TURNOVER OF LOCAL UNION OFFICERS AN ORGAN EZATIONAL ANALYSIS THES IS fer the degree of Ph.D Michigan State University AMNON CASPI 1970 Lymeqnv Ll TH 65‘? Michigan State University 1? 13"“ ‘- This is to certify that the thesis entitled TURNOVER OF LOCAL UNION OFFICERS: AN ORGANIZATIONAL ANALYSIS presented by Amnon Caspi has been accepted towards fulfillment of the requirements for Ph.D. Social Science degree in 5%va H WWW Major professor Date February 26, 1970 local 1 Michel: nation ideutg ‘36:» 0113: of 0: San ¢ the 4 the 1959 int; Sh‘l den ABSTRACT TURNOVER OF LOCAL UNION OFFICERS AN ORGANIZATIONAL ANALYSIS By Annon Caspi This study deals with leadership stability on the local union level. 'Its conceptual framework starts with Michels' theory which deals with the power structure within national political parties. The goal of the study is to identify some of the organizational structural variables and explore their relationship to the leadership turnover on a local union level. The sample is a random stratified one, consisting of one hundred United Automobile‘Workers locals in Michi- gan during 1959-1967, inclusive. The data sources include the annual reportsof the locals, filed in accordance with the "Labor Management Reporting and Disclosure Act of 1959*, Governmental and U.A.W. publications, and field interviews. The dependent variables are local union leader- ship turnover measured by the number of officers (Presi- dents, Vice Presidents, Financial and Recording Secre- taries) who were elected in each local and in all of them. The independent variables are eighteen structural ch CG 8i Amnon.Caspi characteristics of the local. These are grouped into four concepts: Formalisation, Complexity, Attractiveness, and External Elements. * The relationship between the structural variables and turnover are tested with various statistical methods: correlations, multiple correlation, chi-square, and one way analysis of variance. The results indicate that the variables of formalization (age and size of the local), complexity (number of units, payments to local employees, and total resources), and attractiveness (compensation to local officers) correlate negatively with the turnover rates. The growth rates (variables of formalization) cor- relate positively with turnover. The correlations and their significance levels vary considerably among the dif- ferent variables and officers. Generally, the results suggest that.Michels' theory is applicable to the local level, and the larger, older, and more segmented the local, the lower its turnover rate. The external elements (the company, the community, and the time dimensions) are unrelated to the rates of turnover. The multiple correlation calculations show that the total fourteen variables (those continuous variables that can be used in this calculation) account for 7 percent to 3k porcent of the variation, depending upon the office involved. Formalisation accounts for A percent to 17 per- cent; complexity accounts for 2 percent to 7 percent; and attractiveness for 1 percent to a percent. eXp tut it the six tiw so At St. to we of PC l‘E ti de 01 El Amnon Caspi This percentage seems to be somewhat lower than expected, and two reasons for this emerge from the struc- ture and nature of the study. 1. Although the study makes it clear that formaliatic or structural approach can cone tribute to the understanding of turnover in the local union, the results suggest that in order to achieve a comprehenr sive view and a higher degree of statistical and substan- tive explanation, a combination of both structural and socio-psychological approach is required. 