding, AIDS Social Security . . . . Strong Military . . . . Unemployment . . . . . 87 90 93 94 96 98 102 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 CHAPTER I INTRODUCTION Many studies have attempted to politically situate labor. The vast majority of them reveal a propensity on the part of the rank and file to vote Democratic. Those studies which have focused on presidential elections, most notably Delaney, et a1. (1990), Parent, et a1. (1987), Wattenberg (1987), and Norpath (1987) , suggest that while the Democratic "union vote" may have dropped.off somewhat, it remains intact. Because the President of the United States is empowered to appoint members to the National Labor Relations Board (NLRB), it is especially important to organized labor that a Democratic president be elected. Democratically controlled boards are traditionally more likely than Republican controlled boards to vote in ways which support labor. NLRB interpretations of law are often found to be "more favorable to business when Republicans are in power and more favorable to labor when Democrats are in power" (Freeman. and. Medoff, 1984:41). Additionally, as evidenced in the negative public opinion toward unions which followed President Reagan's handling of the 1981 Professional Air Traffic Controllers Organization strike, presidents influence the tone of public opinion toward organized labor. 2 Typically, it is during Democratic presidencies that opinion and legislation have been more favorable for unions. Those studies which have focused on Congressional elections, most notably Delaney, et a1. (1990) and Wilson (1979), also support the notion that labor tends to vote the Democratic ticket. A Democrat-controlled Congress is advantageous for organized labor, since, as Freeman and Medoff (1984:197) point out "...union-favored legislation has done well when Democrats are in power but poorly when Republicans are in power." The AFL-CIO's Committee on Political Education (COPE) primarily endorses Democratic candidates for Senate and Congress (Greenstone 1977). COPE and the AFL-CIO Executive Committee actively support issues and legislation which they deem advantageous for organized labor. And the AFL-CIO endorsed the Democratic candidate for President in 1984, 1988, and is committed to do so again in 1992. Because they actively seek, via promulgation of their agenda, to garner support through voting, union leaders must be aware of the effectiveness of their major strategies. Whether or not the voting behavior of union members reflect the endorsements of union leadership should be of major importance to labor leaders. Voting behavior of union members, therefore, is worthy of scholarly analysis. The relationship between endorsements and issue opinions of labor leaders and the political behavior and attitudes of rank and file union members is also worthy of analysis. 3 Questions surrounding the influence of organized labor leadership on its members are abundant in Industrial Relations and Sociological literature (Masters and Delaney, 1987; Blume, 1973; Harwood, 1981; Hudson and Rosen, 1954; Heldman and Knight, 1980; Patton and Marrone, 1984; Patton, Marrone and Hindman, 1986; Ra, 1978; Form, 1982 and 1983; Lyons, 1969; Nesbitt, 1978; Wilson, 1979). The study at hand advances one previously studied and two relatively unexplored ideas. The first major component explored, one which has received some attention in prior research, is the relationship between endorsements by labor leaders, and voting behavior of rank and file union members in presidential elections. The second major component, which has received little research attention, is the relationship between voting in Congressional elections and COPE/AFL-CIO endorsements. The third component seeks to determine if there are significant differences in opinions of unionized and non-unionized people regarding political issues. A methodological component of this study which should garner meaningful results is its use of political party identification as a control variable to insure that differences found are due to union membership rather than to party identification. This study should be particularly noteworthy for academics and practioners (labor lobbyists, political analysts, political candidates, and labor leaders), since it precedes the November 1992 Presidential and Congressional elections. In 1991 Democratic presidential candidates vied for the 4 support and endorsement of organized labor. Candidates Bob Kerry, Jerry Brown, Tom Harkin, Paul Tsongus, Douglas Wilder and Bill Clinton appealed to labor's 700 delegates at an October labor convention. Shribman (1991) reported that these candidates emerged from the chute hungry for labor's blessing, and with an emphasis on issues. These issues included health care and comprehensive health insurance, tax cuts for the middle class, limits on free trade with Mexico, increases in the minimum wage, extended unemployment benefits, and decreased foreign aid. While in 1984 the AFL-CIO presidential election endorsement of Walter Mondale was looked upon by political analysts as the proverbial "kiss of death," the candidates of 1992 were passionate in their pleas for labor's approval. And in.May 1992, Democratic presidential candidate Bill Clinton was nominated for endorsement in an Executive Council meeting. An official endorsement is forthcoming. LIMITATIONS OF PRIOR RESEARCH The effect of leadership pronouncements on member's attitudes toward social and economic issues is reflected in only a paltry share of Political Science and Industrial Relations literature. Yet three primary vehicles for articulating and disseminating organized labor's viewpoint (AFL-CIO Report on Congress, AFL-CIO Convention Proceedings and Congressional Quarterly), emphasize issues. The research at hand seeks to redress this deficiency. The importance of issues as a major factor in voting has 5 not been ignored. The timeliness of information about issues has been viewed from the perspective of "impression-driven models" (Lodge, et a1. 1989), as has the relationship between political party identification and liberal versus conservative attitude toward issues (Rappaport, et a1. 1990), the function of single issues in voter choice for a particular candidate (Stone and Abromcwitz, 1983), and the inappropriate emphasis of nomination activities on a given candidate's issue preference and.ideology (Polsby, 1983; Soule and.Clarke, 1970; Roback, 1975; Shafer, 1988 c.f. Rappaport, et a1. 1990). A second area of research deficiency surrounds the issue of rank and file voting behavior in Congressional elections. Few studies analyze union member voting behavior as it relates to candidate preference in Congressional elections. The study at hand serves to ameliorate this deficiency by investigating the relationship between the AFL-CIO/COPE 1988 endorsements of Congressional candidates and union member voting behavior. From Masters and Delaney (1987: 346-350) , whose review of literature regarding union political activities is an invaluable source for bibliographic reference, three additional research deficiencies are suggested. The study at hand seeks to redress these deficiencies in one way or another, as follows: 1. Generalizability: Masters and Delaney suggest that generalizability of research findings is impaired for a number of reasons. Studies often utilize samples drawn from one union only, or, at best, from a small set of unions 6 representative of just one region of the U.S. The study at hand utilizes a survey whose population is comprised of U.S. citizens of voting age, who reside in the 48 contiguous United States. 2. Methodology: One methodological flaw, argue Masters and Delaney is one which is a potential hinderance to validity: the collection of data from rank and file members by either their union or a local council of the AFL-CIO. The accuracy of such data may be questionable, based upon the assumption that at least some respondents will answer according to what the survey instrument's administrator prefers. The study at hand is administered not by a union or collection of unions, but by a major university. Another methodological failing disclosed by Masters and Delaney concerns the absence of comparison-group data. Because the National Election Survey (NES) randomly samples the U.S. voting population, the opportunity for making comparisons is rich. 3. Unexamined Issues: The research at hand seeks to address at least one of the unexamined issues brought forth by Masters and Delaney: "the possibility that union strategic choices affect individual members' behavior." Major sources for articulation and dissemination of AFL-CIO pronouncements include (a) COPE ratings of candidates, and (b) AFL-CIO/COPE issue opinions, as presented in the AEL-QIQ Convention Proceedings. the WW and the Qgngressignal_gnartezlie§. A synthesis of information taken 7 from these publications is compared to voting behavior and political attitudes, in the study at hand. JUSTIFICATION FOR FURTHER RESEARCH Should the results of the study at hand indicate a significant, positive correlation between union membership and voting controlling for party identification, it would serve as an indication that labor leader's endorsements are indeed accomplishing the desired goal. Likewise, if the study indicates an ability to discriminate between union and non- union members (the "union vote" or the "union stand on issues"), the signal to the AFL-CIO, COPE and individual unions is that rank and file members are "buying into" their message. Implications of the failure to discriminate or to find these relationships, may indicate one of two things: first, that there may be a "spillover effect" of labor's agenda onto the general population such that non-members as well as members are accepting labor's point of view or, second, that the approach of labor leaders to educating members as to candidates and issues is not a viable one. Ample justification for a study such as this exists. By addressing a number of areas in which prior research is deficient, and by utilizing appropriate methodology, this research has the potential for contribution to academics and practitioners alike, especially when conducted at such a crucial time for the Democratic party: that party which traditionally aligns itself with organized labor. 8 UNION THEORY AND AMERICAN UNION POLITICS The alignment between labor and liberal reform and labor and the Democratic party was evident for the first time on a national basis, during the 1936 presidential election, when President Roosevelt was re-elected to office. The Congress of Industrial Organizations (CIO) had departed from the strict Gompersonian "reward of friends, punishment of enemies, " entering into the political arena during the earlier part of the decade. After the 1936 election, John L. Lewis made evident labor's expectation that promises made to them during pre-election support, be kept. That same pre-election support-post-election expectation relationship is a significant component of labor politics today. Wilson (1979) describes the candidate endorsement scheme of COPE as first, selecting a group of issues deemed important to labor; second, rating'each member of Congress and the Senate with a mark for each time he or she votes the COPE platfonm on these issues (the issues ratings then tally to somewhere between 0% and 100% for the year), and finally, discussing which candidates to vote for based in a large part upon these ratings. The COPE record is paramount in the decision as to which candidates receive financial help and endorsements from.organized labor. Wilson points out that in discussions prior to the 1976 elections, the very first input about favorability of candidates for endorsement, even in Southern states, was, in fact, their COPE rating. What implications does this political process have for the 9 rank and file? COPE issues and COPE rating scales are determined at a national level. Rank and file members, already removed from direct determination of labor's issue criteria, are also removed from both the rating of political incumbents (most of whom come up for re-election), and from the endorsement decisions derived from these ratings. ch reflectant then are issue stands and endorsements of rank and file.members?‘ The study at hand seeks to answer this question by posing six hypotheses regarding the impact of endorsements and issues pronouncements on union membership. Regardless of the final outcome of this relationship, one would have to call into question the appropriateness of this political approach for organized labor. This approach is discussed by Wilson (1979:24), who attended COPE meetings which had as their agenda discussion of candidate endorsements prior to the 1976 national elections. The approach in constructing scales which ultimately lead to endorsements he tells us, can only be "imagined." One approach might be to remain within the Gompersonian framework by concentrating on those issues determined to be labor's own, such as abolition of 14(b). A second approach might be the addition of issues which.would.benefit all working persons, such as unemployment and National Health insurance. The third approach would be to devise issue opinions regarding' social justice, such as womens' rights, preferential hiring and subsidized housing. Between the 1984 and 1988 elections, all three approaches were used.by the.AFL-CIO and COPE. Stands on these issues are 10 presented on pages 23-39. Just as Wilson found the issue stands immediately preceding the 1976 elections to be "liberal" when it came to expanding Federal Government welfare service provisions, civil rights and control of the economy, so are the majority of those positions taken.prior to the 1988 elections. Since no direct feedback is attained from the rank and file, are these issue stands and subsequent endorsements reflectant of the viewpoints of the few, rather than of the many? Michels (1915) theory of organization suggests that large organizations such as the AFL-CIO are often guilty in this regard. He cites trade unions as organizations which develop bureaucratic structures, are hierarchically organized and allot almost exclusive control of power to officers. We can only imagine the magnitude of effort, coordination and expense it would take for the AFL-CIO and COPE to gather and then act upon the opinions of the rank and file. Such administrative difficulties, argues Michel, are inevitable in large organizations and that a price therefore must be paid. That price is paid by rank and file members by way of lost influence. Leadership, in essence, has complete access to and control over resources such as information, communication.and political skill. Leaders, in Michel's opinion, become a power elite whose perspectives and interests emanate from their own privileged positions. These perspectives and interests may very well then not reflect the will and interest of the rank and file, but rather of its leaders. 11 Part and parcel of the relationship between leaders and the rank and file will be according to this theory, a continuous opposition of leaders toward other large ruling elements of society, such as business and the government. An antagonistic relationship with business might, in fact, serve to quell the masses even though their opinions are not regarded, in that at least their leaders stand opposed to a common enemy. Michel goes so far as to say that it is upon self-interest that decisions are made and that the driving force behind these decisions is the maintenance of leaders' privilege and power. Lipset (1962:35) suggests that unions within the AFL-CIO are controlled by what he calls "self- perpetuating oligarchies." If, indeed, individual rank and file opinion is in.a large part disregarded, is there any justification for doing so? It could be argued that an organization with a common ideology or collective sentiment is well represented by leaders committed to these commonalities. The "immense social movement" (Touraine et al. 1983) of Poland's Solidarity and France's Confederation Generale are both highly ideological by nature. Both Solidarity and the Confederation fall within the framework of social movements based upon a strong ideology (Zurcher and Snow 1981). The .American labor' movement, in, sharp) contrast, has evolved within a framework of business unionism and is void of dominant class ideology. Goals and ideologies of the U.S. union movement have transformed. Gone are the loftier goals 12 of the Knights of Labor, replaced by "bread and butter" goals and an "individualist" sentiment (Kessler-Harris 1987). The individualist sentiments of U.S. labor may not be reflected in issues pronouncements and subsequent endorsements which deviate from the market unionism which has developed in response to individualist sentiments. Without collective sentiment, how can leaders assume collective issue opinions and goals? .And within.a business unionism framework, is there room for pronouncements which take liberal stands on social justice issues? Feree and Miller (1985) suggest that in Western society, personal ability and individual effort determine social outcomes. The instrumental (bread and butter) framework of U.S. unionism tends to abrogate solidaristic ties (Kessler- Harris 1987, Dubofsky 1987). According to Kessler-Harris, workers by the 1960's had turned away from collective interests almost entirely "...disassociating themselves from the collective spirit of unionism" (34) , concentrating instead on individual orientations. The collective consciousness is gone and at best, there remains what Kessler-Harris calls an illusionary sense of community; Gone, too, is labors' cautious involvement in the political process--a process viewed as competition for worker loyalties (Rogin 1962). The above sentiments are driven by what Perlman (1928) would call a consciousness on the part of workers, of job scarcity. Perlman (1928) questions the appropriateness of an "...idealistic readiness on the part of the individual to 13 offer... sacrifices for the group as a whole," (Larson and Nissen:168) when the U.S. mode is purported to be one of business unionism. Kerr, et a1. put the argument succinctly: in an industrializing society, there is no monolithic solidarity of workers, nor political uniformity. Without what would seem in light of the above, to be an appropriate political methodology, can we expect to find a significant positive relationship between labor leader pronouncements or endorsements and the voting behavior and political issue opinions of the rank and file? It would seem unlikely. And even though bivariate analysis might show that a higher proportion of union members than non-members voted Democratic in 1988, control for party identification may diminish the significance of the relationship. It is a- primary objective of this research to measure a direct effect between member political behavior and leader endorsements so that the measure will not merely be that of union status to the Democratic party. RESEARCH OBJECTIVE This research seeks first to determine whether or not we may justifiably discriminate voting behavior in presidential and Congressional elections on the basis of union status, controlling for political party identification (as a Democrat, Republican or Independent) and.a number of other correlates of voting behavior. Second, this research probes the impact of union membership on opinions in regard to issues on which the 14 AFL-CIO/COPE have stated the union position. The following criteria are included. in ‘these analyses *whose «dependent variable is presidential vote: (1) union status, (2) gender, (3) age, (4) race, (5) religion, (6) education, (7) occupation, (8) region, (9) income, (10) party identification for the analysis whose dependent variabLe is Congressional vote, incumbency is added to the above list of independent variables. These variables are consistent with those found in both Political Science and Industrial Relations literature as potentially having impacts on voting behavior and opinion holding. DESCRIPTION OF CHAPTERS Chapter II recapitulates extant and traditional literature regarding the political behavior, party preference and political attitudes, voting behavior in presidential elections of rank and file union members, and the role of the AFL-CIO and COPE in candidate endorsements and issue opinions. Chapter III sets forth the major hypotheses to be tested, the data, and methodologies utilized in the research. Chapter IV presents a discussion of the results of the quantitative analyses performed, both by way of descriptive statistics and as pertains to the three major hypotheses, as stated. The dissertation concludes with Chapter V, a summary of research implications and presentation of what the research has added to knowledge in this field, in addition to suggestions for further research. CHAPTER II LITERATURE REVIEW Criteria which affect voting behavior and political issues opinion are examined in this chapter in terms of relevancy to the study and justification of their use based upon prior research. Before discussing literature pertaining to predictor variables, however, justification for using COPE/AFL-CIO endorsements and issue opinions as a measure of labor leader pronouncements is reviewed. AFL-CIO AND COPE ENDORSEMENTS The View taken of the AFL-CIO in this research is that it functions as a voice for national and affiliated local unions and their members. When Fossum (1992) takes this View, he finds politics to be an important component of that voice. The vehicle for political activity within the AFL-CIO is the committee on political education (COPE), which serves to "evaluate legislative records of federal, state and local candidates and WW membership" (117). Lobbying as well as political activity, argues Fossum, are major activities of the AFL-CIO and state and local central bodies. Positions taken by locals in 15 16 national elections must be consistent with those taken by the AFL-CIo.1 Because the AFL-CIO and COPE function, as Fossum attests, as vehicles to provide information to membership, their political stand on issues (found in the Congressional Quarrgrly, the Reporr on Congress, and the anygnrign W) and their endorsement of political candidates are utilized as background material for the dependent variables in analyses contained in this research: presidential vote (for endorsed candidate), Congressional vote (for endorsed candidates), and issue opinions (as related to issue stands taken by the leadership of the union). According to Wilson (1979), COPE is a federation "nominally separated" from the AFL-CIO, empowered to endorse presidential and vice presidential candidates. State and local COPEs reserve the right to endorse all other political candidates. Both levels of COPE endorsements are used in the research at hand. An important COPE strategy is the presentation of organized labor's stand on issues. Issues covered by COPE vary from year to year and from election to election, according to labor's agenda. Normally, social, economic and foreign affairs issues are emphasized over union issues. This emphasis is one way that organized labor is able to advance issues of importance to all working people, not just of those 1cof- Bules_Q2xsrning_A:§.§tate.§entral_ngdie§. Publication 12 (Washington, D.C.: A-C, 1973), p. 21. 17 who are organized. COPE takes a more liberal than conservative stand on most issues. COPE promulgates labor's agenda via the publication of COPE ratings of public officials, leaflets, canvassers, and telephone banks (c.f. Congressional_3e22rd S 3276 (16 March 1971))- The function of COPE records is articulated by Masters and Delaney (1987:223) COPE records are a widely used baro- meter of labor's political success, and we rely on them to show general trends in the political support given to unions. Rehmus, et a1. (1978) believe COPE to be a "sustaining force" (188) in.a.number of labor's political activities. COPE is involved in precinct organization, ratings of candidates' issue opinions, voter registration for union members and their families, fund raising for pro-labor candidates, and making “g sg It‘ll): : £10. gr, £4177:: 5w; : s_ ,=_=-3= egg nnininnn. The COPE index reflects issues relevant for all working persons, not just those who are unionized. Congressional roll call votes included in the index are limited to issues which have been targeted for lobbying. "Labor" and "social" issues are scored separately from one another. Those issues determined to be of greatest importance are weighted more heavily by documentation of several roll calls. A roll call vote may be selected over others by labor leaders and their lobbyists if it has been deleterious to labors' agenda (Fowler 1982). Funds for COPE are solicited.on a voluntary basis from affiliated union members. Sulzner 18 (1972) explains that COPE was, in fact, created.as a political adjunct for the m of garnering voluntary funds from union members. The study at hand addresses the question as to whether or not union member voting behavior reflects AFL-CIO/COPE endorsements. Patton and Marrone (1984) conducted a similar analysis. Their study examined the impact of labor endorsements on the 1980 presidential vote. They found that 53 percent of union members sampled were aware of whom the endorsed candidate was, while 55 percent of the general population knew, revealing no significant ability to discriminate between. the two groups as to awareness of endorsements. The impact of the AFL-CIO/COPE endorsement for Mondale that year was seen to have only a small positive effect on union member vote. Wilson (1979) did not venture a guess as to the impact of COPE on the voting behavior of union members in presidential elections. While Jimmy Carter, for instance, was "energetically" backed by COPE, it would be impossible, he concludes, to measure the magnitude of AFL-CIO/COPE influence. He argues, however, that COPE does impact voter turnout and registration. Delaney, at al. (1990) examined the correlation between COPE candidate endorsements and support for these candidates by union.members and their families. Their research utilized data from 'the 1978 Senate, House of Representative and Gubernatorial races. Results reveal that union members and 19 their families tended to vote for endorsed candidates to a greater extent than did non-members and those living in households where there are no union members. These results, the authors point out, reaffirm an AF___L-§_I_O__Ne£s (1986:8) report indicating that of the union members and union family members who voted in the 1986 elections, 76 percent cast their ballot for an endorsed candidate. Results of their logit analysis, indicate probability of voting for an endorsed candidate. Union membership is found to be a significant predictor of voting for the COPE-endorsed House and Gubernatorial candidates but not union-endorsed Senate candidates. How effective is COPE in accomplishing its objectives? If, as has been suggested above, a primary role of COPE is to disseminate information to union members, the work of Juravich and Shergold (1986:143) may be helpful in determining COPE's effectiveness. Results of their analysis of responses of AFL- CIO affiliated union members regarding union influence indicated that for all issues except one (taxes), a higher proportion of union members who voted for Mondale than for Reagan were aware of the official AFL-CIO policy on important issues. Furthermore, nearly 70 percent of Mondale supporters reported that the effect of the AFL-CIO presidential endorsement was either considered, or of importance; whereas, less than 30 percent of Reagan supporters felt that the endorsement was important, or worth consideration. COPE is unquestionably a powerful political vehicle for 20 the AFL-CIO. Saltzman (1987:71) ascertained there to be a significant relationship between COPE scores of incumbents and labor political action committee (PAC) allocation. Congressional candidates were allocated an average of $25,963 more PAC monies during 1979-1980 if COPE had rated them.at 100 percent than if they had been rated at 0 percent on their stands on issues and endorsement of the labor agenda. Masters and Delaney (1987:225) presented evidence of large PAC contributions made by over 70 national unions and the AFL-CIO, from 1978-1982. In 1982 alone, $20.9 million was "invested" in Congressional campaigns. Monies are allocated to those candidates whom labor leaders believe would best represent labor in terms of issues and of legislation relevant to these issues. The angrggginnnl_gnnrn§r y highlights the record of candidates and incumbents in terms of their efforts on behalf of interest groups. The publication also documents ratings of these candidates by COPE, as well as labor leader stand on issues. Several factors are antecedent to the distribution of labor PAC funds to candidates. Masters and Zardkoohi (1986) investigated disaggregated data in order to find determinants of financial resource allocation to federal candidates between 1974 and 1982. Three variables were found to have a significant, positive effect on labor PAC allocations: legislator liberalness (as to issues), electoral vulnerability (electoral challenge candidate confronted in the 1978 general election), and Democratic party affiliation. 21 Labor groups have "mixed strategies," of support for Congressional candidates according to Herndon (1982) and exhibit more strategic diversity than do business groups. Among these strategies are reliance on candidates' records, supporting those candidates who make themselves accessible, O C 0 ‘31 us 0 s- -; gscr I IIO‘! - no 0 :ssz-s s .‘00 . - ,- __1 1-11.- . . - _ _ - , -A- Q - iggngg, nnd defending thosg incnnnenrn wno nre disposgg n9 rngir r'nterenns. The issue of party identification again surfaces in the Herndon analysis; data shows that where a candidate's record of support for labor's interests either does not exist or is ambiguous, party becomes a surrogate (where Democratic party identification is positively correlated with labor PAC contributions). Wilhite and Theilmann (1986:175) suggested that there is a "pattern" to the contributions of both labor and corporate PACs which may be explained as a function of supply and demand. Demand variables consist of incumbent tenure, tightness of race, and candidate ideology, thereby necessitating publication of candidates' stands on issues. Campaign funds tend to be supplied to anticipated winners, powerful candidates, and ideological allies. A.major component of labor PAC research is analysis of PAC impact on voting and elections. It is the implicit strategy of organized labor, argued Saltzman (1987), to "maximize the number of Democrats in Congress" (175). It seems, though, that to some extent, corporate PAC contributions to Republican candidates counteract the prevailing support for Democratic 22 candidates. In.order to:measure the net effect of labor PACs, Saltzman examined the relationship between PAC contributions and WW d or. The relationship, he argued (163) has been weakened by business lobbies, declining’ membership, and growing numbers of "public interest" groups (e.g. environmental and consumer protection groups) whose issues may take precedence over those of organized labor. He found the impact of the contributions to be substantial, but cautioned labor against the growing power of corporate PACs. Kerr and Rubin (1981) examined issues voting when establishing the relationship between union political action and passage of economic legislation. Union membership significantly influenced voting by Congressional incumbents on most issues examined. The link was explained to be a function, in part, of the voting power of union members, as well as of their campaign contributions. Masters (1985) found that during a period of low resource commitment to political action (1977-1981), the three largest non-postal unions suffered major defeats in Congress regarding issues of importance. Results from Chappell's (1982) simultaneous probit-tobit model analysis of Congressional voting revealed no significant relationship between contributions to Congressional campaigns and voting decisions. Two variables were found to have a significant impact on voting: incumbent ideology and constituent preferences. 23 The research at hand queries much in the same way as does Saltzman, whose study concentrated on what COPE considers to be interests vital to labor and ratings of roll call EOLQS- However, where Saltzman used the same sources as the present study to analyze the impact of PAC contributions on House member voting, the major objective of the research at hand is to utilize these sources to analyze impact on the voting behavior of union members. Saltzman found the relationship between PAC contributions and roll-call voting on issues to be significant. However, were these the issues important to the individual voter? The study at hand addresses this question. Most labor PAC contributions are allocated to Democrats, rather than Republicans (Saltzman 1987). At times, this allocation to the Democratic versus the Republican party exceeds 95 percent. Pohlmann and Crisci (1983) revealed that organized labor contributed more than $5 million to Democratic candidates for Congress in the 1976 elections (c.f. Donnelly, 1978), and that gf 292 elected Democratg, 251 nag negn gngorgeg by 9925 (c.f. Eccles 1976). APL-CIO AND COPE ISSUE OPINION PRONOUNCEMENTS According to Shribman (1991) political strategists are setting about organization of voters for the 1992 elections on the basis of support for issues, and not on the basis of candidates themselves. Ra (1978:122) argued that accelerated issue awareness among voters has become a distinguishing characteristic of contemporary American politics. Candidates' 24 stands on issues was an important reason for voting for or against the Democratic incumbent in 1984, according to Schlicting (1989). Repass (1971) found issue opinions to be an important variable in voting behavior, if only for a few specific issues. The correlation between strength of issue partisanship and presidential vote (.57 tau b) indicated a relatedness of vote to partisanship on certain issues. Borrowing from Converse (1964), Repass calls segments of the voting public who mention a specific issue in the open-ended questioning of the 1960 and 1964 NES "issue publics" (391). Issue publics ranged in size during 1964, from 21.3 percent (racial problems perceived to be an important domestic issue), to 1.0 (union, labor relations perceived to be an important issue). Twenty-three percent of respondents were able to bring forth four or more problematic issues, and the average number of issues brought forth was 2.5 per respondent. Campbell, et al. (1960:170) suggested that the role which any one issue plays in determining partisanship is dependent upon at least three criteria: 1. The issue must be cognized in some form (issue familiarity). 2. The issue must arouse some minimal intensity of feeling (issue intensity). 3. The issue must be accompanied by the perception that one party represents the person's stand on the issue better than do other parties. It was suggested, at the same time, that we not rely solely on the promulgations of issue opinion that the broad-brush.media 25 set forth, These issues may fail to capture the true sense of issue familiarity. The implication for special interest groups, including the AFL-CIO, then is to focus on familiarity and education of the favored stand on issues. From the mass media.the public may, the authors argued, "collect opinions as avidly as small boys collect butterflies" (173). Yet issue familiarity may be lacking. Familiarity must be viewed from a twofold frame of reference (177): nnnrgng§§_nr_i§§n§_nng WWI—$314M Again. implications for the AFL-CIO and other interest group strategic plans are obvious. In her study of the Communication Workers of America (CWA), Thomas (1986) concludes that unions must introduce and explain the issue. Since issue awareness by the public is a goal of lobbying organizations like the AFL-CIO, it is important to know to what degree the organization is successful in creating issue awareness. In order to accomplish this objective, one must first identify the issues and stands on which organized labor would like to establish issue awarenessn To do this, the 1987 and 1988 AEL-QIQ Rengrrg on Congress (Report), the 1985-1988 WW Interest Group Ratings (CQ) . and the Erggnggingn 9f rhg 12§7 nEL-QIO Convention (Proceedings) were examined. What follows is a synthesis of AFL-CIO/COPE issue pronouncements as set forth in these publications. 26 ISSUES IDENTIFIED (1) Egongnig Pglicy Ingues. Economic issues identified by the AFL-CIO and COPE as being important include the budget deficit, farm policy, import quotas, minimum wage, pensions, social security, taxes and unemployment. Budget Deficit. 19§6 CQ The AFL-CIO and AFSCME joined senior citizens groups in lobbying against the proposed balanced budget amendment which would result in social security, medicare and other social program reductions; 1251 QQ The AFL-CIO opposed a balanced budget amendment to the Constitution (Senate vote 45); ggnngnnign__£rgg§nging§ Resolution No. 85 It is resolved that the AFL-CIO in its efforts to help change economic policy, opposes balanced budget requirements, either through the Gram-Rudman-Hollings statutes or the Constitution. Farm Policy/Assistance. angnnnign___£rgg§nging§ Resolution 21 Resolving to strengthen the rural economy and improve farm family quality of life, the AFL-CIO states unequivocally that purchasing of U.S. farmland must be regulated, commodity speculation be "curbed," and the growth of corporate agri-business be limited (178). Also supported were Federal Reserve Board allocations of credit to family farmers at lower interest rates and training and employment assistance programs for displaced agricultural workers. Farmer-owned cooperatives and finnily farms are the objects of support. Cooperatives are encouraged "as an effective balance against corporate domination and control of agricultural 27 products" (179). It is resolved that a farm policy would include price support system for small-medium-sized farms, governmental purchasing of surplus food and distribution to the poor and foreclosure protection and emergency credit programs for family farmers. Import Quotas. 98 Re t o o s The AFL-CIO supported a trade bill to improve a 1% limit on the import of 185 types of textiles and apparel in order to ameliorate such adverse effects as the loss of 300,000+ jobs and numerous plant shutdowns; 1986 C9 The AFL-CIO supported textile and apparel import quotas (vote 305); 195g QQ The AFL-CIO supported limiting the import of textiles and shoes (House vote 319); anventi on Proceedingn Resolution No. 12 The AFL- CIO supports legislation which would "Provide relief from export-oriented industrial targeting practiced by foreign governments that seek to expand their sales and employment at the expense of the United States" (91). The importance of import quotas is illustrated in the following portion of the resolution: The national policy of import restraints on steel must be fully implemented, and the Steel Import Stabilization Act vigorously enforced, particularly with respect to modernization and the training of displaced workers. Should the import restraint program prove ineffective, quota legislation will be necessary... (Ibid). Minimum Wage. 988 e re The AFL-CIO supported a cloture motion to stop the filibuster of a piece of legislation which would increase minimum wage to $4.55 by 28 January 1, 1991; 1112.29 Secretary-treasurer of the AFL-CIO, Thomas R. Donahue, advocates an increase in minimum wage and refutes the assertion by Sen. Hatch that 11 million new jobs since 1982 indicates economic growth. The existing $3.35/hour paid to 6.7 million workers is below a decent living wage, said organized labor; M organized labor representative, Robert McGlotten, declared that a minimum wage bill which included a training wage would not be supported by the AFL- CIO; Convention Prgceedings Resolution No. 24 supports Kennedy-Hawkins Minimum Wage Restoration Act of 1987, which would adjust the minimum wage upwards by 50 cents/hour the first year and 40 cents/hour in each of next two years. Pensions. 1986 Q9 The AFL-CIO opposed a House bill limiting the amount of benefits which may be allotted to employees on a tax-free basis. These limits would decrease the average annual pension by lowering the floor on pensions from $75,000 to $65,000 and the $90,000 ceiling to $77,000; 1217—952 The AFL-CIO supports provisions in tax law which would permit designated employees to receive self-made contributions to their own pension plans on a tax-free basis for three years after retirement (Senate vote 135); 1988 99 A representative for the AFL-CIO referred to pensions as "deferred wages," to which employees are entitled. Organized labor opposed the siphoning off of excess assets by companies; Wan—PW Resolution No. 28 "The AFL-CIO supports making employers fully liable for all promised pension benefits, both under a 'standard' and a 'distress' 29 termination. This would help to provide increased protection for participants... The AFL-CIO is greatly concerned that nothing has been done on termination reversions in which employers terminate well-funded defined benefit plans to secure so-called excess assets, those assets in excess of present plan liabilities" (444). Social Security. ;_9_§§__C_Q The AFL-CIO supports renewal of Social Security COLA (vote 73); Co t' Resolution No. 29 The AFL-CIO stands behind the following: repeal of the increase to 67 (from 65), the normal age of retirement, marital earnings sharing (to resolve women's social security problems), improved.benefits formulas for the disadvantaged in wages, prohibition of Social Security Trust Fund asset divestiture by the Treasury Department, the establishment of a Social Security Administration, and a strengthening of social security protections. Taxes. 128g 99 The AFL-CIO lobbied for contraction of complete deductibility of fringe benefits. It also supported a progressive tax which would more heavily tax upper income earners (votes 26, 27); WW Resolution No. 61 Several avenues of expression concerning tax policy are expressed: That the AFL-CIO deplores the whole approach to economic development based on signaling to foot-loose corporations a state or locality's receptive "business climate." This approach may succeed at stealing jobs from somewhere else, but it does nothing to create new jobs; and, be it further RESOLVED: That the AFL-CIO will vigorously oppose the use of all 30 types of "beggar-thy-neighbor" corporate tax incentives as well as individual tax cuts skewed toward the wealthy to attract business to locate in a state or locality, including property tax abatements, sales tax exemptions, corporate income tax loop- holes or rate reductions, and the establishment of enterprise zones (152). Resolution No. 62 Herein are stated resolutions which address specific measures by which to formulate an equitable state tax system. These include the following: progressive personal and corporate income taxes in order to ensure greater revenue allocation to public services, the removal from tax rolls of families whose income falls below the poverty line, elimination of loopholes for the wealthy, broad-based graduated income taxes, taxing of intangible poverty, and property tax relief for persons with low and middle income levels" Resolution 85 calls for higher taxes for the rich and for corporations. Unemployment. 1986 C9 The AFL-CIO supported benefits for long-term unemployed persons; gonynntion grgggegingn Resolution No. 27 Following its statement that "Unemployment insurance in the United States is failing to provide benefits to nearly seven of every ten of the jobless," the AFL-CIO Executive Council resolved to support funding for programs targeting displaced or dislocated workers, along with the following: - ...improving the funding of the un- employment insurance program by indexing the federal taxable wage base to at least 65 percent of the average annual wage, establishment of federal 31 minimum benefit standards to ensure fair treatment for all jobless workers... - ...[reformation of] the criteria for eligibility and ease the qualification requirements for extended benefits. (2) W. Foreign policy issues identified by the AFL-CIO and COPE as being important include Central America, defense, South Africa, and trade. Central America. anygnnign_2rggnnging§ Resolution No. 142 The AFL-CIO resolved the following: "The AFL-CIO opposes any further funding' of the Contras; and be it further resolved; the AFL-CIO calls for a.moratorium on all military- related.and to the government of El Salvador" (303). Peaceful settlements "through dialogue and negotiations between the governments of Nicaragua and of El Salvador..." (287) is urged. Defense. Convention Proceedingg Resolution No. 110 The AFL-CIO resolves that it will continue its tradition of support for a strong national defense, as well as its support for NATO and balance of forces between NATO and the Warsaw Pact. It is resolved that the AFL-CIO supports "American research, development and limited testing of strategic defenses" and "protection of our nation's nuclear retaliatory capability" (281), while believing' that arms limitation agreements are also essential. South Africa. 12§6 C9 The AFL-CIO supported economic sanctions against South Africa (vote 173); 1981 CQ The AFL- CIO supported both a ban on all U.S. trade and business 32 investment (Senate vote 244) and prohibition of bank loan renewals to South .African. businesses (Senate vote 234); Conynnrign Ergcegdings Resolution No. 37 "The AFL-CIO continues to push for the strongest possible sanctions to be imposed by the United States government...against the apartheid regime" (291). Trade. 19g: nggrt on Congre§§ In response to unpre- cedented U.S. deficits, the AFL-CIO supported trade reform legislation aimed at reducing the deficit, combatting unfair trade practices and restoring a balance. It also supported the Gephardt amendment, designed to expose those trading partners which have excessive trade surpluses; 12§§_3§pgrr_gn angrnnn The AFL-CIO supports the Democratic Congressional- backed expanded trade adjustment assistance program, a "mechanism to reduce the trade deficit by strengthening U.S. trade law, the inclusion of abuse of worker rights abroad as an 'actionable' unfair trade practice, and the plant closing provision" (3); 1986 C9 The AFL-CIO joined the group LICIT (Labor-Industry Coalition for International Trade) lobbying in favor of an omnibus trade package targeting stringent U.S. response to unfair trade practices. Lane Kirkland stated on Capitol Hill that government seeks "cheap and docile labor," in encouraging foreign investment; 987 The AFL-CIO supports retaliation against countries guilty of unfair trading practices (House vote 120). Labor also approves of 8538, requiring 180 days notice prior to plant shutdown. Organized labor backs an amendment to the 1987 trade bill, 33 which would reduce the U.S. bilateral trade imbalances with Japan, South Korea, and Taiwan; 1988 CC For the third year in a row, the AFL-CIO supported a measure to require countries whose trade surpluses are excessive, to eliminate unfair trade practices(House vote 72). In House vote 426, the AFL-CIO opposed dropping the Gephardt amendment to the trade bill; Conygntign Proceedings Resolution No. 12 Those provisions featured in certain House or Senate trade reform bills which are ardently supported, include the following: - Require any major trading partner that maintains excessive surpluses with the United States and trades unfairly to reduce those surpluses. - Provide relief from export-oriented industrial targeting practiced by foreign governments that seek to expand their sales and employment at the expense of the United States. The resolution.also supports therTextile and Apparel Trade Act of 1987. (3) Socinl Policy Issues. Social policy issues identified by the AFL-CIO and COPE as being important include childcare, civil rights, drugs, education, environment, food stamps, healthcare, homelessness, housing, immigration, pay equity, research on AIDS, and welfare. Childcare. Convnntion Procgedingg Resolution No. 33 Citing the fact that 52% of mothers with children under the age of six are members of the workforce, inadequate funding for childcare and cuts in childcare funds, the AFL-CIO urges 34 Congress to take action in the following areas: Passage of the Act for Better Child Care Services, a comprehensive childcare measure; funding Title XX of the Social Security Act; assurance that licensed child care be available for welfare mothers involved in self-improvement programs; and passage of the Parental and Disability Leave Act. Civil Rights/Discrimination. 1288 Report on Congrngg In reaction to the 1984 Supreme Court decision in Grove City, which narrowed Title IX protection, the AFL-CIO upholds the Civil Rights Restoration Act. The Act is designed to restore four civil rights laws to their original intent of barring federal funding to educational institutions which discriminate; Convgnti on Proceedings Resolution No. 35 It is resolved that the AFL-CIO will focus on restoring and maintaining anti-discrimination laws and affirmative action. The call went out for voluntary affirmative action plans, and eradication of gender, age and racial discrimination. It is resolved that Congress be urged to pass the "Civil Rights Restoration Act." Drugs. anvgntion grongngingg Resolution No. 208 The AFL-CIO urges Congress to develop additional legislation aimed at increasing’ efforts in. the areas of law' enforcement, education, drug treatment, drug shipment and trafficking interdiction, and drug crop eradication. Education. 1926 C9 The AFL-CIO supported funding for Head Start, education, and training programs (vote 58); 1221 CQ The AFL-CIO supported maintenance of education programs at 35 levels equal to 1986 (Senate vote 77); anynnnign_2rgggnging§ Resolution No. 34 Federal investment in public education in the following areas is supported: elementary, secondary, higher (citing "shortsighted" budget proposals), vocational, adult, worker, and literacy education. Environment. W The Environment The AFL-CIO urges that the following' measures be followed: expediency in cleanup of hazardous waste sites by the EPA, subsidization of coal-burning facilities in order to deal with acid rain, more powerful pesticide-control law, increased funding for waste water treatment projects, stricter enforcement of rules covering hazardous chemicals, and federal legislation concerning vapor recovery. Food Stamps. Convennign grogegdingg Resolution No. 32 The AFL-CIO urges Congress to increase food stamp benefit levels, broaden eligibility for food and nutritional programs, provide outreach programs for the needy and to pass the Hunger Relief Act of 1987 and.the Urgent Relief for the Homeless Act, which would provide shelter, food and healthcare. Healthcare/Medicare. W The AFL-CIO supports legislation to expand Medicare into the area of catastrophic and long-term illnesses. Legislation mandates coverage of unlimited free hospitalization, payment of high physician bills and drug benefits; 1226 C9 The AFL-CIO joined Republican Senator Grassley in opposing government intrusion into employer health plans. The decision is based upon the federation's belief that decisions should be made during the 36 collective bargaining process, and that limits on flexibility would jeopardize employees' chances of attaining adequate health care benefits and insurance. The AFL-CIO supported Medicare/Medicaid funding of $4.6 billion for 1986-1988 (vote 74); m The AFL-CIO opposed reduction of catastrophic- health care expenditures; Convention Proneegingg Resolution No. 30 The AFL-CIO Executive Council supports the notion that "all Americans have access to quality care they can afford through the enactment of a national health care program" (562) . Furthermore, enactment of the Minimum Health Benefits for All Workers Act of 1987 (requisite health care for all employees), is urged. Homelessness. Convention Proceedingn Resolution No. 55 The AFL-CIO calls for governmental assistance for the homeless by way of supplies and shelter; Resolution No. 64 Further, it is resolved that the AFL-CIO will join with advocacy groups in urging Congress and the government to provide increased funding for social services, housing and to the homeless, specifically. Housing. 