2. The United Auto‘hcrkers International Union, a highly organised and structured union, uses various organisational mechanisms to closely control its locals, in order to assure their 'well defined organisational behavior patterns. Because of the unifying effect of these mechanisms, the variables' potential influence on leadership turnover seems to be reduced. The study also includes an analysis of the rela- tionships among the different variables themselves, a detailed analysis of the special elections, and the analy- sis of the inner mobility of officers. These analyses, together with the major findings relevant to turnover sug- gest the impossibility of treating all four officers inclue dad in the study in the same way. The officers differ from one another in their turnover rates which correlate differ- ently with the various structural variables of the study. These unplanned findings suggest that a refinement of the concept of leadership is required for a better under- standing of the turnover phenomenon. TURNOVER OF LOCAL UNION OFFICERS AN ORGANIZATIONAL ANALYSIS 8! Amnon Caspi A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY College of Social Science 1970 x 5.1; 7g; 7. Ho Marcus, 1 and insi- tee each Kruger, 801' Win: the Sch, State u! “Me 0 the m Auto Wc d‘l‘ity tram” withom £527.47 7~ /~7° ACKNOWLEDGMENTS I wish to extend my thanks to Professor Philip Narcus, my major advisor, for his advice, encouragement, and insight and to all the members of my Guidance Commi- tee each of whom.helped in his area: Professor Daniel Kruger, the Chairman, Professor Albert Blwm, and Profes- sor Winston Oberg. l gratefully acknowledge the financial support of the School of Labor and Industrial Relations at Michigan State University, as well as the good advice and assis- tance of Professor Jack Stieber, the Director. Special thanks are due to Mrs. Martha Soltow and the library staff, to Hr. Douglas Fraser from the United Auto workers Headquarters, to the staff members of Soli- darity House, and to the Labor Management Service Adminis- tration in Detroit. Last but not least, my thanks to Nina, my wife. ‘Without her this study would not have been'written.' 11 LIST 0? CHAPTER CHAPTE] GRAPE CRAP: CRAP TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . CHAPTER I. CHAPTER II. THE PROBLEM, THE PURPOSE AND THE OUTLINES OF THE STUDY . . . The Problesxend Purpose . . Outlines of the Study . . THEORIES, CONCEPTS, VARIABLES AND " HYPOTHESES O O O O O O O O O Michell' 1.13.." e e e e e e The Conceptual Framework . The Concepts, the variables HYPOChOIOSeeee eee Formalisation . . . Complexity . . . . Attractiveness . . External Elements . CHAPTER III.DATA AND METHODOLOGY . . . CHAPTER IV. CHAPTER V. Sources of the Data . . Quality of the Data . . “OthOdclogy e e e e e e m STRUCTURAL VARIABLES or Sise of locals and Rates of Age of the Locals and Other The Type of the Local . . . The Community Factor . . . The Company . . . . a the .....8. THE LOCALS. Growth . variables 0 O O O O 0.... Phyments of the Locals to Employees and ’Offic.r.eeeeeeee PATTERNS OF ELECTIONS, TURNOVER AND INNER MOBILITY OF LOCAL OFFICERS . . . Comparison and Discussion of Turnover HWWChtm070r7e e e e e e e e Appelbaum's findings - comparison . Local and International Turnover . Why the differences in turnover? . 111. 12 12 15 18 18 22 25 29 31 35 37 37 #2 #3 #5 £7 CHAPTER 1 “AFTER 813L100; APPENDI APPE ND] APPENDj Special Election . . . . . Patterns of Re-Elections . Patterns of Inner Mobility amneeeeeeee CHAPTER VI. STRUCTURAL VARIABLES AND Formalisation . . Complexity . . . Attractiveness . External Elements 0... CHAPTER VII.SUMMARI AND CONCLUSIONS TURNOVER O... O... O... O... 0... Suggestions for further research . BIBLIWRAPHY O O O O O O O O O O O O O O O 0 APPENDIX A. POPULATION AND THE SAMPLE . . . . APPENDIX a. pm comer-Jun sonnets AND THE VARIABLES......... APPENDIX C. INTERVIEW GUIDELINES 1's 0 0... 10A 115 120 126 Table Ill-1, IV-2, IV-3. IV-h e “-5. IV-6 IV-7. V-l V-z M O M V‘s v.6 Table IV-l. I'-2 e IVrB. I'.‘e "'5 e “.6. N'Ve V-l e "2 a V-3 e v.‘ . V65. V-6. LIST OF TABLES Distribution of sample locale by 1‘00” .nd ”ab.flh1p e e e o e e e e The correlates of age with other "r1.b1.' Of th. ”tad, e e e e e e e Amalgamated and single-unit locals in the sample and in Michigan . . . . . Relationships between the type of the local and fhcilitiee, newspapers and payments to employees and officers. . Correlations between community index .