1987 Repog on Congrgns Based upon the fact that federally subsidized housing is an essential anti-poverty program, the AFL-CIO opposes an amendment which would decrease federal housing assistance by $1.7 billion and redirect money targeted for construction of new units into existing unit repair; 1222_CQ The AFL-CIO opposed a proposed $1.7 billion cut in monies set for construction of new federal housing units (House vote 169); MW Resolution No. 37 15 Citing lower real income and increased interest rates, the Executive Council calls for revitalization of programs for new housing as well as renovation of existing low-income housing. Immigration. 1926 CC When attempts were made to postpone enforcement sanctions on employers who knowingly hire undocumented workers (thus creating a "pool of low-wage, exploitable workers who undercut job opportunities, wage levels and working conditions for American Workers" (12) ) , the AFL-CIO opposed delaying sanctions and upheld the original immigration reform bill; 1928 CC The AFL-CIO opposed postponement of enforcing government sanctions on employers of illegal aliens (Senate vote 124); W Resolution No. 219 takes a two-pronged approach. First, the AFL-CIO supports sanctions against employers who knowingly hire unauthorized workers. Second, it urges that well- established, contributing (illegal alien) members of communities be allowed the opportunity to enter legalization programs. Pay Equity. anygntion groggegingg Resolution No. 78 The AFL-CIO resolves to support pay equity through elimination of racial and gender-based pay inequities in negotiating labor contracts, pay equity legislation, pursuit of legal remedies and enforcement of Title VII. Research Funding/AIDS Research. anygnnign_£rggngging§ Resolution No. 152 Finding AIDS to be a major health problem, the AFL-CIO urges Congress to increase federal funding for research, services, and education. Financial support is also 38 deemed necessary for those who have contracted AIDS or ARC. Welfare. 1227 Report on Congre§s In an effort to promote family welfare via education and.training programs, childcare and medical subsidies, and improved benefits, the AFL-CIO supported H.R. 1720 (The Family Welfare Reform Act of 1987); 1222 Egpprn Qn Congrens When a Dole amendment threatened to weaken the above bill, the AFL-CIO opposed it; 2222_CQ The AFL-CIO supported a limit on welfare reform bill amendments; t P e ' Resolution. No. 31. The .AFL-CIO supports legislation which would enact a family support program. It also encourages support for AFDC family assistance, including assistance in education, training, childcare, Medicaid, and collection of child support. (4) Labor Issues. Issues identified by the AFL-CIO as being of particular importance to labor include high risk notification, plant closing, safety, striker replacement and training. High Risk Notification. 1986 CC Organized labor supports a measure to monitor occupational job risks via a government health board (HR162, S79). Where workers are at risk, employers must notify them; 1928 CC The AFL-CIO opposed a two-year study of high-risk notification.which.would.postpone passage of the bill (vote 354). It lobbied for Senate bill 79, which would require notification in cases of exposure to toxic substances; Cpnynnnrgn_2rggngging§ Resolution No. 86 The AFL-CIO urges Congress to increase funding for OSHA, to strengthen penalties for violation of health and safety 39 regulations, and to legislate compensation of victims of occupational disease; Resolution No. 26 Includes the provision that Congress should enact the High Risk Occupational Disease Notification Act. Plant Closings. 1 88 Re rt 0 o e The AFL-CIO supported the inclusion in the trade bill of an amendment to mandate 60-day advance notice to employees, by employers, of permanent plant closings. It opposed a motion to eliminate the amendment. It then supported a separate plant closing bill; 1286 C9 The AFL-CIO supported mandatory notification (vote 383) of shutdown, along with requiring owner consulta- tion with employees prior to shutdown (vote 372) ; 1988 CC For the third year in a row, the AFL-CIO supported a 60-day advance notice of plant shutdowns (Senate vote 180); Conygnnipn Erocegdingg Resolution No. 180 The AFL-CIO "Commands the Secretary of Labor's Task Force for its findings on the need for early notification of plant closings and a massive effort to assist displaced workers" (134). Safety OJT. 9 e ort 0 Co r Estimating that approximately 100,000 workers die from occupational diseases and 240,000 more are disabled, the AFL-CIO supported legislation which focused on detection and prevention of occupational disease. Notification of high risk (e.g. exposure to known carcinogens) would be included. The AFL-CIO opposed an amendment which would drop risk notification and substitute a two-year study and enforcement of existing safety laws . 40 Striker Replacement. Cpnvention Eroceegingg Resolution No. 57 "That the AFL-CIO calls upon the Congress to move without delay to protect the jobs of striking workers by enacting legislation that would prohibit the hiring of 'permanent' replacements during economic strikes" (204). Training. Conventign Proceedings Resolution No. 13 Citing job elimination, job changes, new jobs and occupations, the AFL-CIO calls for action in the following areas: full employment through sound fiscal, training, monetary, labor market, and trade policies; protection of workers and communities from adverse effects of plant closings, via plant closing legislation which would include worker adjustment assistance programs; training opportunities for all workers, in.order to keep current jobs or advance to better ones; basic skills training, retraining, and opportunities for mobility; union-based training; employer-based training; public and private sector cooperation in training program design and implementation; on-the-job training; and Trade Adjustment Assistance, Job Corps, and other programs of training assistance. Issue opinions are found by Rappaport, et al. (1990) to vary according to gender and political party identification, when categorized by "foreign policy issues," "social welfare issues," and "women's issues." Foreign policy issues show a greater gap in opinion than do the other areas. Issue opinions are also said to differ according to their perceived level of validity. Flanigan and Zingale (1983) contend that 41 issues surrounding the treatment of African Americans have historically been the most volatile in the United States. Issues surrounding foreign affairs vary greatly in their level of volatility, where periods characterized by wars and hostage crises intensify feedings in the short run. As to economic issues, Flanigan and Zingale find no change "in the American public's persistent willingness to support federal governmental programs intended to solve social problems. In economic matters Americans are more liberal than conservative" (94). Data used in the Flanigan and Zingale study were derived from the 1980 Survey Research Center/Center for Political Studies National Election Studies. For union members, "economic issues [still] occupy the most important position," with social issues "narrowing the gap," according to Ra (1978:134). Political W of union members of the stands of local labor leaders on issues is an important component in measuring the impact of the attempts by leaders to influence their membership. Blume (1970) took this perspective in his study of United Auto Workers, Local 12 in Toledo, Ohio. Members were asked whether or not they were aware of the union's stand on issues. Members were better able to identify stands on issues (44 percent identified the union stand on "open housing") than were able to identify which candidates had been endorsed (15 percent). Only 22 percent of respondents reported that they would adhere to union stand on issues and candidate endorsements "always" or "most of the time," and 23 percent, 42 "occasionally." Further results of Blume's study support conclusions previously drawn by TTnman (1951) that members tend to believe that the appropriate role of union leadership is tending to the furtherance of "bread and butter" issues. When Juravick (1986:147) looked beyond the "myths" of the relationship between union members and voting in the 1984 presidential election, he found mixed results as to union members' views on the issue positions of candidates. A major element of issue involvement which may lead to issue-oriented political behavior is what Campbell, et al. call "issue intensity". This refers to the intensity of supporting or opposing sentiments surrounding an issue, rather than the intensity of its importance, on an aggregate basis, to national welfare. There appears to be a relationship between familiarity and intensity, they argue in that "the most familiar issues are also the issues that evoke most strong feeling. . . nevertheless it is clear that familiarity is not synonymouswwith intensity" (176). Where awareness of issues and of their political salience are important in the context of issue familiarity, issue intensity relies less on such qualitative variations than on the quantitative (170). ‘Variation.in intensity is measured in the W on a Likert Scale for those issues questions which are close- ended. Campbell, et a1., set forth.the:notion.that.competition.between.issue importance is a key element of variations of intensity. Where one issue 43 is given preference, another is denied. The national Elngnionn Snryey includes a.block.of questions concerning the importance of governmental funding (some of which are included in the research at hand) touching upon such key issues as funding for welfare, national defense, and affirmative action. Competition for governmental resource allocations is a reality in American society. Respondents are asked to quantify (via a Likert Scale) the importance of any given issue. BIOGRAPHICAL CHARACTERISTICS Strong (1977:2) argues that "once you know various demographic details about a person, your prediction of how he will vote will be accurate most of the time." Weisberg (1987), in his discussion of a voting gap, operates from the assumption that differences in voting behavior result not only from group mobilization and divergent issue appeals, but also from demographic factors (335). Parent, et a1. (1987), believe the core of Democratic coalitions to be comprised of minorities, union members, the poor, Catholics, urban dwellers (from the largest 12 U.S. metro areas), and Southerners (c.f. Axelrod, 1972; Abramson, Aldrich, and Rohde, 1983; Petrocik, 1981). Because of the importance placed on demographic factors, independent variables utilized in the research at hand include gender, age, race, religion, education, region, income, party identification, and incumbency. In order to conduct an appropriate analysis of the impact of these variables, the criterion used in each separate 44 analysis has been chosen for its ability to measure correlation with either voting or issues opinion pronouncements of the AFL-CIO and COPE. W Union.members tend to identify with the Democratic party, and, according to Delaney, et al. (1990:625), are predisposed toward voting for Democratic candidates. Delaney, et al., in fact, suggest that a union voting block cuts across Senate, House and gubernatorial elections. Coleman (1988:689) is unequivocal when speaking of the relationship: The electoral basis of labor's position in American politics is anchored in the Democratic Party. Halloway (1979:119) merely states that "labor tends to favor Democrats," while Aronowitz (1973:251) speaks of labor's "dependence" on the Democratic party and Sullivan (1984) cites the fact that America's unions focus their political spending priorities on Democratic Party support. Numerous analyses have been conducted utilizing union status as a potential correlate of support for the Democratic party. The overwhelming majority of these studies indicate there to be a positive and typically significant relationship between status as a union member and identification with the Democratic party. Analyses of voting behnvior tend also to reveal a positive relationship between status as a union member and voting for a Democratic rather than a Republican candidate, even though the relationship is not as strong as it 45 is for party identification. Berelson, et al. (1966:53), assert that the relationship between union membership and voting behavior is dependent upon the degree of commitment to unionism. Petrocik (1981), cited in Parent, et al. (1987), lays out the Democratic voting coalition: Catholic and Jewish voters, urban voters and labor union voters. Likewise, Norpath (1987) speaks of a Democratic coalition which includes union households. Results of both party and electoral coalition analyses by Wattenberg (1987) , reveal there to be strong union membership, in both the 1956 and 1984 presidential elections, within the Democratic party coalitions. Sullivan, et al. (1977), reminds us that there was actually a Labor Coalition Clearinghouse, comprised of nine labor unions (including the United Auto Workers, American Federation of State, County and Municipal Employees, and the Communication Workers of America) which rigorously supported Democratic candidates and union stand on issues. Parent, et a1. (1987), argue that union members and their families are among those sociodemographic groups which comprise the Democratic coalition. Results of their aggregate-level study of the 1984 Democratic primaries and caucuses along contextual and structural dimensions show that blacks and union. members are dominant among sociodemographic influences on voting behavior. Mondale's position, in fact, was positively affected by the percent of unionized voters in each state. Wilson's (1979:93) study of politics in 1970 indicated 46 there to be three areas where the tendency of the union vote is toward the Democratic party: voting for Senators (67.8 percent union families versus 58.1 percent of general population voted Democratic), House of Representatives (65.9 percent union families versus 54 percent of general population voted Democratic), and party identification (53.8 percent union families versus 43.5 percent general population are Democrat or strong Democrat). Dye (1973:113) found that in all three national elections of the 1960's, union members voted Democratic at a higher rate than did non-members. For national elections, in fact, many researchers have revealed a distinct "union" vote, according to Masters and Delaney (1987:343). In their review of empirical literature the authors cite Axelrod (1972, 1982); Form (1982, 1983); Campbell, et al. (1980); Ra (1978); Wolfe (1969); Kornhauser, Sheppard and Mayer (1956); and Sheppard and Masters (1959) as researchers whose findings support the notion of a "union vote." There has been some concern about shifts in political loyalties of union members. Wolfe's (1969) analysis of m in union member voting propensities resulted in his concluding that members, while exhibiting a distinctiveness in voting behavior, did n93; vote as a "bloc" in any of the six presidential elections between 1948 and 1968. Brown and Catlett (1983:46) deny the existence of a pro-Democratic "labor vote" in the 1980 elections, based upon the fact that a relatively higher proportion of union households voted the 47 Republican ticket. There is also evidence in the analysis of a decline in union preference for the Democratic party. The Cnnnn_flgrigzflgrn;g (Monday, September 2, 1991:9), in an article entitled "Labor Is Irked With Democrats," discusses the apparent disillusionment of rank and file members with union leaders as a result of a "series of defeats" in Congress. Where unions have continued to support their Democratic allies, the Democratic Party is perceived as not substantially reciprocating. Over the last two years, the article states, several key labor issues have been poorly handled by Congressional Democrats. Prominent Democrats defied the firm anti-free trade (with Mexico) stance of the AFL-CIO when they supported fast tracking negotiating authority pertaining to the matter. Furthermore, Congress also failed to override vetoes of measures which would have helped settle an Eastern Airlines strike. Congress also failed to get 1990 civil rights legislation passed. Health care and education improvement measures are slow in coming. And, the vote over the striker replacement bill fell short of the majority needed to override a veto. Borrowing evidence from the 1980 and 1984 presidential elections, Fields, et al. (1987) cited the fact that a significant proportion of union members (over 40 percent) voted for the Republican candidate. The researchers suggested that union members are politically heterogeneous, and borrowing from Lyons (1969), Seidman, et al. (1951 and 1952) and Masters and Delaney (1986) reveal that members fall along 48 divergent points of the political spectrum between liberal and conservative, often disagreeing with the more liberal positions of their leadership. It is suggested that the leadership might align themselves with opinions truly representative of union members. Fields, et al. (1987) concluded that, in order to enhance national election results, appropriate union agendas should be represented at the local level. Supporting this notion of a politically heterogeneous union membership, Masters and Delaney (1987) concluded that "union members and leaders often differ in political orientation, though the extent of disagreement varies across political issues" (336). Kochan and Wever (1991:380) in discussing conditions for change in the industrial relations arena have this to say: Even while the New Democratic Party has helped sustain labor's influence in Canada (Bruce, 1989), the declining role of U.S. labor movement in the Democratic Party (and, indeed, the weakening of the Democratic Party itself) continues to constrain American labor power and influence. While the trend may be disturbing to union leaders, it should not be taken as evidence of a significant decline in labor's political record. Masters and Delaney (1987) studied union legislative records during the period of 1980-1984, a period representative of a marked turn toward political conservatism. Results of data obtained from both the 97th Congress (1981-1982) and the 98th Congress (1983-1984) reveal 49 the following facts: - In the 97th Congress, all but two of the eight unions reviewed, lost a majority of votes pertaining to their legislative positions, even though Democrats controlled the House 243-192. - In sharp contrast, the 98th Congress resulted in each union winning a significantly greater proportion of roll calls, and a higher average percent "right" votes (per their legislative positions). Masters and Delaney concluded from their analysis that "unions have weathered the Reagan presidency and a Republican Senate without suffering a substantive diminution in labor relations laws and general labor legislation" (15). They remind us, however, that unions have suffered political defeats in the last decade, that Congress has a propensity to concentrate on non-labor issues and that there are interunion variations in successes in obtaining "right" votes from Congress. Delaney, et al. (1990:625) revealed. that COPE primarily' endorses Democratic candidates (c.f. Greenstone, 1977), and that union members are predisposed to vote Democratic. Wattenberg's (1987) inquiry into group realignment included a comparative analysis of 1956 and 1984 Presidential Elections. Union member electoral and party coalitions analysis disclosed that while the direction of union member alignment is toward the Democratic party, a healthy percentage of union vote in both years (48 percent and 43 percent respectively in electoral coalitions) went toward the Republican candidate. Party coalition information, however, reveals that 50 the cleavage between union and non- union voters on party identification remains intact; the Democratic party has suffered losses from the decline in union membership -- not from any movement by union members toward Republican identification (63). Norpath (1987:379) likewise focused on group composition of party electoral coalition. Norpath explored realignment of core groups, to the Democratic or’ Republican. party, and included.union households as one of singroup identifiers. IHe found that in 1986, 66 percent of union members and poor people identified. as IDemocrats, more than. Catholics (59 percent) or southerners (56 percent), but less than city dwellers (71 percent) or African Americans (90 percent). Findings of the study supported those of Wattenberg, in that not a single one of the six groups commonly counted as the components of the Democratic coalition has moved into the Republican fold (379). The union household coalition has not realigned with the Republican party. When Ladd (1985) examined the issue of group realignments in the 1984 presidential election, he fell short of declaring that a realignment in union member voting had occurred. Of those groups which he designated as "mainstays of the New Deal Coalition," white southerners and blue collar votes were said to have dealigned with the Democratic party, but the union vote was seen to give Mondale a margin (albeit a modest one). Labor leaders, in contrast, had, in the author's words, gone "all out" for Mondale (13). 