“ 9thCr "ri‘bICI a e e a e e e e a "Big Three" locals as compared to Oth0"eeeeeeeeeeeeeee Distribution of locals according to their yearly payments to officers and employees............. Distribution of the hundred locals by the Bulb” 0: OfILCCr. a s e e e e e thIOftmovgreeeeee eeee Comparison of turnover rates--Applebaum's and current .tUdy e e e e e as e a ‘Means of officers by the type of the office and the amount of personal contact with rank and file members . . Distribution of special elections by thCLPfflunHCYeeoeeeeeeee Preguency of special elections by the e fLCCeeeeeeeeeeeeeeee Page 38 #2 £3 #5 A6 A7 52 5k 56 62 .N V-7. What happened to those officers who were elected in special election, by office. . 6A V-8. The results of the vice president auto- matic movement to the presidency . . . . 65 V59. Inner mobility patterns . . . . . . . . . 67 V-lO. Analysis of presidents' answers about the relative importance of the four offices. 68 V-ll. Rearrangement of mobility patterns to , examine upward and downward mobility . . 69 V-lz. nobility between groups of officers . . . 70 VI-l. Correlations of turnover and formalisation V'TLCbICC a e e e e e e e e e e e e e e e 71 VI-Z. Overall statistics-~turnever and the fourteen structural variables . . . . . . 79 YI-B. MMlti le correlation of formalization var ables and turnover . . . . . . . . . 80 VI-A. Correlations of turnover and complexity "TLCbICQ e e e e e e e e e e e e e e e 85 VI-5. Average number of officers and type of 100.1 a e e e e e e e e e o e e e e e e 86 VI-6. Average number of officers and the exis- tODc. 0f fCCLILtLCC e e e e e e e e e e e 87 VI-7. Average number of officers and the exis- tent. of . H.”'p.p°re e e e e o e e e e e 87 VI-S. .Multiple correlation of complexity '.r1.b1.. .nd turnflVOT. e e e e e e e e e 89 VI-9. Turnover and payments of officers . . . . . 92 VI-lo. Average number of officers by the company . 9A VI-ll. Correlations of turnover and community indexeeeeeeeeeeeeeeeeeeg‘ VI-12. Number of new elected officers by the yar010000000000000000095 V1913. Number of new'elected officers by the 133’s Ila e e e e e e e e a e e e e e e e 95 vi . D . u‘lhe D e . . . k 1 I 0 O E U Q Q a e ‘ e e ‘0'20 ( I I I r, 1 . I ' C a Q ‘ ‘ . . . . O ' ' x g g . . g ‘ . . a A O ‘ \ , . 9 I e .' I § 9 ‘ a s e v y . \ . g e e s 9 e o e e e s m ' . . 0 e e ‘9 9 e G ‘ Q 9 '~ 0 v . e ‘ . ‘ ' 2 Q Q h e s e ~ e i e e e e ' s ‘1 . D O O O I C e e e O ’ . l q - s a _ t ‘ 6 t i ' V I I ‘ § 5 C O F 3 D ‘ ' N ‘ ' \ ‘ ‘ VI-lA. Distribution of changes of officers as occurred in the same year of the new labor contract ora year later . . . . . . . . .97 A.-l. Population and sample . . . . . . . . . . ..118 A.-2. Comparison of the 8 interviewed locals . . .119 vii CHAPTER I E P O LEM ' OUTLINES OF THE §IUQI TEE PROBLEM.AND PURPOSE Michels' theory known as'the "Iron Law of Oligarchy",1 originated as an attempt to explain the internal processes within democratic organizations, especially those processes which influence the leadership and distribution of power. The subject of Michels' observations were political parties in Europe, but his theory was extended to include analysis of other types of organizations. These findings have be- come an important part of organizational theory. various aspects of thhels' theory were empirically tested and, with few'exceptions, were found correct. These empirical tests were conducted on the national level of political parties and the international level of labor unions. The theory, however, has not been tested on the local levels of either of these organizations. Michels' theory is a fairly large body of assumptions and hypotheses which can hardly be empirically tested in a single research project. The one aspect we are to examine inois, . 