51 i: 3.19:! .0 -_ ,gonme er O__l't a. “1:. '0 sea --3it’,=!,‘ EIQDQHDQEEQQLfi Many writers have questioned the impact of labor endorsements. Patton and Marrone (1984) assessed the phenomena by asking selected union members a series of open- ended questions referring to AFL-CIO endorsements for the 1980 presidential elections. Results clearly indicate that union members were aware of endorsements at a higher rate than were nonmembers, but that only 12 percent of those sampled stated that they were influenced in their vote by the endorsement. Form (1973:238) likewise found little evidence of rank and file support for the political promulgations of their leadership: The main finding of this research is that nowhere do workers endorse official pronouncements that unions should seek political solutions to their problems. Rather, they feel that the union should address itself almost exclusively to wages and working conditions, and, on occasion, to social solidarity. Union officers, however, are more radical than the rank and file, and they want to involve them in politics. Blume's (1970) analysis of one United Auto Workers local supported this contention somewhat. Detroit Local 12 members indicated strong support for leadership efforts to increase wages and benefits, as well as to improve working conditions, each thought to be the "bread and butter" for the rank and file. However, results also revealed a propensity toward supporting leader pronouncements on issues closely related to traditional political activities. Funding for public 52 education, local candidate endorsements*workers compensation, and increased minimum wages were supported by members, as endorsed by leaders, while open housing and Medicare increases were not. Wolfe (1969) concluded, based upon his analysis of unions voting behavior, that "closeness" of feeling of union members to labor unions positively correlates with tendency to vote in the direction of Democratic orientation of labor leadership. Some researchers ascertain that there is a strong relationship between leadership pronouncements and rank and file members' political behaviors. Zuravich and Shergold (1988:374) reported that in‘their’survey'of.AFL-CIO.affiliated unions, union pronouncements had a strong effect on presidential choice in the 1984 elections. In their study of Detroit autoworkers, Sheppard and Masters (1959) reported there to be a majority of Detroit UAW workers who believed in the recommendations of labor groups and supported leadership's endorsement activities. Based upon their review of extant and traditional literature, Masters and.De1aney (1987:338) have concluded the following: - Most studies' findings support the notion that the rank and file supports union political involvement "in general." - There is frequent objection to npggifiig forms of political involvement, "snch gs instrugting the rnnk nnd file Qn hgw to votg on glecnign 22y," - The rank and file propensity is toward more support for traditional union functions (i.e. 53 collective bargaining) than for what is per- ceived to be nontraditional (political activity). - Rank and file issues opinion differs by issue topic. Leader political involvement in issues pertaining to economic or labor legislation receives more support than involvement with issues of a social nature. we: Fields, et al. (1987) suggest that female union members may be a major support group for political action strategies aimed at revitalization of labors political position. Their 1987 study revealed a propensity on the part of female union members to support national political involvement on the part of organized labor. When Wattenberg (1987) examined potential realignment of rank and file political identification, females were found to be less likely' to vote for' the Republican. presidential candidate than were male voters. Ladd (1989) discussion of the 1988 elections revealed that high status females identified.with the Democratic party (40%) by a slight margin over identification.with the Republican.party (38%). Twenty- one percent said that they were Independents or otherwise identified (16). In 1985, Ladd had noted a "gender gap" in the 1984 presidential vote, when in four separately conducted exit polls (NBC News, Los Angeles Times, ABC News/Washington Post and CBS News/New York Times), a higher percentage of female voters than male voters, had voted for the Democratic candidate, Walter Mondale. Yet Pomper (1976), in his analysis of three presidential 54 elections from 1952 to 1964, found that in two of the three elections (1952 and 1960), a higher percentage of women than men, voted for the Republican Candidates (Eisenhower and Nixon, respectively). And Gallup Opinion Index No. 183, December 1980 shows that in five of the eight elections between 1952 and 1980, an equal or higher proportion of women than men voted for the Republican candidate. Danielson and Murphy (1983:238) ascertained that the gap between men and women in voting participation which once existed, had closed by 1982. Citing Gallup Poll results, Danielson and Murphy (1983:238) pointed out that of the five U.S. national elections conducted between 1964 and 1980, in all but one (1976) election, a.higher percentage of women.than men voted for the Democratic candidate. Rappaport, et al. (1990:781) utilized the variable issues when analyzing gender differences in political attitudes. Sex differences were found to exist in all four foreign policy issues (defense spending, military aid to Central America, nuclear freeze and U.S. military presence in the Middle East), a finding supported in prior research. Delaney, et al. (1988) discovered a positive correlation between a given union's political involvement and the proportion of its membership that is female. They ascertained, during a study of interunion differences in political action, that the proportion of women members of a given union was positively correlated with that union's political involvement. The researchers suggest, therefore, 55 that one component of successful organizing drives targeting women might be an emphasis on political participation roles. Research into the role that women play in the presidential nomination process was investigated by Rappaport (1990). Female caucus participants in the presidential nomination process were found to be more liberal than were male participants. Furthermore, over the period of 1970-1990, women were found to play a larger role than previously, especially within the Democratic arena. Delaney, et al. (1990:626) suggest that women as a group may tend toward choosing a COPE-endorsed candidate since endorsed candidates normally support policies favored by women, such as funding for daycare and the Equal Right Amendment . Ase Danielson and Murphy (1983) believe age to be a stabilizing variable in.party preference, since allegiance to one party tends to strengthen as a person ages. Older voters are more likely than are younger voters, to vote Republican. Younger people tend to identify themselves as Independents. Campbell et al. (1960) noted the same results after analyzing Survey Research Data from 1952 to 1957. Younger people were more apt to be Independents, while those in their late twenties and thirties were beginning to solidify an identification as either a Democrat or a Republican. The Democratic Party is advantaged in its ability to attract 56 younger voters, while Republican identity appears to increase in older age categories. The results of their research rendered another interesting pattern: identification (by age group) was found to be a function "of the proportion of a person's life he has been associated with the group" (163). Such a notion would suggest that groups like the AFL-CIO may do well in disseminating information as to labors' agenda early in rank and file members' careers. Wattenberg's (1987) research into electoral coalitions yielded somewhat different results by vote. In the 1984 presidential election, three age categories (under 30; 30-59; over 60), all show a Republican vote of 57 to 59 percent of the electorate. Party identification in 1952 followed the conventional path, according to Ladd (1989), who reviewed Gallup Poll data from 1952 (April-July) and 1985 (January-July). There is a clear indication in 1952 figures that older respondents tended to identify with the Republican party, while younger workers were more apt to identify with the Democratic party. Figures for 1985 support the contention that with age, Republican party affiliation increases. In Pomper's (1968) analysis of three presidential elections between 1952 and 1964, it was found that in all instances, younger voters tend to vote for the Democratic candidate. The effect of age on political attitude and voting is explored by Niemi and Weisberg (1976:289) . Results support the notion that older voters (60 and above) have a strong 57 propensity to vote along party lines, as well as having a greater tendency than do younger voters to vote Republican. Delaney, et al. (1990) postulate that older voters may have a greater propensity than.do younger voters to choose the COPE-endorsed candidate since COPE ratings are usually high for candidates supporting social security system and housing subsidation policies favorable to the elderly. RACE According to Pae (1986), race is a significant predictor of voting behavior in all eight presidential elections which took place between 1952 and 1980. The dominant socio- demographic influences on voting choice during primaries according to research conducted by Parent, et al. (1987) , are, in fact, proportions of black and labor union memberships. Dye (1973:112) argued that a "disproportionate" amount of support for the Democratic party comes from (among other groups) African Americans. Although there have been fluctuations in the strength of ties between African Americans and the Democratic party (Niemi and Weisberg 1976), the majority of their vote has gone to Democratic candidates. It is typical, argue Flanigan and Zingale (1983) for racial minorities to vote for the Democratic candidate at a higher rate than do whites. Party identification in the Flanigan and Zingale study of 1952 and 1980 NES data shows a definite trend whereby African Americans tend to identify with the Democratic party. The 58 study also revealed a somewhat divergent vote by region, whereby a higher proportion of southern than northern African Americans identify with the Democratic party. Berelson, et al. (1948) noted the possibility of another divergence, this one relating to degree of member identity with their minority group. According to their research, the authors found a correlation between degree of affiliation.with.one's minority group and voting behavior. The more closely identified a minority member is to his or her minority group, the more likely that person was to vote for a Democratic candidate. Such was the case with 1956 and 1984 electoral coalitions identified by Wattenberg (1987). In the 1956 presidential election, 61 percent of the Republican vote was cast by whites and 36 percent by blacks. In the 1984 presidential election, 63 percent of the Republican vote was cast by whites, while only 9 percent was cast by blacks. Research of Ladd (1989) reveals a variability by state in 1988 (based upon NBC and Wall Street Journal election day surveys) in percent of vote according to race. Of the nine states studied, however, the greatest difference was only 10 percent; in Mississippi 87 percent and in North Carolina 97 percent of African American voters voted for Dukakis. Nationwide, 86 percent of the African American presidential vote went to Dukakis and 11 percent went to Bush. Seventy- eight percent of the African American Congressional vote went to Democrats, and 9.7 percent to Republicans. Perhaps Danielson and Murphy (1983:248-249) best summarize 59 the impact of race on political behavior when they state the following: Minority racial status, like religion, correlates strongly with party allegiance. From emancipation to the New Deal, blacks, at least those who voted, were strongly Republican; then Franklin D. Roosevelt won them away from the party of Lincoln. Although the Republicans have made a number of attempts to woo support and black leaders frequently urge their followers to vote more selectively, blacks have remained since 1932 solidly Democratic. It was in part Republican despair at cracking the black vote in northern metropolitan areas that made it attractive for Nixon, Ford and Reagan to make appeals to whites. Delaney, et al. (1990:626) assert ‘that COPE-endorsed candidates actually count on a substantial black vote. COPE endorsements normally go to those candidates supporting policy favorable to blacks, such as anti-discrimination in housing and civil rights legislation. 89.113129 According to Danielson and Murphy (1983), religion is a correlate of political loyalty. The relationship breaks down thusly: - white Protestants residing in the northern U.S. are more likely to vote Republican. - Catholics tend to vote Democratic, regardless of their economic status. - the Jewish vote, since the 1920's, has tended to be Democratic. (247) 60 Wattenberg's (1987) analysis of the 1984 presidential election found a similar pattern for Jewish and white Protestant voters, but does not strongly adhere to the pattern for Catholic voters. Seventy-two percent of the ‘white Protestant vote went to the Republican candidate, while only 31 percent of the Jewish vote did; however, fifty-four percent of the Catholic vote went for the Republican candidate--a larger proportion than is traditional. This anomaly may be accounted for by the fact that Ronald Reagan openly opposed abortion. Danielson and Murphy (1983) attribute Richard Nixon's success with Catholic voters to his pro-life stance. The stronger the ethnic ties of Jewish and Catholic voters, the stronger the allegiance tends to be to a traditional Democratic coalition (Petrocik (1981) cited in Parent, et al. (1987)). Danielson and Murphy (1983) have noted there to be a more pronounced tendency among Irish and Polish Catholics than among Italian Catholics to vote Democratic. A stronger vote may also occur due to the issues or candidate involved. For instance, in the 1952 elections, 70 percent of Jewish voters cast their ballot for Adlai Stevenson. In 1960, 80 percent of Catholic voters cast their ballot for John F. Kennedy (Pomper 1968). Berelson, et al. (1948) noted that the more Catholic one "feels" (or the more closely one associates with ones religion), the more likely that one's vote will be cast for the Democratic candidate. 61 mm Danielson and Murphy (1987) find that in general, the higher their education and income and the more professional the nature of their work, the more likely voters are to be Republican. Conversely, the lower their income and education the more closely their occupations are related to manual, unskilled labor, the more likely they are to be Democrats. (246) There are, however, prominent exceptions to this generality, according to the authors. Intellectuals tend to affiliate with the Democratic party. And a large base of support for the Republican party comes from middle-class white-collar workers. Education has been associated with the propensity to participate in voting. Number of years of education is a good predictor of likelihood to vote (Flanigan and Zingale 1983). Attainment of a college degree has been found to be a variable of Republican vote. Pomper 1976 found this to be true in the 1960 and 1952 presidential elections. Dye (1973) found. it to be true in. the 1960 and 1968 jpresidential elections. And Flanigan and Zingale (1983:69) found it to be "typical." Pae's (1986) results from presidential elections between 1952 and 1980 show education to be significantly correlated with voting behavior in seven of the eight elections. More highly educated voters may tend to vote for COPE- endorsed candidates because typically those endorsed support increased funding for education, according to Delaney, et al. 62 (1990) . Benign Although the variable region does not play as prominent a role in predictive studies relating to voting behavior as does, for instance, race or religion, a number of studies attest to mild correlations as well as direction of correlation, When Wattenberg (1987) looked for the existence of electoral coalitions in the 1984 presidential elections, little was gained from regional difference. Results showed the following percent Republican vote, by region: Northeast 55.1 percent, Midwest 58.4 percent, South 62 percent, Mountain 66.9 percent Pacific 57.7 percent. Pae (1986) did not find region to be a significant factor in Presidential Elections between 1960 and 1976. In.Pomper's (1976:72) investigation which divided data.by Northeastern, Midwestern, Southern, and Western regions of the U.S., no vast differences in the 1964 and 1960 presidential elections were noted, outside of the tendency for Western voters to favor Kennedy over Nixon.by 10 percent, while in the Northeast, Midwest and South, Nixon was the preferred candidate. A. Los Angeles exit poll, November ~8, 1988 indicates that Southern voters voted the Republican ticket in the 1984 and 1988 national elections at a higher rate than they voted the Democratic ticket. Gallup Opinion Index No. 183, December 1980 shows that in five of the nine presidential elections from 1952-1984, including 1980 and 1984, a higher 63 proportion of Southern voters voted for the Republican, than for the Democratic candidate. Flanigan. and Zingale (1983) utilizing' NES data successfully differentiated region by broadening categories, dividing them into the nation, the non-south, and the south. The south was significantly more strongly Democratic in its party identifications between 1952 and 1980 than was the nation as a whole or the non-south. The authors said this about regional differences: For many years the advantage that the Democrats enjoyed nationally was largely a result of the overwhelming majority of Democrats in the south, as shown in [their results]. Today, Democrats are still considerably stronger in the South than in the North, though the proportion of Republicans in the South has increased to about the same level as the North. Delaney, et al. (1990:627) used southerners and non- southerners as dummy variables for region, based upon the propensity for southerners to vote the Democratic ticket. The research at hand likewise aggregates data from the NES, originally presented in a state-by-state basis, into southern, non-southern. IDQQEQ Danielson and Murphy (1987:246) believe that there is some truth to the cliche that "the rich are Republicans and the poor Democrats," all things being equal. Disaggregated, though, there will be exceptions to the rule, the most glaring 64 of which may be the tendency of highly paid entertainers to be more liberal and to be aligned with the Democratic party, and the propensity of farmers (some of whom fall into middle or low income categories), to identify with the Republican party. Economic distinctions may have been at their broadest point during the New Deal. Campbell, et al. (1960) in their quintessential analysis of New Deal politics contained in m Wen, attributed the distinction to a purposeful appeal on the part of the Democratic party for the support of disadvantaged groups (as well as verbal assaults of "economic royalists") (156) . New Deal politics created alignments with the unemployed, the disadvantaged, and organized labor (Key (1958) ) . Whether or not income still has quite so profound of an effect on political affiliation or voting behavior will be examined in the study at hand. REEL! Iggnrifignrion Flanigan and Zingale (1983) remark that early studies regarding voting behavior exclude partisanship as a variable, concentrating instead on social variables. It was the Survey Research Center, in fact, whose 1952 General Election survey included party identification for the first time in a major study, which brought out the importance of including it in political behavior studies. And it was Campbell, et al. (1954), who initiated the inclusion of the party identification variable, which, according to Flanigan and Zingale, assumes a "central role in all voting behavior 65 analysis" (44). Any review of literature surrounding the notion of party identification would be remiss if not including the definitive writing of Campbell, et al. (1976), in Thg Anericnn ther. Setting aside an.entire chapter for analysis of "The Impact of Party Identification" (120-145), the authors conclude, innnr nlin, the following: - Party identification is antecedent to voting behavior. - The allegiance to one party over another is a psychological identification, and does not necessarily rely on formal party membership or support. - Most people in the United States have at least a "sense" of party attachment. - There is evidence of "great stability" in party inbetween elections. - The correlation between party identification and the vote tends to be high. - This high correlation is relatively undisturbed by a change in candidates or the nature of issues. - Partisanship "persists through time" for many individuals. - Few other variables have as great an impact on national elections in the U.S. as does party identification. - The strength and direction of political partisanship are central to political behaviors and attitudes. - The method of measurement of partisanship is an important consideration. It should be based upon self-reported loyalties (as with the SRC study at hand), rather than voting record. Pae's (1986) analysis of presidential elections from 1952- 1980, revealed party identification to be significantly 66 related to voting behavior in all national elections. Pomper's (1976) analysis of three presidential elections uncovered a high correlation between the proportion of voters who identified with either the Democratic or Republican party and vote for that same party. Axelrod's (1972) analysis of electoral coalitions between the years 1952 and 1968 revealed that unions are one of three coalition partners (along with the poor and blacks) in the Democratic party. By examining party identification, inferences may be drawn regarding suggested "shifts" in group ties to a particular party. The majority of political analysts have predicted a Republican victory in 1992, for the fourth consecutive presidential election in a row. Democrats have failed to garner the presidency in five of the last six elections, and the last Democratic president left the White House over eleven years ago. The W122], (Volume 73, No. 15:1) reported in its November 1, 1991 edition that polls then showed a 63 percent approval rating of President Bush's performance. This is a higher rating than normally needed one year prior to elections in order to predict a good chance at winning in 1992 (Devine, 1990). Masters and Delaney (1987) conducted interviews with lobbyists representing 10 unions as to their perception of the political performance of labor. Results of their findings show a nearly unanimous conclusion on the part of lobbyists that labor was politically disadvantaged in the 19803, as it has been since the political agenda-setting of President Reagan. One factor believed to 67 contribute to labor's political dilemma is a "conservative trend" in general public opinion. The public, it is contended, is "disenchanted" with Democratic party liberalism. Republicans are often perceived by the general public to be more capable than the Democratic party at creating conditions for a strong economy. In his analysis of CBS News and New York Times polls of 1984 (September 30 - October 4), Ladd (1985:8), finds that registered voters by a margin of two-to-one (54 to 27 percent) answered "the Republicans" when asked "Which party is...better able to insure a strong economy?" There is a plethora of "best guesses" as to why the Democratic party has been unable to garner the Presidency in five of the last six elections. One extreme View holds that it is in the party's "electoral interest" to lose the presidency, in a calculated attempt to maximize their "real goal" of controlling Congress (Erickson 1989:30). Other writers take a dimmer view, as does Schneider (1984:19) Democrats are in the position as the French aristocracy at the time of the Bourbon Restoration. We can pretend that nothing has happened, but the ancient regime is dead.. Things will never be the same. The problem for the Democrats is to avoid the fate of the Bourbons, of whom Talleyrand said, They learned nothing and they forgot nothing. Schneider's explanation for Democratic party success in House races is different than that of Erickson. Republicans, he writes, garnered a significant number of open seats -- two to one, as a matter of fact. Results of exit polls reveal that 68 in.the 1984, nearly as many votes were cast for Republicans as for Democrats. Yet due in part to a redrawing of Congressional district. boundaries under ‘the direction. of Democrats, a three to two House lead for the Democratic party resulted. Schneider concludes that a shift has indeed occurred, based upon results from exit polls. Four nationwide polls reveal that 36 percent of voters identified themselves as Democrats in 1984, whereas in 1980, 42 percent polled said.the same. 'Thirty-two percent in. the 1984 polls identified themselves as Republicans, whereas only 28 percent had said the same in 1980 (21). In what appears to be a common baseline election upon inquiring into shifts in party identification, the author compares the nature of the 1956 Adlai Stevenson vote to that of the 1984 Mondale vote. Although the percent of national vote stayed nearly constant (42 percent for the former, 41 percent for the latter candidate), the sources of support differed. Schneider lists these differences in terms of demographics (21): - Sources of support, Mondale Black voters College graduates Women Jews Professionals - Sources of support, Stevenson Whites Southerners Men Blue-collar workers Union members Catholics 69 The most devastating cause of the shift, asserts Schneider, is a loss of credibility for the Democratic party, in the area of economic issues. "Shifts in group ties," writes Ladd (1989:1) "often startle us." In.his analysis of the 1988 elections, variables affecting the favored status of Republicans in the presidential contest are examined. A. popular two-term Republican president and favorable public opinion as to the state of the economy appear’ to have precipitated overwhelmingly positive feelings toward.continued GOP support across divergent groups of voters. lPolls conducted in 1988 by CBS/New York Times and Roper indicate that a significant proportion of Americans felt that they, their family, and the economy were "better off" than previously. (Lad, 1989:3). Democrats were facing the proverbial uphill battle following eight years of perceived prosperity. This perceived prosperity may have contributed greatly to offset a historical tradition of party voting. Inmmpencx Incumbent Congressional candidates seeking re-election have won nine out of ten times since World War II, according to Banaian and.Luksetich (1991). In 1988, 98.3 percent of all Congressional incumbents were re-elected to another term in office, and of 409 Congressional incumbents seeking re- election, only one was defeated in the primaries (Cook 1990). Alger (1991) points out that one reason.why over 90 percent of 70 Congressional incumbents ‘were re-elected. in 1990 is the significantly higher "issue coverage" afforded incumbents by the media, a fact uncovered by Clarke and Evans in 1983. Alston (1990) attributes the over 90 percent re-election rate of Congressional incumbents, to Democratic Congressional Campaign (DCCC) spending strategies, which give "top priority" to incumbent protection (4235). Contributions of the DCCC to incumbents typically greatly exceed funds available to non- incumbents. Banaian and Luksetich (1991) found that such campaign spending advantages are significant determinants of votes received by candidates. In looking toward the 1992 Congressional elections, there are those critics who argue that the 90 percent plus re- election rate will not be repeated. Dwyer (1991) predicts retribution by way of election loss or limited legislative terms, for incumbent Congressional candidates. The bounced- check fiasco, refusal to pay Capitol Hill restaurant tabs, ticket fixing and other "scandals," he believes, will increase the odds of a lower win-rate for incumbents in 1992. And since "constituent trust" has been found to have significant influence and better predicts vote for an incumbent than does his or her party identification (Parker 1989) , such "scandals" may have significant negative impact on incumbent re-election. Cook (1991) argues that incumbent House members have already faced an uphill battle for re-election, when in 1990 52 incumbent Congressional candidates were re-elected "with their lowest winning percentage ever, an even greater number (57) 71 won with their lowest vote share since they were first elected" (483). Yet, as Gibbs (1990) points out, voters face a system which is stacked in favor of incumbents, so much so that the "anti-incumbent mood" of 1990 succeeded in decreasing the re- election rate by a mere two percent (from 98 percent to 96 percent). CHAPTER I I I METHODOLOGY In this chapter, three separate sections describe the data and methodology used, as well as hypotheses to be tested. Section one, entitled DATA, sets forth the origins of the data, the sample and sample design, purpose for the data collection and the advantages of its utilization in addressing the hypotheses contained in the study at hand. Section two, entitled THEORETICAL EXPECTATIONS discusses the proposed analysis of variables in the research. For the first analysis, the proposed analysis of the relationship between the independent variables (union status, gender, age, race, religion, education, occupation, region, income and party identification) and the criterion presidential vote will be discussed. For the second analysis, the proposed analysis of the relationship between these same independent variables plus an incumbency variable, and the criterion Congressional vote will be discussed. For the third and final analysis, the proposed analysis of the relationship of issue opinion to the criterion will be discussed. Section three, ANALYSIS, outlines the analytical approach.which has been determined to be most appropriate in this research. 72 73 DATA The data set was obtained from the Inter-University Consortium for Political and Social Research (ICPSR) located at the Survey Research Center and the Center for Political Studies, University of Michigan, Ann Arbor. During each presidential election.year pre- and post-election surveys are conducted in order to attain information.pertaining to voting behavior, public opinion, and participation. Core questions for these surveys comprise such items as likes/dislikes of presidential candidates, party identification, vote (intention, and/or actual) for President, vote for members of Congress, campaign activities, and various measures of values and predispositions (Introduction and Codebook, ICPSR #9196). The research at hand utilizes a number of new items which appeared in the 1988 survey. These are issue opinion items designed to measure respondents' attitudes toward issues such as taxes, the poor, health insurance, and foreign policy. The survey instrument is comprised of 1301 variables, most of which are questions directed to respondents, and the remainder are interviewer checkpoints, summarizations, and identifications. 'Variables selected for analyses include the following: 1) Demographic characteristic variables. 2) Respondent vote variables, which include respondent vote in 1988 Presidential and Congressional elections. 3) Issue opinion variables, which measure 74 respondents' opinions of various social and economic issues. Issues were selected based upon results of a qualitative analysis of formal AFL- CIO, COPE, and lobbying efforts in soliciting support for organized labors' stands on issues. Those issues of major importance to organized labor were drawn from the grggnnging§_gr_nnn_1221 AEL‘CIQ Convnntion: 1986, 1987, and 1988 C 'ona te 'e : and the 1987 and 1988 AEL-CIO Repprns on Congress and are analyzed as to their relationship with rank and file opinion of these same issues. The Report on Congress, published by the AFL-CIO each year, reviews the record.of Congress as it pertains to efforts to either support or deny support to labor issues. Efforts may include support/veto for labor-related legislation, or addendums/deletions to labor-related legislation. There are two major components to each.Reportn The first, "Major Issues in the House," analyzes Congress from the standpoint of bills introduced which are important to labor (e.g. Plant Closing Protections, Trade, Pay Equity, and Catastrophic Health Insurance). The second, "How Your Representatives (or Senators) voted," presents state-by-state coverage of exactly how each Congressman or Senator voted on.bills relating to the issues identified in "Major Issues." Salisburg (1984:64) remarks that the Congressional Quarterly is the "inside press" of American politics. These 75 publications indeed give the inside track on the lobbying agendas of four interest groups: The AFL-CIO, The Chamber of Congress, The Americans for Democratic Action, and the American Conservative Union. .Additionally, the Congressional Quarterly provides the "group ratings" (of candidates seeking election or re-election to Congressional, Senatorial, Vice Presidential or Presidential Office), of each of the aforementioned interest groups. Proceedings of the 17th AFL-CIO Convention set forth, among other things, organized labors' perspective on myriad social and economic issues. Analyses of the above sources allows me to identify a list of issues on the union agenda. Items in the NES issues section which matched this agenda were set aside for analysis. Unfortunately, not all union issues were included in the NES. However, the number' of union-relevant issues ‘which.‘were included are more than adequate to make for a rich study. The Sample and Sample Design The Survey Research Center provides the following description of the population and sampling techniques of the NES. STUDY POPULATION The study population for the 1988 NES is defined to include all United States citizens of voting age on or before the 1988 Election Day. Eligible citizens must have resided in housing units, other than military reservations, in the forty-eight contiguous states. This definition excludes persons living in Alaska or Hawaii and requires eligible 76 persons to have been both a United States citizen and eighteen years of age on or before 8 November 1988. MULTI-STAGE AREA PROBABILITY SAMPLE DESIGN The 1988 NES is based on a multi-stage area probability sample selected from the Survey Research Center's (SRC) National Sample design. Identification of the 1988 NES sample respondents was conducted using a four stage sampling process--a primary stage sampling of U.S. Standard Metropolitan Statistical Areas...and counties, followed by a second stage sampling of area segments, a third stage sampling of housing units within sampled area segments and concluding with the random selection of a single respondent from selected housing units (American National Election Study, 1988: Pre- and Post-Election Survey, Introduction and Codebook, X). THEORETICAL EXPECTATIONS The review of extant and current literature previously provided brought forth a gathering of research which nearly unequivocally ascertains that members of labor unions tend to identify with the Democratic party. AFL-CIO has endorsed the Democratic presidential candidate in each of the last three presidential elections. Further evidence from.the literature suggests strongly that union status is a significant variable in voting behavior and that union members tend to vote for Democratic candidates more than do non-members. These assumptions are elemental to formulation of hypotheses to be tested in the research at hand. Specifically, the relationship between union status and vote, as well as union status and stands on issue, will be examined. 77 Dependent Variables Three separate analyses are performed in this research each utilizing a distinct dependent variable. The criterion for the first analysis is presidential vote. This analysis allows us to examine whether union members are more likely to vote for the endorsed presidential candidate (Dukakis) than are non-members. Presidential vote is a categorical variable, coded as Bush = 0 and Dukakis = 1. The criterion for the second analysis is Congressional vote which allows us to examine the impact of endorsements on the voting behavior of union members and non-members in Congressional elections. Congressional vote is a categorical variable, coded as not endorsed = 0, union endorsed = 1. Care was taken to separate Congressional candidates according to COPE endorsements for 1988. COPE endorsements for Congressional candidates are found in Appendix A. The criterion for the third analysis is support of the union issue agenda. This analysis allows us to examine to what degree union members hold the opinions on issues that labor leaders attempt to disseminate throughout membership ranks. It also allows us to discern whether union members differ from non-members in support of union issues. Issues were treated individually (disagree with union stand = 0, agree with union stand = 1) and in combination. For the aggregated variable, respondents were given a score representing their percentage agreement with union stands. Respondents who answered fewer than one-half of the issue 78 questions were eliminated from the analysis. Independent Variables Independent variables utilized in the analyses are common to research of political behavior and attitudes. Union status, gender, age, race, religion, education, region, income, and party identification have been widely used in prior research. In ‘the third. equation. whose dependent variable is Congressional vote, the independent variable incumbency is added. 0 t tu Union status is coded union member = 1; non-member = 0. Although the NES also collects data on the union status of spouses and other household residents only actual union members were coded as such. ngder Gender is a binary variable, coded female = 0, male = 1. Age Age is measured as a continuous variable, beginning at the voting age of 18 years. 79 BEES There are two categories of race: white, and nonwhite. Both categories are represented as a dummy variable, with white, that category with the highest number of respondents, used.as the comparison.group. .Although most of the literature addresses a black versus white vote, some prior research focuses on all minorities. That information, therefore, will be available for the reader. Religion The religion variable includes four categories: Protestant, Catholic, Jewish, and other. Each of these is represented as a dummy variable, and is compared to Protestant, that category with the greatest number of respondents. Education The researcher followed the coding procedure used by Delaney, et al. (1990) who utilized data from the 1990 American National Election for their logit analysis of voting for COPE-endorsed candidates. In this way the education variable becomes continuous by assigning year values to the original categorical breakdown (e.g. high school = 12 years, some college = 13 years). 80 Region Region was coded into two categories: South = 0; Non- South = 1. lBQQEQ Income has been converted to an equal-appearing interval measure by recording each income category to its central value. Rnrry Identifigation Party identification is measured on a seven point ordinal scale ranging from Strong Republican through Strong Democrat. Codes are as follows: Strong Republican = 0, Republican = 1, Independent leaning Republican = 2, Independent = 3, Independent leaning Democrat = 4, Democrat = 5, and Strong Democrat =6. Party identification is important because union members tend.to identify with.the Democratic party and.to vote Democratic. Therefore, union member support for endorsed candidates may be indistinguishable from party identification. W Incumbency was coded into two categories: no incumbent in the race = 0; incumbent running for re-election = 1. 81 Research Hypotheses With the antecedent literature before us, the following three major null-hypotheses will be tested: H(1): Union members are no more likely to vote for AFL-CIO/COPE endorsed presidential candidates than are non-members. H(2): Union members are no more likely to vote for AFL-CIO/COPE endorsed presidential candidates than are non-members once party identification and other variables are controlled. H(3): Union members are no more likely to vote for AFL-CIO/COPE endorsed Congressional candidates than are non-members. H(4): Union members are no more likely to vote for AFL-CIO/COPE endorsed Congressional candidates than are non-members once party identification and other variables are controlled. H(5): Union members and non-members do not differ in their opinions regarding political issues. H(6): Union members are no more likely to agree with AFL-CIO/COPE stands on political issues than they are to disagree. Results will also indicate the significance of the variables in the models. The effect of individual independent variable on presidential vote, Congressional vote, and issue opinion will be discussed in Chapter IV. 82 ANALYSIS In order to determine whether or not union membership is a significant contributor to models that seek to explain respectively, presidential voting’ behavior, Congressional voting behavior, and opinions on selected issues, logistic regression (logit) has Ibeen selected. as the appropriate 2 Logistic regression assumes statistical tool for analysis. a non-linear function, and therefore a log-linear model is presented. Log-linear models are considered to be logit models when, as in the analysis at hand, each dichotomous dependent variable (presidential vote, Congressional vote, issue opinion) is assumed to be dependent upon a set of interval, ordinal, or nominal predictor variables (Knoke and Burke 1980). The technique results in logit coefficients, which in their raw form, are log odds. For interpretation these coefficients are converted into odds ratios, and may also be converted into probabilities (similar to the regression coefficients of ordinary least squares regression). The significance of each logit coefficient is tested using the t statistic. Because the SPSS X program which produces the logit analysis does not determine the proportion of variance explained, an ordinary least squares regression on the same set of predictor variables will be run separately, and the R“2 will be used as an estimate of multiple correlation. 2The analysis is conducted using the LOGIT model of the PROBIT procedure of SPSSX. 83 Because in the analysis which follows, maximum likelihood coefficients are converted into odds ratios regarding the association of the variable of interest (union versus nonunion), a discussion of the decision to use odds ratios seems appropriate. Morgan and. Teachman (1988) set forth. a concise and comprehensive presentation on the use of logistic regression in quantitative research. Their discussion of odds and odds ratios may be applied to the research at hand. Odds ratios, or the measures of association between components of nominal- level variables (e.g., presidential vote, Congressional vote, and issue opinion), are the basis for understanding logistic regression and log linear models. Logit allows one not only to determine the ratio of two odds (e.g. , union member odds/non-member odds) of some occurrence (e.g., voting for a particular candidate) but also to examine odds for each group (or category of a dichotomous variable, e.g., odds for union; odds for nonunion), as tnnnnnlyes representing ratios of events/nonevents. Simply, odds reflect the likelihood that a member of a group will (or will not) perform a particular behavior. This is reflected in the simple proportion of the group that performs the behavior. For example, if 55 percent of union members voted for Dukakis, the odds that a particular union member voted for Dukakis is 55/100 or .55. If the odds that a.non-member voted for Dukakis was .45, then the odds ratio of union member to non-members would be .55/.45 or 1.22. That 84 is, the odds that a union member would vote for Dukakis is 1.22 times the odds that a non-member would. On the other hand the odds ratio of non-union to union voting for Dukakis would be .45/.55 or .82. (That is, the union/non-union and the non-union/union odds ratios are inverses of one another, 1/.82=1.22 and 1/1.22=.82.). Odds ratios, then, can be interpreted as measures of association, except that an odds ratio of 1.0 indicates 0 or no correlation; the odds for one group is identical to that for the other; Odds ratios greater than.one indicate that the members of the group represented in the numerator are more likely to perform the behavior in question than members of the group represented in the denominator. Odds ratios less than one indicate just the opposite. As well, the value of the odds ratio indicates the multiplier or how much more (or less) likely members of one group are to perform the behavior than are members of the other group. In the example above, union members were 22% more likely to have voted for Dukakis than were non-members. The advantage of using this measure of association in analyzing the hypotheses at hand is obvious. Morgan and Teachman cite Feinberg (1985) in this respect. Odds ratios, Feinberg points out, are clearly interpreted, revealing the multiplicative change necessary for moving from one odds (of voting for Dukakis) to another odds (odds of voting for Dukakis for union members being 3 times the odds of voting for Dukakis for non-members). If the odds ratio is greater than 85 one there is evidence that union members are more likely to vote for Dukakis than non-members while an odds ratio less than one is evidence that non-members are more likely than members to vote for Dukakis. Values are easily interchanged to facilitate an understanding of the relationship between components of the nominal variable of interest (UNION STATUS may simply be = union or non-union), since Odds /Odds union non-union interchanged as 1/(odds /odds This may be seen as union non-union) ‘ odds ratios being invariant to changes in rows or columns. The advantage here, according to Feinberg, is that percentages across independent variables (member/non-member) may be clearly examined, even where multiplication occurs. Odds ratios, unlike mere analysis of percentage differences (as with crosstabulation analyses), are invariant to multiplication. Therefore, one will not misinterpret measures of association due to differing levels in subpopulations. Clearly, this type of analysis lends itself to the present study examining union member versus non-member behavior. SUMMARY Discussions of the data utilized, theoretical expectations of the researcher, and the type of analysis chosen to be most appropriate were discussed in Chapter III. In the first section, data origins and the advantages of utilizing it in testing the hypotheses were discussed. In the second section, theoretical expectations were set forth and the hypothesized 86 relationship between the criterion and independent variables forming three separate models were discussed. The last section introduced logistic regression as the type of analysis utilized, examined. its. appropriateness, and. detailed. its capabilities as they relate to testing the hypotheses. Chapter IV will present results of the analyses performed. CHAPTER IV Analyses of findings regarding the hypotheses as well as supplemental inferential statistics and interpretations regarding variables comprising the election models are presented in this chapter. AEEEAL_YQEQ Table 4.1 presents results of a bivariate analysis of the actual union vote for the 1988 presidential candidates. Table 4.1 Actual Vote in 1988 Presidential Elections Union Nonunion Endorsed Candidate 56.4% 45.2% Nonendorsed Candidate 42.1 53.3 n = 1227 xz== 6.07 p = .048 Results indicate that.56 percent of the union.vote went to the endorsed candidate, Michael Dukakis, while only 45 percent of the nonunion vote went to Dukakis. The results are significant (12 = 6.07, p = .048). The null hypothesis, that 87 88 union. members are no more likely to vote for endorsed presidential candidates than are nonmembers, is rejected. The crosstabulation and chi-square analyses include candidate and union status variables only and do not control for other variables, such as party identification, as does the logistic regression. With bivariate analysis, therefore, the union status variable does account for differences in voting behavior. However, when the logistic regression is performed, union status is no longer significant. Once other variables have been controlled for, union status does not explain variation in presidential vote. The null hypothesis, that union members are no more likely to vote for the endorsed candidate than are nonmembers once party identification and other variables have been controlled for, is accepted. Table 4.2 reports the results of the logit analysis for actual vote. A number of independent variables are significantly related to the criterion. These include party identification (logit. coef = - .8792, odds ratio = .42); region (logit coef. = .7024, odds ratio = 2.02); nonwhite (logit coef. = 1.4924, odds ratio = 4.45); income (logit coef. = 0020, odds ratio = .99); Catholic (logit coef. = .3438, odds ratio = 1.41); Jewish (logit coef. = 1.2912, odds ratio = 3.14). The odds ratio for the political party identification variable indicates that as one moves up the scale from Democrat (=0) to Republican (=6), the odds of voting for the endorsed presidential candidate significantly decrease. 89 Northerners were twice as likely as were Southerners to vote for the endorsed presidential candidate. The odds ratio for the region variable indicates that Southern voters tended to vote Republican in 1988. The odds of a Southern voter voting for the endorsed candidate were .49 that of a Northern voter . 90 Table 4.2 Logistic Regression of Actual Support for the Union Endorsed Presidential Candidate Independent Logit Odds Variables Coefficient Ratio Party Identification .8792 .42* Region .7024 2.02* Age .0012 1.00 Education .0416 1.04 Sex .1848 1.20 Nonwhite 1.4924 4.45* Income .0020 .99* Catholic .3438 1.41* Jewish 1.2912 3.63* Other Religion .7430 2.10 Union Member .0482 .9529 n = 1010 pseudo R2 = .52 * significant at the .05 or below. 91 The odds ratio for the race variable shows that nonwhites were almost 435 times more likely to vote for the endorsed candidate, Dukakis, than were white voters. The odds of a white voter voting for the endorsed candidate was .23 that of a nonwhite voter voting for him in the 1988 presidential election. Voters whose incomes were higher tended not to vote for the endorsed candidate. The odds ratio for the income variable indicates that as income increases, the odds of voting for the nonendorsed candidate also increase. Income was converted to an equal-appearing interval measure by recoding income categories to their central value. The odds of voting for the endorsed candidate decreases by a factor of .99 for each increment of income. The odds of voting for the nonendorsed candidate increases by a factor of 1.002 for each increment of income. The odds ratio for religion indicates that Catholics were 1% times more likely and Jews 3% times more likely to vote for the endorsed candidate than were Protestants. Conversely, Protestants were .80 as likely as were Catholics and .29 as likely as Jews to vote for the endorsed candidate. A regression analysiS‘was performed using the same set of variables as those in the logit analysis for actual presidential vote. The pseudo R2 is significant and shows that 52 percent of the proportion.of variance in.the dependent variable (actual vote) is explained. 92 a t n d V t Actual vote was combined with intended vote since actual vote response was taken during post-interview only, and only a portion (59 percent) of the original survey participants were sampled. Before these variables were combined for logistic regression analysis, the intended vote was compared to the actual vote for those responding to the question regarding their intended vote and the question regarding their actual (post-election) vote. The relationship between intended vote and actual vote indicates substantial agreement between these variables. That is, the intended vote is a substantial predictor of the actual vote (lambda = .83). It is argued that using the intended vote of those not asked about actual vote allows examination of the gnnirn sample. Ninety-two percent of those in ‘the complete sample 'who intended to vote for the endorsed candidate Dukakis did, indeed, vote for him. Ninety-three percent of those who intended to vote for the nonendorsed candidate, Bush, actually did. The relationship between intended and actual vote is significant (x2 = 782.6, p < .0005). Table 4.3 presents results of a bivariate analysis of intended union members votes for the 1988 presidential candidate. When asked the question: "Who do you think you 93 Table 4.3 Intended Vote in 1988 Presidential Elections Union Nonunion Endorsed Candidate 56.1% 47.1% Nonendorsed Candidate 43.9 52.9 n = 2040 xz== 5.47 p = .019 you will vote for in the election for president?" 56 percent of union members responded that they would vote for the endorsed candidate, Dukakis. Forty-four percent indicated that they would vote for Bush. The results are significant (12 = 5.47, p = .019) and nearly match the results of the actual vote (56 percent and 47 percent, respectively). Table 4.4 reports the results of the logit analysis for actual plus intended vote. Independent variables which are significantly related to the criterion are the following: party identification (logit. coef = - .8778, odds ratio = .42); region (logit coef. = .5636, odds ratio = 1.76); nonwhite (logit coef. = 1.01, odds ratio = 2.75); income (logit coef. = 0012, odds ratio = .99); Catholic (logit coef. = .3616, odds ratio = 1.44). 