1 in this study is the phenomenon of leadership turnover in the local union. This aspect of turnover is only implied in Michels' presentation, as he does not use the term or a direct equivalent of it. However, the concept of "sta- bility of leadership" can be operationaliaed as lower turnover. The local union is used as the subject of this study for two reasons. (1) The individual's relation- ship'with an organisation is generally at a local level, rather than an international; and, therefore, it is impor- tant to study this aspect in order to test whether Michels' theory can be applied to the local union. (2) Because the international organisation is composed of and influ- enced by local unions, a better understanding of the local union will facilitate the understanding of the total or- ganisation. There is no doubt about the importance of the local union, even though some of its functions have been removed to the International. Brooks emphasises the importance of our study: 1‘ The sources of vitality'within the American Labor movement are to be found in the local unions and local organisations. The bulk of our present able national leadership - of which we have a great deal - sprang out of active locals and the only real danger to the LaborMovement is that we shall dry up these sources of leadership and inspiration.1 ”—“__ 1 G.W. Brooks, The Source of Vita it of t e Amer Labor movement, (CorneII University,I§%Ii, p. 3. In spite of the declared importance of the local unions, very little has been published about them or their leadership. on. available documentation in of a case study nature, a comparison and analysis of relatively few locals, and above all, of descriptive nature. Furthermore, these available studies, with one exception, do notdeal direct- ly and solely with the phenomenon of turnover of local leadership. The one exception is of descriptive value only without any attempt to study functional relations.1 0n the other hand, although sophisticated analysis of the local union israre, and therefore it is almost impossible to substantiate statements, the union govern- ment's literature includes many statements about the dif- ferences of rates of turnover between International and local unions. None of these statements is supported by empirical and/or comparative study, although sometimes a case study or a long acquaintance with the Labor Movement is the source of insight upon which the statement is based. One example of these attempted explanations of varying turnover rates is the analysis of Bloom and Northrup, who argue that the reasons for higher rates of turnover in locals are: l. Local.members are willing to change leadership. 2. Local officers are elected for a period of one or two 1L. Appelbaum, "Officers Turnover and Salary Struc- ture in Loca Unions," Industrigl 3nd Labgr fielgtiogp figzigg, XIX, No. 2 (January, , 22 -23 . e ' I - A J . ‘ - e I r I ‘ . ‘ ’ I . ’ L ' P ' ' . . . i I A I‘ ' . A ' - . e ' ' C e ‘ . l .v ’ y . _ . I ' I . I D ‘ I ‘ .Ae . . . ' . . , r P g ' ' .. \ . ‘I ‘ I - ‘ \ ~ . a. fi F I ~ 1 ’r 'c‘ ' , \‘ .«tJ " A I A l‘ J ‘ . e .- " V " - ‘ ' ‘J . ,. Q \ l U -. , . . I r A ‘ L 1‘ ‘ f . n l ‘I J O . 'L . r t._ ’ .‘ p- " ‘\ r - - t - v . . ‘ c' n-l" "' 1 ~ ' . I ’ av ‘ ‘ ~‘ I " (a V' 4 ,I ’ ” ‘ ' I I. ' . . . t J I . I . 1 . q I- . ;, 1 ‘h a ne'f ‘ ' " Y ’ ' . .D . A . . ‘ o . 4 ‘ O ' " “ . r . , _ e P . I I . . . ‘ . ' years only._ 3. There is no need for a big campaign budget in the local, which makes it easier for new candidates to challenge the officers in power. A. There are close relations between the local leader- ship and members. 5. Local officers are moving to higher positions in the International union. 6. Local officers are not highly compensated in mater- ialistic terms.1 This type of explanation can be found in.a few other cases. Seidman, for example, emphasises two reasons for higher turnover in locale: (l) the local officers' positions are not of any personal advantage to them; (2) it is impossible for local officers to build machinery which.will back them.in power.2 None of these studies however, suggest any theoretical framework for research and explanation. This study will focus on the phenomenon of leader- ship turnover at the local union level. The term.leader- ship refers to the g;ggp_g£_pggp;g who lead the organisa- tion. This group exerts more influence than any other within the organisation. This influence can be summarised 10.P. Bloom, and H.R. Northrup, Be on c of La r w, 5th ed. (Richard D. Irwin, lflfl‘l‘“. “'92" 2J. Seidman, "Democracy in Labor anions," o gf Politigal Ecopogy, DXI (June, 1953), 221-231. along five major lines: 1. The leadership defines and modifies the goals of the organisation. 