94 Table 4.4 Logistic Regression of Actual plus Intended Support for the Union Endorsed Presidential Candidate Independent Logit Odds Variables Coefficient Ratio Party Identification Region Age Education Sex Nonwhite Income Catholic Jewish Other Religion Union Member - .8778 .5630 .0048 .0372 .0172 1.0122 - .0012 .3616 .8898 .5650 .0510 .42* 1.76* 1.00 .99* 1.44* 2.43 n = 1541 pseudo R2 = .48 * significant at the .05 or below. 95 The odds ratio for party identification indicates again that as one moves up the scale from. Democrat (=0) to Republican (=6), the odds of voting for the endorsed presidential candidate significantly decrease. The odds ratio for the variable region shows that BS‘With the equation for actual vote, Northerners were more likely than.were Southerners to vote for the endorsed candidate. In the equation of combined vote, Northerners were 1 3/4 times more likely to vote for the endorsed candidate. Conversely, Southern voters were only .57 as likely as were Northern voters to vote for the endorsed candidate. Race is once again found to be a significant predictor of vote for the endorsed candidate, although at a lower odds ratio than for actual vote. Nonwhites were 2 3/4 times more likely to vote for the endorsed candidate than were whites. Conversely, the odds for white voters were only .36 of the odds for nonwhites to vote for the endorsed candidate. The odds ratio for the income variable indicates that voters whose incomes were higher tended not to vote for the endorsed candidate. Results parallel those found for actual vote. ‘The odds of voting for the endorsed.candidate decreases by a factor of .98 for each increment of income. Conversely, the odds of voting for the nonendorsed candidate increase by a factor of 1.002 for each increment of income. The odds ratio for the religion variable indicates that Catholics were 1 2/5 more likely to vote for the endorsed candidate than were Protestants. Conversely, the odds of a 96 Protestant voting for the endorsed candidate was .71 that of a Catholic. The pseudo R2 for this logistic regression is significant and shows that.48 percent of the proportion of variance in.the dependent variable (combined vote) is explained. n te Table 4.5 presents results of a bivariate analysis of union vote in 1988 Congressional elections. Results indicate Table 4.5 Actual vote in 1988 House Elections Union Nonunion Endorsed Candidate 65.2% 52.4% Nonendorsed Candidate 34.8 47.6 n = 588 12 = 7.01 p = .008 that 65.2 percent of the union vote went to COPE-endorsed candidates, while a smaller proportion (34.8 percent) of the nonunion vote went to endorsed candidates. The results are significant (x2 = 4.02, p = .045). The null hypothesis of no difference between union and nonunion vote in the Congressional elections, is rejected. The crosstabulation and chi-square analyses include candidate and union status variables only and do not control for other variables, most 97 notably, party identification, and incumbency. With bivariate analysis, therefore, the union status variable does account for difference in voting behavior. When logistic regression is performed, however, union status is no longer significant. Once other variables have been controlled for, union status does not explain variation in Congressional vote. The null hypothesis, that union members are no more likely to vote for the endorsed candidates than are nonmembers once party identification and other variables are controlled for, is accepted. Table 4.6 reports the results of the logit analysis for the 1988 Congressional elections. Those variables which are significantly related to the criterion include party identification (logit. coef = -.5200, odds ratio = .59; region (logit coef. = .6564, odds ratio = 1.93); incumbency (logit coef. = 1.1490, odds ratio = 3.16); Catholic (logit coef. = .4204, odds ratio = 1.52); nonwhite (logit coef. = .9698, odds ratio = 2.64). 98 Table 4.6 Logistic Regression of Support for the Union Endorsed Candidate for the House of Representatives Independent Logit Odds Variables Coefficient Ratio Party Identification - .5200 .59* Region .6564 1.93* Age - .0058 .99 Education - .0648 .94 Sex - .0134 .99 Nonwhite .9698 2.64* Incumbent 1.1490 3.16* Income .0004 1.00 Catholic .4204 1.52* Jewish - .0930 .91 Other Religion - .6652 .51 Union Member .2984 1.35 n = 532 pseudo R2 = .33 * significant at the .05 or below. 99 The odds ratio for the party identification variable indicates that Republicans are less likely than Democrats to vote for endorsed congressional candidates. As on moves up the scale from Democrat (=0) to Republican (=6), the odds of voting for the endorsed congressional candidate significantly decrease. Northerners were nearly twice as likely as were Southerners to vote for an endorsed candidate. 'The odds ratio for the region variable indicates that southern voters tended to vote for nonendorsed Congressional candidates. The odds of a Southern voter voting for an endorsed Congressional candidate were .56 that of a Northern voter. The odds ratio for the incumbent variable indicates that if the endorsed candidate is an incumbent, the odds of voting for the candidate were three times that of when they were not an incumbent. The odds of a Catholic voting for an endorsed Congressional candidate were one and one-half that of a Protestant voting for the endorsed Congressional candidate. The likelihood of a Protestant voting for an endorsed Congressional candidate were .66 that of a Catholic. The odds of a nonwhite voter voting for an endorsed Congressional candidate were over two and one-half times that of a nonwhite voter. Although union members are one and one-third more likely to vote for an endorsed Congressional candidate than are nonmembers, the effect of union.membership found in bivariate 100 analysis is no longer evident once party identification and other variables are controlled. Union status is not significant in the logistic regression. The pseudo R2 for this regression is significant and shows that 33 percent of the proportion of variance in the dependent variable (Congressional vote) is explained. Inrge ypting Models Cnipn_ugnpnr§nip. Table 4.7 combines the three voting models for the purpose of comparison. Regarding the independent variable of interest, union membership, it proved to be significantly related to voting behavior in all three models where controls were not applied. Once party identification and other variables were entered into a logistic regression, union membership ceased to be a significant predictor of either presidential or Congressional voting behavior. Logistic regression involving both intended.presidential vote and Congressional vote, however, do show a tendency on the part of union members to vote for the endorsed candidate. Even in the actual presidential vote model, the odds are close enough to 1 (log odds = .95) to indicate nearly equivalent odds of either a union or nonunion voter voting for the endorsed presidential candidate. The apparent absence of a "distinct union vote" does not support the findings of Axelrod (1972, 1982) Form (1982, 1983), Campbell, et al. (1980), Ra (1978) and.others (as cited 101 in Masters and Delaney, 1987). Its absence, however, does conform to the findings of Wolfe (1969) and to Brown and Catlett (1983) assertion that a labor vote was absent in the 1980 presidential elections. 102 Table 4.7 Combined Logit Results of Individual Voting Behavior Dependent Variables Odds of Voting for the COPE-Endorsed Candidate in the 1988 Elections Actual Actual Plus Independent Presidential Intended House Variables Vote Pres. Vote Election Party Identification .4151* .4157* .59* Region 2.0186* 1.7570* 1.9278* Age 1.001 1.0048 .9942 Education 1.043 1.0379 .9373 Sex 1.2030 1.0173 .9867 Nonwhite 4.4478* 2.7516* 2.6374* Incumbent - - 3.1550* Income .9980* .9988* 1.0004 Catholic 1.4103* 1.4356* 1.5226* Jewish 3.6371* 2.4346 .9112 Other Religion 2.1022 1.7594 .5142 Union Member .9529 1.0523 1.3477* * significant at the .05 or below. 103 92322;. Gender is not a significant predictor of voting behavior in any of the three models tested. The "gender gap" suggested by Ladd (1985) in the 1984 national election is not apparent in the 1988 national elections. Divergent results emerge regarding gender in the study at hand. In keeping with much of the literature prior to the 1980 national elections, females tended to vote for the Democratic candidate, albeit by a small margin, in the 1988 presidential election. in the 1988 House election, however, the odds of a female voting for the Democratic candidate were less than the odds of a male doing so. The Gallup Opinion Index NO. 183 indicated that in five of the nine national elections from 1952 - 1984, an equal or higher proportion of females than males voted for the Republican candidate. Mg. Age is not a significant predictor of voting behavior in any of the three models tested. The results of both presidential voting models are contrary to common findings from past elections (see Ladd (1989) and Pomper (1968)), which report that as one ages, the tendency to vote Republican increases, in that these models show equal likelihood of voting for either candidate as one ages. In the Congressional voting model, results indicate that as one ages, the odds of voting for the endorsed (Democratic) candidate increase. 229;. The odds being with nonwhites to support endorsed (Democratic) candidates is consistent with extant and 104 traditional literature. Findings support those of Pae's (1986) , wherein race was significantly related to a Democratic vote in all presidential elections between 1952 - 1982 , and Parent, et a1. (1987), Dye (1973) and Niemi and Meisberg (1976) in their findings that black voters tend strongly to vote Democratic and to be aligned with the Democratic party. 2211511913. The odds being significantly higher for a Catholic voter than for a Protestant voter to vote for the endorsed (Democratic) candidate in both the presidential and Congressional elections is in keeping with much of prior research. That Catholics and Jews tend to vote the democratic ticket supports the findings of Danielson and Murphy (1983) and Petrock (1981). The odds being significantly higher for a Jewish voter than for a Protestant voter to vote for the endorsed (Democratic) candidate is also supported in the literature. In the actual presidential voting model, Jewish voters were significantly more likely than were Protestant voters to vote for the endorsed candidate. And although the odds of Jewish voters casting a vote for the endorsed candidate was not significantly higher than it was for Protestants in the combined vote model, the direction is toward the endorsed candidate. WH- In the presidential election models, the logit analyses indicate that as one's education increases, so do the odds of their voting for the endorsed candidate. These 105 results support the notion of "intellectual" affiliation.with the Democratic party. Yet in none of the three voting models is the relationship between education and voting significant. In the Congressional voting model, results indicate that as one becomes more educated, they tend to vote for the nonendorsed (Republican) candidate, thereby supporting Pomper (1976) and Dye (1973). M911. In keeping with Flanigan and Zingale (1983) and Delaney et al. (1990), region was divided into North and South. Like those findings of Delaney, et al. regarding Senate and Gubernatorial elections, North was found to be a predictor of higher odds of voting for the COPE—endorsed candidate in the 1988 presidential elections. Unlike the Delaney, et al. finding of higher odds for Southerners voting for the endorsed Congressional candidate, results of the 1988 Congressional voting model indicate that being a Southern voter significantly increased. the odds of voting for’ a nonendorsed candidate. lngpmn. The odds that one's income would effect their voting behavior is significant in both presidential voting models. Results indicate that as income increases, the odds of voting for the endorsed candidate decrease. Rarty_1dentifignrign. Results of all three voting models are consistent with past findings, most notably Pae (1986) and 106 Pomper (1976) in that party identification is significantly related to voting behavior. The odds that a Democrat will vote the Democratic ticket are significantly higher than those of a Republican voting the Democratic ticket. The odds that a Republican will vote the Republican ticket are significantly higher than are those of a Democrat voting the Republican ticket. Issues A univariate hypothesis test utilizing the t-distribution indicates that when sixteen social, economic and foreign policy issues are aggregated, the difference between union member and nonmember opinion is significant (t = -2.75, p = .006, df = 2038). Therefore, when issues are combined, the null hypothesis, that union member and nonmember issue opinions do not differ, is rejected, When issues are combined, union. members tend. to agree ‘with .AFL-CIO/COPE opinion more than do nonmembers. Therefore, when issues are combined the null hypothesis, that union.member opinion is no more likely than is nonmember opinion to reflect the views of the AFL-CIO and COPE, is rejected. Results of analyses utilized to test for significant differences of opinion between members and nonmembers regarding ingiyignnl issues are presented in Tables 4.8 - 4.23. IDiscussion.following'each.table indicate whether'or not hypotheses pertaining to issues are to be accepted. Results from Table 4.8 indicate that there is no 107 significant difference (x2 = .0757, p = .78) between the two groups regarding opinion as to the appropriateness of Table 4.8 Central America Union Nonunion Agree AFL-CIO/COPE 61% 60% a ee AFL-CIO/COPE 39% 40% n = 1836 providing military aid to the Contras. The null hypothesis, that there is no difference of opinion regarding this issue is, therefore, accepted. The union member group of respondents is found to agree (61%) more than disagree (39%) with the notion that military aid to contras should be decreased, and the non-union group is found to agree (60%) more than disagree (40%) that it needs to be decreased. Proceedings resolution. 142 resolves ‘that the .AFL-CIO opposes further funding of Contras. Union group opinion is, therefore, commensurate with that of the AFL-CIO. The null hypothesis that union member opinion.will not reflect that of labor leaders is, therefore, rejected. Results from Table 4.9 indicate that there is no significant difference (x2== 2.985, p = .08) between the two groups regarding opinion as to government funding for 108 Table 4.9 Child Care Union Nonunion Agrgg AFL-CIO/COPE 64% 57% Qinngrgn AFL-CIO/COPE 36% 43% childcare. The null hypothesis, that there is no difference of opinion is, therefore, accepted. The union member group of respondents is found to agree (64%) more than to disagree (36%) with the proposal that federal funding for childcare should be increased. According to resolution 33 of the Proceedings, the AFL-CIO urged Congress to pass a comprehensive childcare measure, fund Title XX of the Social Security Act and insure childcare for mothers in need. Union group opinion is, therefore, commen- surate with that of organized labor. The null hypothesis, that union member opinion will not reflect that of labor leaders is, therefore, rejected. Results from 'Table 4.10 indicate that there is no significant difference (x2== .0240, p = .88) between the two groups regarding opinion as to governmental involvement in assistance to African.Americans. The null hypothesis, that 109 Table 4.10 Civil Rights Union Nonunion Agree AFL-CIO/COPE 24% 25% anngrnn AFL-CIO/COPE 76% 75% n = 1945 there is no difference is, therefore, accepted. The union member groupvof respondents is found.to)oppose (82%) more than to favor (18%) increased federal spending for programs to assist African Americans. The non-union group is found to oppose (80%) more than to favor (20%) an increase. The 1988 Report on Congress (1988 Report) shows the AFL- CIO to be in full support of the Civil Rights Restoration Act. The Proceedings, resolution 35 resolves that the AFL-CIO will work to restore and maintain both anti-discrimination laws and affirmative action. The AFL-CIO also calls for voluntary affirmative action plans. Union group opinion regarding government funding of programs to assist blacks is contrary to that of the AFL-CIO. The null hypothesis, that union member opinion does not reflect that of labor leaders is, therefore accepted. Results from Table 4.11 indicate that there is no 110 Table 4.11 Defense Spending Union Nonunion ree AFL-CIO/COPE 30% 28% Disagree AFL-CIO/COPE 70% 72% n = 1631 significant difference (x2==.2528, p = .62) between union.and non-union groups regarding opinion as to the issue of defense spending. The null hypothesis, that there is no significant difference is, therefore, accepted. The union member group of respondents is found to disagree with the proposal that funding for defense be increased, as is the non-union group. Resolution 110 of the Proceedings resolves that the AFL- CIO will continue a tradition of support for strong national defense. Furthermore, labor leaders are found to support research and development in the area of strategic defense systems. Union group opinion is, therefore, contrary to that of labor leaders. The null hypothesis, that union member opinion does not reflect that of labor leaders is accepted. Results from Table 4.12 indicate that there is no 111 Table 4.12 Drugs Union Nonunion ngrng AFL-CIO/COPE 79% 76% Qipngrng AFL-CIO/COPE 21% 24% n = 1983 significant difference (x2 = 1.1359, p = .29) between union and non-union group opinion regarding the issue of drugs. The null hypothesis that there is an no significant difference between the two groups is, therefore accepted. The union group is found to favor (79%) more than.oppose (21%) increased funding for the war on drugs. The non-union member group is found to favor (76%) more than oppose (24%) increases. The AFL-CIO (Proceedings, Resolution 208) urges Congress to develop additional comprehensive drug programs in the areas of drug crop eradication, interdiction of trafficking and shipments, education, treatment and law enforcement. Union member opinion is, therefore, commensurate with that of the AFL-CIO, and the null hypothesis of there not being' a similarity of opinion, is rejected. Results from Table 4.13 indicate that there is no 112 Table 4.13 Education Union Nonunion Agree AFL-CIO/COPE 70% 65% ninngrgg AFL-CIO/COPE 30% 35% n = 1994 significant difference (x? = 2.2774, p = .13) between union and non-union groups regarding opinion as to funding for education. The null hypothesis, that there is no difference is, therefore, accepted. The union member group of respondents is found to agree (70%) more than disagree (30%) with the proposition that federal budget allocation to public schools be increased, and the non-union.member group is found to agree (65%) more than disagree (35%). Both the 1986 and 1987 CQs review AFL-CIO opinion as to education. The AFL-CIO is found to support funding for Head Start and education and training programs. The Proceedings, resolution 34, supports federal funding for elementary, secondary and higher education. Union group opinion is, therefore, commensurate with that of the AFL-CIO. The null hypothesis, that it will not reflect the opinion of labor leaders is, therefore, rejected. 113 Table 4.14 Environment Union Nonunion Agree AFL-CIO/COPE 73% 62% Qipngrgn AFL-CIO/COPE 27% 38% n = 1979 Results from Table 4.14 indicate that there is a significant difference (x2== 9.0186, p = .002) between union and non-union group opinion regarding environmental issues. The null hypothesis, that there is no significant difference is, therefore, rejected. The union-member group is found to support (73%) increased federal spending on the environment more than to oppose it (27%). The non-union member group is also found to support increased federal spending (62%) at a higher rate than to oppose it (38%). In the Proceedings, the AFL-CIO urges the government to expedite waste site cleanup and supports myriad programs aimed at protecting the environment. Union group opinion is, therefore, commensurate with that of the AFL-CIO. The null hypothesis, that.member opinion.does not reflect that.of labor leaders, is rejected. Results from Table 4.15 indicate that there is no 114 Table 4.15 Food Stamps Union Nonunion Agnes AFL-CIO/COPE 22% 25% Qinngrgn AFL-CIO/COPE 78% 75% n = 1937 significant difference (x2 = .5798, p = .45) between union and.non-union.group»opinion regarding the issue of funding for food stamps. The null hypothesis, that there is no significant difference is, therefore, accepted. The union- member group is found to oppose (75%) more than support (25%) increased funding for food stamps. The non-union group is found to oppose (78%) at a greater rate than support (22%) increased funding. The AFL-CIO (Proceedings, Resolution 32) urges that relief for those in need of nutrition be brought about through increases in benefit levels of food stamps, increased emergency food assistance for the homeless, enactment of the Hunger Relief Act of 1987 and assistance for outreach programs. Union group opinion is, therefore, opposed to that of the AFL-CIO. The null hypothesis, that union member opinion does not reflect that of labor leaders, is accepted. 115 Table 4.16 Healthcare Insurance Union Nonunion 59222 AFL-CIO/COPE 45% 42% Qinngrnn AFL-CIO/COPE 55% 58% n = 1707 Results from 'Table 4.16 indicate that there is no significant difference (xz== .3923, p = .53) between the two groups regarding opinion as to government-funded healthcare. The null hypothesis, that there is no significant difference, is accepted, .A higher portion of union.members (56%) is found to agree than disagree ((44%) with the proposal that health insurance should be left to private enterprise rather than being government-funded. A higher proportion of the non-union member group of respondents is found to agree (58%) than disagree (42%) the idea.that insurance is best left out of the hands of the government. The 1988 Report, 1988 CQ and the Proceedings disclose an unequivocal support on the part of the AFL-CIO and COPE, for catastrophic health insurance, unlimited free hospitalization and payment of physician bills and a national health care program for those who are unable to afford adequate healthcare. The 1988 CQ shows support by labor leaders for the enactment of the Minimum Health Benefits for All Workers 116 Act. Union group opinion is, therefore, contrary to that of the AFL-CIO. The null hypothesis, that union member opinion will not reflect that of labor leaders, is accepted. Table 4.17 Homelessness Union Nonunion Agree AFL-CIO/COPE 70% 67% nigngrng AFL-CIO/COPE 30% 33% n = 1967 Results from Table 4.17 indicate that there is no significant difference (x2==.5257, p = .47) between the two groups regarding federal funding for the homeless. The null hypothesis, that there is no significant difference between the groups as to this issue, is accepted. The union member group of respondents is found to agree (69%) more than disagree (33%) with the proposal that fund allocations should be increased. The non-member group of respondents is found to agree (67%) more than disagree (33%). Two resolutions from the Proceedings deal with the homeless. Resolution 55 calls for governmental assistance in sheltering and providing supplies for the homeless. Resolution 64 urges Congress to increase funding for housing and social services. Union group opinion is, therefore, 117 commensurate with that of the AFL-CIO. The null hypothesis, that union members opinion will not reflect that of labor leaders, is rejected. Table 4.18 Import Quotas Union Nonunion AQIEQ AFL-CIO/COPE 71% 67% Qinngrgg AFL-CIO/COPE 29% 33% Results from 'Table 4.18 indicate that there is no significant difference (xz==1.0362, p = .31) between.the two groups regarding opinion of import quotas. The null hypothesis, that there is no difference of opinion regarding this issue is accepted. The union member group of respondents is found to favor (71%) more than oppose (29%) limits on foreign imports. The non-union member group is found to favor (67%) more than oppose (33%) limits. The AFL-CIO position on the subject of import limitations is expressed in the 1988 Report, the 1986 and 1988 CQs and the Proceedings. In its Report, the AFL-CIO supports a 1% limit on the import of 185 types of textiles and apparel; in the CQs, it supports textile, shoe and apparel import quotas, and in Resolution 12 of the Proceedings, the AFL-CIO seeks relief 118 from export practices of foreign governments seeking to profit at our expense. Union group opinion is, therefore, commensurate with that of the AFL-CIO. The null hypothesis, that union member opinion will not reflect that of labor leaders, is rejected. Table 4.19 Preferential Hiring Union Nonunion Agree AFL-CIO/COPE 17.8% 20.3% Qinngrng AFL-CIO/COPE 82.2% 79.7% n = 1668 Results from 'Table 4.19 indicate that there is no significant difference (x2== .5913, p = .44) between the two groups regarding opinion as to giving preference in hiring and promotion to African Americans. The null hypothesis, that there is no difference is, therefore, accepted. The union member group of respondents is found to oppose (82.2%) more than to favor (17.8%) preferential hiring and promotion. The nonunion group is found to oppose (79.7%) more than to favor (20.3%) preferential hiring and promotion. The Proceedings, resolution 35 resolves that the AFL-CIO will work to resolve and maintain affirmative action, and in fact calls for voluntary affirmative action plans. Union group opinion is, therefore, contrary to that of the AFL-CIO. 119 The null hypothesis, that union member opinion does not reflect that of labor leaders is accepted. Table 4.20 Research Funding, AIDS Union Nonunion Agnes AFL-CIO/COPE 74% 74% nipngrnn AFL-CIO/COPE 26% 26% n = 1960 Results from ‘Table 4.20 indicate that there is no significant difference (xz = .0510, p = .82) between the opinion of the two groups regarding‘ the issue of AIDS research. The null hypothesis, that there is no significant difference is, therefore accepted. The union member group of respondents is found to favor (74%) more than oppose (26%) increased spending for AIDS research. The non-member group is found to favor (74%) more than oppose (26%) increased spending. In the Proceedings, the AFL-CIO urges there to be additional funding for.AIDS research, education and services. It also urges there to be increased funding for research into its cause, development and method of transmission. Union group opinion is, therefore, commensurate with that of the AFL-CIO. The null hypothesis, that union member opinion on 120 AIDS research does not reflect that of labor leaders is, therefore, rejected. Table 4.21 Social Security Union Nonunion ngrng AFL-CIO/COPE 63% 59% ninngrng AFL-CIO/COPE 37% 41% Results from 'Table 4.21 indicate that there is no significant difference of opinion between the two groups regarding the issue of social security (xz==1.5326, p = .22). The null hypothesis, that there is no significant difference, is, therefore, accepted. A higher proportion of the union member group (63%) is found to favor more than oppose (37%) increased federal spending for social security. The non-union member group is found to favor increased federal spending at a rate of 59 percent. In its Proceedings, the AFL-CIO urges the repeal of raising minimum age requirements for receiving social security, a marital earnings share and improved benefits formulas. Union member opinion is, therefore, in agreement with that of the AFL-CIO. The null hypothesis, that union member opinion does not reflect that of labor leaders, is 121 rejected. Table 4.22 Strong Military Union Nonunion Agrng AFL-CIO/COPE 80.8% 79.3% Qigagree AFL-CIO/COPE 19.2% 20.7% n = 1765 Results from 'Table 4.22 indicate that there is no significant difference (xz==.2158, p = .64) between union.and nonunion groups regarding opinion as to the issue of a strong military. The null hypothesis, that there is no significant difference is, therefore, accepted. The union member group of respondents is found to agree (80.8%) with the assertion that it is important to have a strong military force in the U.S., as is the nonunion group (79.3%). In Proceedings, resolution 37, the AFL-CIO "reiterates its traditional support for a strong national defense" (281) in order to defend democracy. Union group opinion is, therefore, commensurate with that of labor leaders. The null hypothesis, that union member opinion does not reflect that labor leaders is accepted. 122 Table 4.23 Unemployment Union Nonunion Agrgn AFL-CIO/COPE 34% 32% nrnngrgn AFL-CIO/COPE 66% 68% n = 1964 Results from Table 4.23 indicate that there is no significant difference (12 = .4196, p = .52) between the opinions of the two groups regarding the issue of unemployment. The null hypothesis, that there is no difference is, therefore, accepted. The union member group is found to be opposed (66%) more than to favor (34%) increased federal assistance to the unemployed. The non-union member group is found to be opposed (68%) more than to favor (32%) increased federal assistance to the unemployed. In the Proceedings (Resolution 27) the AFL-CIO finds unemployment benefits programs to be underfunded, and urges additional funding for unemployment insurance and establishment of federal minimum benefit standards. Union member opinion is, therefore, contrary to that of the AFL-CIO. The null hypothesis, that union member opinion will not reflect that of labor leaders, is accepted. CHAPTER V SUMMARY, CONCLUSIONS, AND IMPLICATIONS The purpose of this research was twofold: first, as a descriptive study to present (a) the proportion of union members who voted for the COPE-endorsed presidential and Congressional candidates in 1988; (b) the proportion of members compared to nonmembers who voted for the COPE-endorsed presidential and.Congressional candidates in 1988; and (c) the proportion of union members who in that same year, agreed with labor's stand on issues. Second, as a predictive study, to analyze empirically (a) the effects of union status, gender, age, race, religion, education, region, income and party identification on voting behavior in the 1988 presidential and Congressional elections; (b) whether' or not there is a significant difference between member and nonmember voting behavior; and (c) whether or not there is a significant difference between member and nonmember issue opinion. To fully investigate each of these areas, extant and traditional literature and prior quantitative research was reviewed in Chapter II. In Chapter III, methodology and data utilized in the research were set forth, and theoretical expectations outlined. Analyses of findings regarding the 123 124 hypotheses as well as supplemental inferential statistics and interpretations regarding variables comprising each model were presented in Chapter IV. Summary Presidentialiete Bivariate analysis of the relationship between union status and presidential vote indicates that a higher proportion of union members than nonmembers voted for the endorsed candidate, Michael Dukakis, in the 1988 presidential elections. Chi-square analysis reports that the relationship between union status and presidential vote is significant at the .05 levele A significantly larger proportion.of the union vote was cast for the endorsed candidate than for the nonendorsed candidate. In an attempt to predict the relationship between union status and actual vote, a number of control variables common to voting studies were included in a logistic regression equation. Once gender, age, race, religion, education, region, income and party identification were controlled for, the significance and direction of the relationship between union status and actual presidential vote changed. The relationship was not significant and the direction of the union vote is extremely mild toward the nonendorsed candidate. Variables which are significantly related to the criterion include party identification, region, race, income and religion. 125 Bivariate analysis of the relationship between union status and intended presidential vote indicates that a larger proportion of union members than nonmembers intended to vote for the endorsed candidate, Michael Dukakis, in the 1988 presidential election. Chi-square analysis reports that the relationship between union status and intended presidential vote is significant at the .05 level. Once again, a logistic regression was performed in an attempt to examine predictive relationships between a number of independent variables and the criterion (combined vote). When gender, age, race, religion, education, region, income and party identification were controlled for, the significance between union status and vote disappeared. The relationship was not significant, however, the direction of union member vote remained (albeit extremely mild) toward the endorsed candidate. Variables which are significantly related to the criterion are party identification, region, race, income and religion. s'on vote Bivariate analysis of the relationship between union status and Congressional vote indicates that a higher proportion of union members than nonmembers voted for COPE- endorsed Congressional candidates in the 1988 elections. Chi- square analysis indicates that the relationship between union status and Congressional vote is significant at the .05 level. A higher proportion of union voters voted for the endorsed 126 candidates than for the nonendorsed candidates. In an effort to predict the relationship between union status and Congressional vote, a number of control variables common to voting studies were included in a logistic regression equation. With gender, age, race, religion, education, region, income, incumbency and party identification controlled for, union.atatus was no longer significante ‘Union status in.the Congressional voting model was not a.significant predictor of voting behavior in House elections, even though the odds of union.members voting for COPE-endorsed candidates were higher than those of nonmembers voting for COPE-endorsed candidates. Variables which were significantly related to the criterion included party identification, region, incumbency, nonwhite and Catholic. ISSSSS Majority of union member opinion was commensurate with that of labor leaders' in approximately 63% of the designated issues. A higher proportion of union members agreed than disagreed with labor leader pronouncements regarding the following issues: budget deficit, Central America, childcare, drugs, education, environment, homelessness, import quotas, AIDS research, social security, and a strong military. A higher proportion of union.members disagreed than agreed with labor leader pronouncements regarding the following issues: civil rights, defense spending, food stamps, healthcare insurance, preferential hiring and unemployment. When the 127 issues are tested individually, the differences between member and nonmember opinion were significant for only one issue -- the environment. However, when all 16 issues were tested together, union status was a significant variable. Union members supported labor leader pronouncements regarding issues overall, to a higher and more significant degree than did nonmembers in 1988. Results of the aggregate analysis may be due to the additive effect of small individual issue differences coupled with the significance on the environmental stand. CONCLUSIONS AND IMPLICATIONS Bivariate analyses show a significant relationship between union status and voting behavior. However, because party identification and other relevant variables are not controlled, the direct effect can not be measured in bivariate analyses. What they show us is, in part, the relationship of union status to voting Democratic or for incumbent candidates. Once controls are added, the direct effect of union status on voting for the endorsed candidate disappears. Union membership has no effect on voting for endorsed candidates either for president or for Congressional seats. And on all issues except one, the environment, there is no significant difference between member and non-member opinion. For nearly 40% of issues, union member position is not commensurate with that of the AFL-CIO and COPE. These are for the most part, social justice issues upon which the stands are liberal. 128 Alternative explanations of the results present themselves. It may be that labor leader pronouncements and candidate preferences do not reflect those of the majority of the rank and file. In this regard, Michel's (1915) theory of organizations may be applied. the assumed inevitability of a power elite and decision making by the relative few mirrors a scenario jpresented. by ‘Wilson (1979) of COPE discussions regarding .issues. stands, incumbent ratings and. candidate endorsements. An alternative explanation may be that communication between leaders and.members about the significance of backing certain issues and candidates needs strengthening. The 1985 AFL-CIO report on."The Changing Situation.of‘Workers and.Their Unions" acknowledges the need for strong-interaction between members and national leaders, so that leadership remains "...attuned to the desires of the members (23)". Members would also remain attuned to desires of leaders. Political conferences, local, regional and national events and conferences are means, suggested in the Report, of increasing communication. More needs to be done, according to the Report, by way of distributing opinion surveys. Of all voting coalitions reviewed in the literature and used as independent variables (e.g. African Americans, Southerners, lower income voters), unions have a distinct advantage. They are a social movement grganigntion (Zald and McCarthy 1980) as well as a social movement. There is no formal "leader" of women, "leader" of minorities or "leader" 129 of the South, even though there may be organizations, such as the National Organization of Women, that represent segments of voting coalitions. But there nrg labor leaders, elected to positions which empower them to represent and "lead" labor. The opportunity is available, therefore, to increase interaction between leaders and the rank and file through formal channels--to acquaint members with issue stands and to gather opinions from members. Suggestions for Further Research None of the multivariate analyses showed union status to be a significant variable of voting behavior in the 1988 elections. In including control variables common to voting behavior research in the logit analysis, the significance found in bivariate analysis disappeared. This suggests that in predictive studies of union member voting behavior, certain variables (most notably party identification) must be included in future models. Most union "bloc" vote conclusions were derived from bivariate, descriptive statistics and may instead reflect the tendency of union members to vote Democratic and to be affiliated with the Democratic party. In the Congressional voting model an incumbency (controlling for the presence of an incumbent in the race) variable was added, and.may'need.to be carefully considered in any future analysis of ‘voting’lbehavior' in. Congressional elections. Over 90% of Congress persons are re-elected to office and myriad studies provide evidence of its significance 130 to voting behavior. A.valuable component of this research.was its analysis of the relationship between labor leader pronouncements and member issue opinion. Without research of this type, it has been difficult to assess the effectiveness of AFL-CIO/COPE strategies regarding support for issues of importance to labor; Findings indicate that union.:member' opinion. is commensurate with that of labor leaders when issues are aggregated. By analyzing issues separately, this research allows one to probe into which issues are supported by union members and which are not. Three features of this research addressed those deficiencies set forth by Masters and Delaney (1987:346-350) regarding former analyses of union member voting behavior. Each should be considered when conducting future research on this topic. First, a deficiency regarding generalizability of findings was dealt with by choosing for analysis a sample of voting age citizens randomly drawn from the 48 contiguous states. Research using as a data base information derived from one or a small set of unions may be lacking in potential for generalizability to the population. A second methodological flaw of prior research and a potential hinderance to validity has been the collection of voting behavior and issue opinion data from union members by their union or a local council of the AFL-CIO. Because data for the research at hand.was gathered by a major university and not.by a union or collection of unions, members may have been more 131 inclined to respond as they actually felt rather than along "union" ideological lines. IFinally, the research.at hand used comparison-group data, thus allowing comparison between groups--an analytical dimension lacking in much of prior research. Logistic regression is an appropriate analysis technique when the criterion is assumed to be dependent upon a set of interval, ordinal or nominal predictor variables, according to Knoke and Burke (1980). Logit should be particularly useful in future models of voting behavior, since it allows for determination not only of the ratio of two odds (union.member odds/non-member odds) of an occurrence (voting for the endorsed candidate), but it also allows for determination of odds for each group (odds for union; odds for nonunion), as themselves representing ratios of events (voting for the endorsed candidate/not voting for the endorsed candidate). Odds ratios derived from logistic regression are, as attested to by Feinberg (1985), capable of revealing multiplicative change necessary for moving from one odd (of vote for the endorsed candidate) to another odd (odds of voting for the endorsed candidate for union members being 3 times the odds of voting for the endorsed candidate for nonmembers). Suggestions for the Practitioner Because 1992 is a national election year and is the national election immediately following that from which the NES data for this research is derived (1988), findings should 132 be: of interest for labor leaders, political candidates, political analysts and labor lobbyists. As stated in Chapter I, a Democrat-controlled Congress is advantageous for organized labor. Union members voted for endorsed Congressional candidates at a higher rate than did nonmembers, and also at a higher rate than they did for non- endorsed candidates. However, once party identification and other variables are controlled for, there is no effect. The bivariate analysis may reflect the tendency on the part of union members to identify with the Democratic party. The implication here may be that COPE/AFL-CIO strategy needs to target union and nonunion voters alike, better disseminating its message to the population as a whole, on a state-by-state basis. Apparently, existing strategy swayed union voters, but individual choice dictated a larger vote for nonendorsed candidates in 1988. Because both union and nonunion voters agreed with labor leader pronouncements regarding the majority of issues (aid to Contras, childcare, drugs, education, environment, homelessness, import quotas, AIDS research, social security and a strong military), this research suggests a strategy focused on tying COPE-endorsed candidates to issues. As pointed out by Delaney, et al. (1990), COPE usually rates favorably those candidates who support the AFL-CIO/COPE issue agenda (e.g. subsidization of housing for the elderly, daycare spending, civil rights legislation). Making voters aware of endorsed-candidates' stands on issues through mass media and 133 candidate meetings prior to election and election day distribution of issue information may capture a larger share of the vote. When confronted with a full platform of issues, union members tend to follow labor leader pronouncements. Therefore, exposing union members to endorsed candidates and their stand on issues well ahead of election day appears to be focused strategy worth consideration. Such a strategy may be extremely worthwhile in light of the fact that Senator Al Gore, Vice Presidential candidate on the Democratic ticket for the 1992 national elections, has been a "friend" of labor for many years. Because the current political process of labor relies on COPE ratings, the exemplary record of Mr. Gore in this regard should be made known to the rank and file. In the seven years from 1985 to 1991, he never once fell below 83 in AFL-CIO ratings. In 1987 his score totalled to a perfect 100 (CQ July 11, 1992:2025). Whether or not individual voters are aware of his ties to individual issues of importance to working persons, just knowing that he rated highly within labor's realm may at least be of interest to those voters who separate labor and business into two distinct and opposing camps. Targeting education of the rank and file as to the importance of certain issue stands would help eliminate perceived divisiveness among union members -- a necessary strategy for receipt of legislative support subsequent to elections, according to Delaney and Masters (1991:317) . Where currently political activity varies greatly by issue and union 134 (333), a more unified perspective may become the most practical objective of national and state AFL-CIOs and COPE in coming elections. Candidate stand on issues is no small consideration in Congressional elections. In 1988, Jolene Unsoeld (Democrat, Washington State) narrowly won election into Congress, according to Idelson (1990), and is on shaky ground with loggers and those involved with the timber industry, for not taking a strong stand against environmental protection for spotted owls. Alston (1990) reports that Democrat Sidney Yates (Illinois) won in 1990 with 70 percent of the vote, took a strong environmental protection stand, and was backed by $750,000 in campaign contributions. Campaign contributors included labor unions ($75 ,000 in PAC contributions) and environmentalists. Of the 17 issues investigated in the study at hand, one--the environment--was significantly related to union status (union members favoring environmental protection), when analyzed separately. Results from the bivariate analysis of actual presidential vote indicated that in 1988, a higher proportion of union members than nonmembers voted for the endorsed candidate, Michael Dukakis. Juravich and Shergold (1988:374) found that in 1984, unions had "surprisingly strong influence" on President choice by union members. Factors which influenced member voting behavior in their study included active participation in unions and the receipt of literature or telephone calls from union representatives regarding the 135 election, In‘order to increase the odds of:members voting for the endorsed candidate, COPE/AFL-CIO strategy should include intensive efforts to educate voters through the distribution of candidate-centered literature and telephone calls regarding the endorsed candidate and their stand on issues. It could be that, in final analysis, the current labor political process will prevail. Perhaps Michel's (1915) assessment of the inevitability that power will reside in the hands of the relatively few in leadership positions, is apparent. It could be that, in the words of Pizzorno: Ordinary voters must, at least in part, accept the political definition of their interests suggested to them by the members of organizations, although these latter must, at least in part, accept the definitions of political goals handed down by activists and leaders (Berger:255). If so, then an even stronger argument for increased interaction between members and leaders is present. APPENDIX APPENDIX CO ITTE N POLITICAL EDUCATION SENATE, HOUSE AND GOVERNOR ENDORSED ENCUMBENTS AND CHALLENGERS £1212 C2 CANDIDATE P AL 01 John Tyson AL 04 Tom Bevill AL 05 Ronnie Flippo AL 06 Ben Erdreich AL 07 Claude Harris AR 01 Bill Alexander AR 02 Tommy Robinson AR 04 Beryl Anthony AZ 02 Morris Udall AZ 03 David Moss AZ 05 Judith Belcher A2 S1 Dennis DeConcini CA 01 Douglas Bosco CA 02 Wayne Meyer CA 03 Robert Matsui CA 04 Vic Fazio CA 05 Nancy Pelosi CA 06 Barbara Boxer CA 07 George Miller CA 08 Ronald Dellums CA 09 Fortney Stark CA 10 Don Edwards CA 11 Tom Lantos CA 12 Anna Eshoo CA 13 Norman Mineta CA 14 Patricia Malberg CA 15 Tony Coelho CA 16 Leon Panetta CA 17 Vincent Lavery CA 18 Richard Lehman CA 19 Gary Hart CA 22 John Simmons CA 23 Anthony Beilenson CA 24 Henry Waxman CA 25 Edward Roybal CA 26 Howard Berman CA 27 Mel Levine 136 UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU 137 ITT O T C SENATE, HOUSE AND GOVERNOR ENDORSED ENCUMBENTS AND CHALLENGERS CA CA CA CA CA CA CA CA CA CA CA CA CO CO CO CO CO CT CT CT CT CT CT CT DE FL FL FL FL FL FL FL FL FL FL FL FL FL GA GA HI HI HI IA IA IA $22 28 29 30 31 32 34 36 38 41 44 45 $2 01 02 03 04 06 01 02 03 04 05 06 $1 04 06 08 10 11 13 14 15 16 17 18 19 81 04 06 01 02 $2 01 02 03 QANDIDAIE Julian Dixon Augustus Hawkins Matthew Martinez Mervyn Dymally Glenn Anderson Esteban Torres George Brown Jerry Yudelson Dan Kripke Jim Bates Pete Lepiscopo Leo McCarthy Pat Schroeder David Skaggs Ben Campbell Charles Vigil Martha Ezzard Barbara Kennelly Sam Gejdenson Bruce Morrison Chris Shays Joseph Marinan James Griffin Lowell Weicker Jacob Kreshtool Bill Chappell Jon Mills Bette Wimbish David Higginbottom Bill Nelson Jack Conway Harry Johnston Mike Kuhle Lawrence Smith William Lehman Claude Pepper Dante Fascell Buddy MacKay Ben Jones David Worley Mary Bitterman Daniel Akaka Spark Matsunaga Bill Gluba Eric Tabor Dave Nagle E UUUUUUUUUUUUUUUUUUUUDUWOUNUUUUUUUUUUUUUUUUUUUU ASEEEESESEEEEEEEFfiEEEfiEE 138 E PO SENATE, HOUSE AND GOVERNOR ENDORSED INCUMBENTS AND CHALLENGERS QANDIDAIE Neal Smith Gene Freund David O'Brien Jeanne Givens Richard Stallings George Sangmeister Jerry Costello Glenn Poshard Peter Visclosky Philip Sharp Tom Ward Jill Long Jim Jontz George Thomas Holland Mark Waterfill Frank McCloskey Lee Hamilton Andrew Jacobs Evan Bayh Jack Wickes Jim Slattery Dan Glickman Carroll Hubbard Faye Williams Silvio Conte Richard E. Neal Joseph Early Barney Frank Chester Atkins Nicholas Mavroules Edward Markey Joe Kennedy Joseph Moakley Gerry Studds Brian Donnelly Edward Kennedy Roy Dyson Ben Cardin Tom McMillen Steny Hover Kweisi Mfume Peter Franchot Paul Sarbanes Joseph Brennan Kenneth Hayes George Mitchell U E UUUUUUUUUUUUUUUUUUUUUSUUUUUUUUUUUUUUUUUUUUUUUUU C2 01 02 03 04 06 07 08 09 10 11 12 13 14 15 16 17 18 81 02 03 04 05 06 07 08 SI 01 04 05 06 07 08 09 81 02 04 $1 02 $1 01 04 05 06 07 139 CO TT N PO ITIC D CA ION SENATE, HOUSE AND GOVERNOR CANDIDATE John Convers Lana Pollack Howard Wolpe Norman Rivers Bob Carr Dale Kildee Bob Traxler David John Gawron Mathias Forbes Mitch Irwin David Bonior George Crockett Dennis Hertel William Ford John Dingell Sander Levin Gary Kohut Donald Riegle Doug Peterson Dave Carlson Bruce Vento Martin Sabo Gerry Sikorski Mary Hanson James Oberstar Hubert Humphrey III William Clay Ike Skelton Alan Wheat Doug Hughes Max Bacon Wayne Cryts Harold Volkmer Betty Hearnes Jay Nixon Mike Espy Mike Parker Wayne Dowdy Buck O'Brien Tom Judge John Melcher Walter Jones David Price Steve Neal Tom Gilmore Charlie Rose ENDORSED INCUMBENTS AND CHALLENGERS "D UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU a K: 140 QQHHlIIEE_QN_£QLIIIQAL_EQQQAIIQN SENATE, HOUSE AND GOVERNOR ENDORSED INCUMBENTS AND CHALLENGERS ififififiifififififififififififififiéé 08 09 11 $1 01 02 03 $2 01 02 01 02 03 04 06 07 08 09 10 14 82 01 $2 01 02 81 01 03 04 06 07 09 11 12 14 16 17 18 20 21 22 23 QANDIDATE W. G. Hefner Mark Sholander Jamie Clarke Bob Jordan Byron Dorgan George Sinner Quentin Burdick Corky Jones Peter Hoagland John Racek Bob Kerrey Joseph Keefe James Donchess Paul McEachern James Florio William Hughes Frank Pallone Christopher Smith Bernard Dwyer Matthew Rinaldo Robert Roe Robert Torricelli Don Payne Frank Guarini Frank Lautenberg Tom Udall Jeff Bingaman Jim Bilbray Jim Spoo Richard Bryan George Hochbrueckner Robert Mrazek Francis Goban Floyd Flake Gary Ackerman Thomas Manton Edolphus Towns Major Owens Jay O'Donovan Charles Rangel Ted Weiss Robert Garcia Nita Lowey Hamilton Fish Ben Gilman Michael McNulty E U'a'U”UUUUUUUUUUUDUUUUUUUUUUUFUUFUUUUUUUUUUUUUUUUUU 35333532333 E 141 SENATE, HOUSE AND GOVERNOR ENDORSED INCUMBENTS AND CHALLENGERS CANDIDATE Sherwood Boehlert David Martin Rosemary Pooler Frank Horton Louise Slaughter David Swarts Henry Nowak Daniel Patrick Moynihan Thomas Luken Chuck Stidham Tony Hall Tom Murray Gordon Roberts Jack Schira Marcy Kaptur John Buchanan Dennis Eckart Michael Gelpi Don Pease Tom Sawyer Mark Froehlich James Traficant Douglas Applegate Edward Feighan Mary Rose Oakar Louis Stokes Howard Metzenbaum Kurt Glassco Mike Synar Les AuCoin Larry Tuttle Ron Wyden Peter DeFazio Mike Kopetski Thomas Foglietta William Gray Robert Borski Joseph Kolter Gus Yatron David Landau Peter Kostmayer Joseph McDade Paul Kanjorski John Murtha Bernard Tomkin William Coyne C E UCUUNUUUUUUUUUUUUUDUUUUUUUUUHUUUUUUUUUUUUUNURIN C 142 SENATE, HOUSE AND GOVERNOR ENDORSED INCUMBENTS AND CHALLENGERS SANDIDAIE Ed Riebman Ernest Eric Guyll Doug Walgren Joseph Gaydos Austin Murphy Fernand St Germain Richard Licht Jim Leventis Butler Derrick Elizabeth Patterson John Spratt Robin Tallon Tim Johnson Dudley Taylor Marilyn Lloyd Jim Cooper Bob Clement Bart Gordon John Tanner Harold Ford Jim Sasser Jim Chapman Blake Cowden Ralph Hall John Bryant Pat Kendrick Diane Richards Bill Sarpalius Greg Laughlin Gerald McCathern Henry Gonzalez Elmer Wayne Walker Albert Bustamante Michael Andrews Jo Ann Reyes Lloyd Bentsen Gunn McKay Wayne Owens Ted Wilson Brian Moss James Ellenson Owen Pickett Jim Olin David Brickley Frederick Boucher Robert Weinberg E UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU 143 CO I T E O OLI I T SENATE, HOUSE AND GOVERNOR ENDORSED INCUMBENTS AND CHALLENGERS EEEEEEEE ANDIDATE Chuck Robb Paul Poirier Madeline Kunin Bill Gray Reese Lindquist Al Swift Jolene Unsoeld J. Richard Golob Thomas Foley Norman Dicks Jim McDermott Jim Kean Booth Gardner Mike Lowry Les Aspin Robert Kastenmeier Carl Krueger Gerald Kleczka Jim Moody Joseph Garrett David Obey Tom Hickey Herb Kohl Alan Mollohan Harley Staggers Robert Wise Nick Rahall Gaston Caperton Robert Byrd Brian Sharratt John Vinich E UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU LI ST OF REFERENCES LIST OF REFERENCES Abramson, Paul R. 1975. e e 'o a e ' Entitinn. Lexington, MA: Lexington Books. Abramson, Paul R., Aldrich, John H., and Rohde, David W. 1983. Cnange nnd Continuity in tne 1282 Eiggtipng. Washington, D.C.: CQ Press. Alexander, Herbert E. 1976. C ai o e . New York: The Free Press. Alger, D. E. 1991. "Schizophrenia: The Media and the Incumbent Reelection Issue." Epiitigni_2gi§ng§_nng Entitigg (June): 188-193. Alston, C. 1990. "A Political Money Tree Waits for Incumbents in Need." Congressionni anrteriy, Vol. 48 (June 30): 2023-2026. Alston, C. 1990. "Those Who Needed It Least Often Got Campaign Help." Congressional Qunrterly, Vol. 48, No. 52 (December): 4235-4237. American Federation of Labor and Congress of Industrial Organizations. 1987. Proceedingg of the AFL-CIQ Contention. Washington, D-C- American Federation of Labor and Congress of Industrial Organizations. 1988. Report on angresg, Washington, D.C. American Federation of Labor and Congress of Industrial Organizations. 1987. nepgrt on Congress, Washington, D.C. Anderson, John C. and Delaney, John T. 1990. "The Involvement of Canadian Unions in Political Activities: An Exploratory Analysis." 2pnrnni_gt LSDQI_ESSSS£SB: 11, 4: 353-379- Aronowitz, Stanley A. 1973. se i e . New York: McGraw-Hill. 144 145 Axelrod, Robert. 1967. "The Structure of Public Opinion on Policy." gublic Opinion Quarterly, 31: 51-60. Axelrod, Robert. 1972. "Where the Votes Come From: An Analysis of Electoral Coalitions, 1952-1968. P l 'cal Sci e v ew, 66: 11- 20. Banaian, K. and Luksetich, W. 1991. "Campaign Spending in Congressional Elections." flgntgrn_lngniry, Vol. 29 (January): 92-100. Barbash, Jack. 1947. "Unions, Government, and Politics." In u r' nd or Re at'o e iew, 1, 1 (October): 66-79. Barbash, Jack. 1965. "The Structure of Union Political Action," Labo Law Journal, 16, 7 (July): 491-497. Bartlett, Alton C. 1966. "How Rank and File Leaders View Union Political Action." Labor Law Journal, 17, 8: 483-494. . Berelson, Bernard R., Lazarsfeld, Paul F. and McPhee, William N. 1954. Votin : A Stud 0 O i '0 W Toronto ° The University of Toronto Press. Berry, J. M. Lppnying for the Eggpie. 1227. Princeton, NJ: Princeton University Press. Blume, Norman. 1970. "The Impact of a Local Union on its Membership in a Local Election." The Wentern Egiiticai Quarterly, 23, 1 (March): 138-150. Blume, Norman. 1973. "Control and Satisfaction and Their Relation to Rank-and-File Support for Political Action." Wegtgrn Eoiitical Quarterly, 26, 1 (March): 51-63. Bok, Derek C. and Dunlop, John T. 1970. anor and tng Angrignn_annnnity. New York: Simon and Schuster. Borrelli, Stephen, Lockerhie, Brad and Niemi, Richard. 1987. "Why the Democrat-Republican Partisanship Gap Varies From Poll to Poll." Public Opinion Quarterly, 51, 1 (Spring): 115-119. Bradley, Phillip D. 1959. The Public stake in Union Power. Charlottesville: University of Virginia Press. 146 Brown, Edwin L. and Catlett, Judi L. 1983. "Targeting the Political Image: The Case of Amendment 4." Lah9r_§tndies_ionrnal 8. 1 (Spring): 34-47- Bruce, Peter G. 1989. "Political Parties and Labor Legislation in Canada and the U. S. " lngnntrini Bgintipnfi, 28, 2 (Spring): 115-141. "Bush Enters the Election Campaign in a Politically Commanding Position." Wnl; §tre§t Journni, 73, 15 (Friday, November 1): 1. Campbell, Angus, Converse, Phillip E., Miller, Warren E. and Stokes, Donald E. 1960. Inn_nngrignn_ypt§r. University of Michigan: Ann Arbor, Survey Research Center. Chamberlain, John. 1984. "Unions and Politics: A Journalist's View." gonrnnl 0t Lnnor gnnnargn, 5, 3 (Summer): 309-313. "Changing Situation of Workers and Their Unions." Report by the AFL-CIO Committee on the Evaluation of Work, February 1985. Chappell, Henry W. 1982. "Campaign Contributions and Congressional Voting: A Simultaneous Probit-Tobit Model." Tne Envigw of Economics 2nd 2tatisticn, 64 (February): 77-83. Clarke, P. and Evans, S. 1983. Cov 'n ai . Stanford, CA: Stanford University Press. Coleman, Vernon. 1988. "Labor Power and Social Equality: Union Politics in a Changing Economy." Enlitignl Stien2e_Qnarterlx 103. 4: 687- 705- a t . 1986. Washington, D.C.: Congressional Quarterly, Inc. C 'o e . 1987. Washington, D.C.: Congressional Quarterly, Inc. Congressional_9uarterlx- 1988- Washington. D-C-: Congressional Quarterly, Inc. Cook, R. 1990. "Tracking Primaries: Dissension Index." Congr_ssional_ouarterlx Vol 48-. NO- 10 (March): 798. Cook, R. 1991. "Incumbents' National Status Breeds Local Distrust." Congressional Qnartgriy, (February 23): 483-487. 147 Danielson, Michael N. and Murphy, Walter F. 1987. Angrigan nemocrncy. New York: Holmes and Meier Publishers, Inc. Delaney, John T., Fiorito, Jack and Masters, Marick F. 1988. "The Effects of Union Organizational and Environmental Characteristics on Union Political Action." Anericnn Journnl of Pgliticnl Sgiengn, 32, 3 (August): 616-642. Delaney, John and Masters, Marick. 1991. "Unions and Political Action," in The St te f Un'o , George Strauss, Daniel Gallagher and Jack Fiorito, eds: Madison, Wisconsin: Industrial Relations Research Association. Donnelly, Harrison H. 1978. "Campaign Spending: New Record Set in 1976." on es a r W finpgrt, 36 (January): 119. Dubofsky, M. 1987. "The Extension of Solidarity Conflicts with the Spirit of Individualism." Mgntniy_angr 22yin_: 36-37. Durden, Gary C. and Silberman, Jonathan J. 1976. "Determining Legislative Preferences for the Minimum Wage: An Economic Approach-" Journal_ot_2211tital Eggnpny, 84 (April): 317-329. Dwyer, P. 1991. "Incumbent Could be a Four-Letter Word Next Year." 2n§iness Week (October 21): 45. Dye, Thomas R. 1973. olitic i S d Co ti . New Jersey: Prentice Hall, Inc. Eccles, Mary E. 1976. "Labor Fares Well in House." ' 34 (November): 3136-3144. Epstein, Edwin M. 1976. "Labor and Federal Elections: The New Legal Framework." Industrial__elat121§. 15 3 (October): 257-274. Erikson, Robert S. 1989. "Why the Democrats Lose Presidential Elections: Toward a Theory of Optimal Loss," _2l1t1tal_§tient__and_£211tits 22. 1 (March): 30-34. Feinberg, Stephen. 1985. e s' o - . Cambridge, MA: MIT Press. 148 Feld, Scott L. and Grofman, Bernard. 1988. "Ideological Consistency as a Collective Phenomenon." AESEiSSS Boliticai Sgienge Review, 82, 3 (September): 773- 788. Feree, M. and Miller, F. 1985. "Mobilization and Meaning: Toward an Integration of Social Psychological and Resource Perspectives on Social Movements." ggcinl Inguiry, 55 (1): 38-61. Fields, Mitchell W., Masters, Marick F., and Thacker, James W. 1987. "Union Commitment and Membership Support for Political Action: An Exploratory Analysis." gonrnal of anor Rnggnrgn, 8, 2 (Spring): 143-157. Fiorito, Jack. 1987. "Political Instrumentality Perceptions and Desires for Union Representation." ggurnnl of Labor Rnnenrch, 8, 3 (Summer): 271-289. Flanigan, William H. and Zingale, Nancy H. 1983. i B or o e o . Boston: Allyn and Bacon, Inc. Form, William H. 1973. "Job vs. Political Unionism: A Cross-national Comparison." ingnntrini_2gintipn§, 12, 2: 224-238. Form, William H. 1982. "Self-Employed Manual Workers: Petty Bourgeois or Working Class?" 2gpini_£grgn§, 60, 4 (June): 1050-1069. Form, William H. 1983. "Sociological Research and the American Working Class." The.ggtiolog1tal_enarterl¥. 24, 2 (Spring): 163-184. Fossum, John A. 1992. Labor Relations: Dnygigpngntr Structure, Progess. Boston, MA: Irwin. Foster, James C. 1975. Zhe Union Politic: Ihe CIO Poiitical Action Committee. Missouri: University of Missouri Press. Fowler, Linda L. 1982. "How Interest Groups Select Issues for Rating Voting Records of Members of the U.S. Congress." Legisintivg Studie§ Quarterly, 7, 3 (August): 401-413. Freeman, Richard B. and Medoff, James L. 1984. wnat up Unions 20? New York: Basic Books, Inc. Gibbs, N. 1990. "Keep the Bums In." Timg, (November 19): 32-42. 149 Greenstone, David J. 1969. Lnnor in Amerigan Enlitigs. New York: Random House. Greenstone, J. David. 1982. Enniic anugs and Privnte Epwgr in flericnn Politicg. Chicago: The University of Chicago Press. Gulick, Charles and Bers, Melvin. 1953. "Insight and Illusion in Perlman's Theory of the Labor Movement," in industrial and Labor Relations View -- as condensed in Ihgorign of tne anor Movemen , Simeon Larson and Bruce Nissen, eds. Detroit: Wayne State University Press. Halbway, Harry. 1979. "Interest Groups in the Post- partisan Era: The Political Machine of the AFL-CIO." Politicn; Science Qunrterly, 94, 1 (Spring): 117-133. Harwood, Edwin. 1981. "Comment." Jou a of r'R a c , 2, 2 (Fall): 299-305. Heldman, Dan C. and Knight, Deborah L. 1980. Cniong and Lopbying: The Representation Eunction. Arlington, VA: Foundation for the Advancement of Public Trust, Inc. Herndon, James F. 1982. "Access, Record, and Competition as Influences on Interest Group Contributions to Congressional Campaigns." Journai of Politicg, 44, 4 (November): 996-1019. Himmelstein, Jerome L. and McRae, James A. 1984. "Social Conservatism, New Republicans, and the 1980 Election." gnpiic Opinion Qunrteriy, 48, 3 (Fall): 592-605. Holly, W. and Jennings, K. 1980. The anor Relntions Erpggnn. Illinois: The Drydon Press. Hudsen, Ruth A. and Rosen, Hjalmar. 1953. "Union Political Action." Ingustrial and Labor gelations 2eview, 7, 1 (October): 404-418. Hudson, Ruth A. and Rosen, Hjalmar. 1954. "Union Political Action: The Member Speaks." industriai and Lnnor Relations Review, 7, 3 (April): 404-418. Idelson, H. 1990. "Vulnerable Incumbents Prevail, Remain on Shaky Ground." Congressional Quarterly, Vol. 48 (September 22): 3036. 150 Juravich, Tom. 1986. "Beyond the Myths of '84: Union Members and the Presidential Election." anpr Studies_lournal 11. 2 (Fall):135-148- Juravich, Tom and Shergold, Peter R. 1988. "The Impact of Unions on the Voting Behavior of Their Members." Ingnstrini and Labgr gelations geyigw, 41, 3 (April): 374-385. Karr, James B. and Rubin, Paul H. 1981. "The Impact of Labor Unions on the Passage of Economic Legislation." Jsurnal_of_LahQr_Eesearth. 2. 1 (Spring): 133-145- Kerr, Clark, Dunlop, John, Harbison, Frederick and Meyers, Charles. 1960. Indnstrialism nng Industrini unn, Cambridge, Mass.: Harvard University Press, as condensed in The2r1es_21_the_1ahgr_onenent. Simeon Larson and Bruce Nissen, eds. Detroit: Wayne State University Press: 315-336. Kessler-Harris, A. 1987. "Trade Unions Mirror Society in Conflict Between Collectivism and Individualism." nontniy Lgtpr Review, (August): 32-35. Key, V. O. (1958). ' 'c Crgnpn. New York: Thos Y. Crowell Company. Knoke, David and Burke, Peter. 1980. ngiinnnr_nggni§. Beverly Hills, CA: Sage Publications. Kochan, Thomas and Wever, Kirsten. 1991. "American Unions and the Future of Worker Representation," in Inc Stntn gr_tnn_2nign§, George Strauss, Daniel Gallagher and Jack Fiorito, eds: Madison, Wisconsin: Industrial Relations Research Association. Kornhauser, Arthur, Sheppard, Harold and Mayer, Albert J. 1956. fingn_anpr_ygt_§. New Hyde Park, NJ: University Books. "Labor is Irked with Democrats." 1991. annn_flprig; 321212, (Monday, September 2): 9. Ladd, Everett C. 1985. "On Mandates, Realignments, and the 1984 Presidential Election." Epiitigni_§gigngn Quarterly 100, 1 (Spring): 1- 25- Ladd, Everett Carll. 1987. "Party Reform and Public Interest." in Elections_1ner12an_§txle, J- Reichley. ed. Washington, D.C.: The Brookings Institution: 222-239. 151 Ladd, Everett C. 1989. "The 1988 Elections: Continuation of the Post-New Deal System." S nc Quarterly. 104. 1: 1-18. Larson, Simeon and Nissen, Bruce. 1987. Iheories pf tne vem . Detroit: Wayne State University Press. Levitan, Sar and Cooper, Martha. 1984. SSSiSSSS_LSSSiSSo Baltimore: Johns Hopkins University Press. Levitan, Teresa and Miller, Warren. 1979. "Ideological Interpretations of Presidential Elections." The American Politicni Sgiencn Egyiey, 73, 3 (September): 751-771. Lipset, Seymour M. 1981. Eg]j;j9§] Men° The S;:. J E 91.2911L12S. Baltimore: The Johns Hopkins University Press. Lodge, M. Hon, McGraw, Kathleen and Stroh, Patrick. 1989. "An Impression-Driven Model of Candidate Evaluation." Angriggn Political Science Review, 13, 2 (June): 399-419. Love, Thomas M. and Sulznar, George T. 1972. "Political Implications of Public Employee Bargaining." ingustrini Relntions, 11, 1: 18-33. Luttbeg, Norman R. and Gant, Michael M. 1985. "The Failure of Liberal/Conservative Ideology as a Cognitive Structure." Pnniig Qpinion Qunrteriy, 49, 1 (Spring): 80-93. Lyons, Schly R. 1969. "Labor in City Politics: The Case of the Toledo United Auto Workers." 2pgini_§gigngg Qnartgrly, 49 (March): 816-828. Marsel, Louis and Cooper, Joseph. 1977. The Inpngt pf tne Electorni Ergcess. London: SAGE Publications. Masters, Marick F. 1985. "Federal-Employee Unions and Political Action." Indnstr1a1_and_LahQr_Relat1on§ Bgyiew, 38, 4 (July): 612-628. Masters, Marick F. and Delaney, John T. 1984. "Inter- union Variation in Congressional Campaign Support," Ingnstrini gelations, 23, 3 (Fall): 410-416. Masters, Marick F. and Delaney, John T. 1986. "Labor in American Politics: An Appraisal of a Neglected Research T0piC-" anthlished_m222§tript- 152 Masters, Marick F. and Delaney, John T. 1987. "Contemporary Labor Political Investments and Performance-" LSSQI_§LSS1SS_QSSISS17 11. 3 (Winter): 220-237. Masters, Marick F. and Delaney, John. T. 1987. "Union Legislative Records During President Reagan's First Term." Jsurnal_of_Lahor_Researth. 8. 1 (Winter): 1-18. Masters, Marick and Zardkoohi, Asghar. 1986. "The Determinants of Labor PAC Allocations to Legislators." ingnstrinl Relntigng, 25, 3 (Fall): 328-338. McAdams, Alan. 1964. Bower nnd Eoliticn in 2229: ngifiintign. New York: Columbia University Press. Michels, Robert. 1915. "Democracy and the Iron Law of Oligarchy." in Pgiiticai Parties: A §ociologigni d o ' ' e ' ngnggrngy. New York: The Free Press (1962): 342- 356. Michels, Robert. 1915. "The Conservative Basic of Organization." in Epiitical garties; A ociolo Stu of t e of Modern Democrngy. New York: The Free Press (1962): 333-341. Milton, David. 1982. e o ' 'c U. . New York: Monthly Review Press. Moe, T. M. 1980. e O aniz ' f s . Chicago: University of Chicago Press. Moore, David M. 1987. "Political Campaigns and the Knowledge-Gap Hypothesis." Euhlit_opin1gn_guarterlr. 51, 2 (Summer): 186- 200. Morgan, S. Phillip and Teachman, J. D. 1988. "Logistic Regression: Description, Examples, and Comparisons." I2nrnal_2f_Earrisge_and_lhe_tam111 50 (Nov ): 929- 936. Nesbitt, Frederick H. 1978. "Labor Endorsement of a Republican Senator: A Case Study." In Charles M. Rehmus, Doris B. McLaughlin and Frederick H. Nesbitt, eds., Lnnor nng Anerignn Eoiitics, Ann Arbor, Michigan: University of Michigan Press: 409-420. Niemi, Richard G. and Weisberg, Herbert F. 1976. e can Vo e ' . San Francisco: W. H. Freeman and Company. 153 Norpoth, Helmut. 1987. "Under Way and Here to Stay: Party Realignment in the 19808?" W1! 51: 376- 391. Olson, Mancur J. The L0 'c o 0 le ive c ' n. 1965. Cambridge, MA: Harvard University Press. Pae, Sung M. 1986. est’ emo rat' or'e 'n o . Lanlam, MD: University Press of America. Pahlmann, Marcus D. and Crisci, George S. 1982-1983. "Support for Organized Labor in the House of Representatives: the 89th and 95th Congresses. Parent, T. Wayne, Jillson, Calvin C., and Weber, Ronald E. 1987. "Voting Outcomes in the 1984 Democratic Party Primaries and Caucuses." Ame 'c n oliti c' Review, 81, 1 (March): 67-84. Parker, G. 1989. "The Role of Constituent Trust in Congressional Elections." j2nniig_Qpinipn_Qn_rtgriy, Vol. 53: 175-196. Patton, David B. and Marrone, John J. 1984. "The Impact of Labor Endorsements, Union Members and the 1980 Presidential Vote." Lflor 2tndie§ Jpnrnai, 9, 1 (Spring): 3-18. Patton, David.B., Marrone, John.J. and.Hendman, Hugh.D. 1986. "Unions and Politics: 1984 and Beyond." d i a Lnnpr gelationn gesearch Association Procnedingn: 490- 494. Perlman, Selig. 1928. A heo of t o ovem New York: MacMillan, in T e Simeon Larson and Bruce Nissen, eds. Detroit: Wayne State University Press. Petrocik, John R. 1981. garty Cpalitigng nnd the Qggline pf tne New Deni Party Systemn. Chicago: University of Chicago Press. Petrocik, John R. 1981. a C0 ' ' ° a ' 3 nd c ° e of t New Deal Pa S ems. Chicago: University of Chicago Press. Pizzorno, A. 1981. "Interests and Parties in Pluralism" in Qrganizing Interests in Western Europe: Eiurnlisn, Co orat'sm t Transformat'o o it'c , Suzanne Berger, ed. Cambridge: Cambridge University Press: 247-284. 154 Polsby, Nelson. 1983. T e Co e ces o t e o . Oxford: Oxford University Press. Pomper, Gerald M. 1968. Electionn In nnericn. New York: Dodd, Mead and Company. Ra, Jong O. 1978. Labor at the Polis. Amherst: University of Massachusetts Press. Radosh, R. 1969. erica or U.S. o e' P ' . New York: Random House. Rappaport, Ronald.B., Stone, Walter J. and Abramowitz, Alan 1. "Sex and the Caucus Participant: The Gender Gap and Presidential Nominations. " WM grinngg, 34, 3 (August): 725-740. Rehmus, Charles M. 1986. "The Future of Industrial Relations in the United States." Labour nng Sggiety, 11, 2 (May): 149-158. Rehmus, Charles M. and McLaughlin, Doris B. 1967. Lnnpr nng_nngriggn_£giitig§. Ann Arbor: The University of Michigan Press. Rehmus, Charles M., McLaughlin, Doris B., and Nesbitt, Frederick H. 1978. W. Revised Edition. Ann Arbor: The University of Michigan Press. Reichley, A. James. 1987. Elections Anerignn §ty1§. Washington, D.C.: The Brookings Institution. RePass, David E. 1971. "Issue Salience and Party Choice." IDS Angrignn Political Science Review, 65 (June): 389- 400. Roback, Thomas H. 1975. "Amateurs and Professionals: Delegates to the 1972 Republican National Convention." W, 37: 436-468- Robinson, John P. and Fleishman, John A. 1984. "Ideological Trends in American Public Opinion." ARESS1S_Q£_SAS W. 472 (March): 50- 60. Robinson, John P. and Fleishman, John A. 1988. "Ideological Identification: Trends and Interpretations of the Liberal-Conservative Balance . " MW, 52: 134-145. 155 Rogin, Michael. 1962. "Volunteerism: The Political Function of an Antipolitical Doctrine." Ingustrial ang anor Relations Review, 15 (July): 521-535. Rules Governing A-C State Central Bodies (1973). Publication of the AFL-CIO, Washington, D.C.: 21. Sabato, Larry J. 1984. PAC Power. New York: W. W. Norton and Company. Salisbury, Robert H. 1969. "An Exchange Theory of Interest Groups." Midwest Journal of Eolitical 2cignce, 13: 1-32. Salisbury, Robert H. 1975. "Interest Groups." In N. Polsby and F. Greenstein, eds. Hnndbook of Political Science. Reading, MA: Addison-Wesley, 1975. Salisbury, Robert H. 1984. "Interest Representation: The Dominance of Institutions." The Anerican Politigal Scienge Review, 78, 1 (March): 64-76. Salisbury, Robert H., Heinz, John P., Laumann, Edward O., and Nelson, Robert L. 1987. "Who Works With Whom? Interest Group Alliances and Opposition." AESEiSSE Political Science Review, 81, 4 (December): 1217- 1234. Saltzman, Gregory M. 1987. "Congressional Voting on Labor Issues." Industrial and Labor Relations 2eview, 40, 2 (January): 163-179. Schlichting, Kurt C. 1989. "Democratic Incumbents and the 1984 Presidential Election." Public Cpinign Quarterly, 53, 1: 83-97. Schmidt, Emerson P. 1973. Union Power and the Public Interest. Los Angeles: Nash Publishing. Schneider, William. 1984. "Incumbency Saved the Democrats This Time But What About The Next?" Lg§_nnggln§_1inn§, November 11, part V: 1 and 3. Seidman, Joel, London, Jack and Karsh, Bernard. 1951-1952. "Political Consciousness in a Local Union." gublic Opinign Quarterly, 15 (Winter): 692-702. Shafer, Byron. 1988. Bifurcated Eolitigs. Cambridge: Harvard University Press. 156 Sheppard, Harold L. and Masters, Nicholas A. 1959. "The Political Attitudes and Preferences of Union Members." Tne Amerignn Pgli ti cnl Science Beview, 53, 2 (June): 437- 447. Shribman, David (1991, November 13). "Democratic Presidential Assail the Establishment but Not Each Other." Tng_flnll Strget Jonrnnl, p. A18. Shribman, David (1991, November 4). "New Breed of Activist Travels U.S. Pitching Issues, Not Candidates." Tng Wnll street ggnrnal, p. A1. Soule, John and Clarke, J. W. 1970. "Amateurs and Professionals: A Study of Delegates to the 1968 Democratic National Convention." American Politigal Science Rgview, 77: 945-956. Stone, Walter J. and Abramowitz, Alan I. 1983. "Winning May Not Be Everything But It's More Than We Thought." Amnrican Politicnl Science gnvinw, 77: 945-956. Strong, Donald S. (1967). Issne Voting nng gnrty , University of Alabama: University of Alabama Press. Sullivan, Denis G., Pressman, Jeffrey L., Arterton, Christopher F., Nakamura, Robert T. and Weinberg, Martha W. 1977. "Candidates, Caucuses, and Issues: The Democratic Convention, 1976," in Tne lmpngt or tne Electornl Erocess, L. Marsel and J. Cooper, eds. London: SAGE Publications: 81-132. Sullivan, James M. 1984. "Unions in Politics: A.Madisonian Faction in Our Federal System." Qpnrnnl_pr_tnnpr ggngnrgn, 5, 3 (Summer): 275-285. Thomas, Patricia. 1986. "Pay Equity: Educating for Union Action." Labgr §tugi§§ Jonrnnl, 10, 3 (Winter): 278- 289. Touraine A., Dubet F., Wievioika.M., and Strzelecki J. 1983. 't e al sis of a Soc a ov P l 9 0- . Cambridge: Cambridge University Press. Walker, Malcolm J. and Lawler, John J. 1979. "Dual Unions and Political Processes in Organizations." lngnntrinl Belntignfi, 18, 1 (Winter): 32-43. Wattenberg, Martin P. 1987. "The Hollow Realignment: Partisan Change in a Candidate-Centered Era." gnplig Qpinion_Quarterlx 51 1 (Spring): 58- 74- 157 Weaver, W. 1978. "Special Interests Spend $60 Million." new XQEEJILESS, 128 (November 7): A27. Weisberg, Herbert F. 1987. "The Demographics of a New'Voting Gap." gublic Opinion Qunrterly, 51, 3 (Fall): 335-343. Welch, W. P. 1980. "The Allocation of Political Monies: Economic Interest Groups." Enplic Cnoice, 35: 99-120. Wilhite, Allen and Theilmann, John. 1986. "Unions, Corporations, and Political Campaign Contributions: The 1982 House Elections." Jonrnal.of_Lahgr_Researth. 7. 2 (Spring): 175-185. Wilson, Graham K. 1979. U 'o ' 'c t' Eolitics. New York: St. Martins Press. Wolfe, Arthur C. 1969. "Trends in Labor Union Voting Behavior, 1948-1968." Indngtrinl gnlationn, 9, 1: 1-10. Wright, Samuel. 1982. "Clipping the Political Wings of Labor Unions: An Examination of Existing Law and Proposals for Change-" Har2ard_I2trna1_2f_Las_and_znhl1t_2911tx_1. citing Federal Election Commission press release date March 29, 1981. Zurcher, L. and Snow, D. 1981. "Collective Behavior: Social Movements," ' s o : oc'o Egrnpggtiynn, M. Rosenberg and R. Turner, eds. New York: Basic Books. MICHIGAN STATE UNIV. LIBRARIES ill1|millI“||||\llillilWl‘lllNlllllHliliHlilHlIlWl 31293008977591