2. The leadership determines policies, methods, allocation of resources, rules, and regulations, aimed at achie— ving the goals of the organisation. 3. The leadership is the most influential factor of all organisational-administrative processes, i.e., plane ning, coordination, communication. A. The leadership is responsible for solving internal con- flicts in the organisation. 5. The leadership represents the organisation.and serves as a link to other organisations, the community, and the whole society.1 A The importance of turnover is implied by the impor- tance of the leadership: changes of leadership might inter- fore with the smooth operation of the organisation, or fac- ilitate achievement of its goals, depending on the particu- lar situation in a given organisation. Turnover has been studied under different names, in various contexts, and in different organisations. A survey of the literature suggests that the available material about turnover can be classified in four categories: A.‘ 1P. Selanick, dershi in Administrat on (Evanston, Illinois: Row Peterson 0., , o are a s milar presentation. 1. 1 2. 1 3. f A. ' the. some this 5960 391V The 6 l. Turnover of blue-collar workers (hourly-paid employ- cos). 2. Turnover of managers and salaried employees. Here the terms used are generally "mobility" or "succes- sion", and rarely "turnover".2 3. Turnover of executives, usually top executives (one definition: 1 percent to 2 percent of the highest paid executives in the organisation). Here the term used is "succession". A. Turnover of leadership in political or "voluntary organ- isations". The first category is not covered in this study, as the emphasis here is on Egggovgr of legderghip. However, some theories, empirical results, or methodological tools of this category will be used if found to be of value. The second and third categories are relevant to this study and serve as the basic source of ideas, theories and hypotheses. The fourth category is our main concern. we can now define the term "turnover" as it will be used in the study: turnover is the change of leadership in an organisation, as eXpressed by the number of leaders who 1It is interesting to mention Levenson's footnote: ”Sociologists refer to succession on lower levels by the negative toned term 'turnover'". In "Bureaucratic Success- ion", in Etsioni Com le Or anisat o A Sociolo ica , Rogdgr, (Holt, Rinehirt and Winston, 13325), p. 363. 21 id., p. 362. This passage provides a short analy- sis of t e metamor osis of the term from "change of reli- gions and politics leaders" to "turnover of executive". 7 served in a given local during a given period of time. The leadership group in this study includes the President, Vice President, Recording and Financial Secretaries of the local union. In spite of the importance of leadership turnover, as discussed above, little has been done in this area, and Carlson's footnote would appear true today: "The meager systematic literature on succession and its conse- quences is largely descriptive in nature and tends to over- emphasise the disruptive aspects. Propositions are seldom developed or tested."1 It is possible, however, to indicate some basic approaches or partial theories which deal with the ques- tion of turnover. These approaches vary in degree of sophistication, levels of discussion, and the particular aspects with which they deal. One attempt to cope with the issue of succession was made by Grusky. Grusky starts with Gouldner's theory,2 conducts some empirical studies, and suggests the following theoretical framework: Succession is a universal phenomenon which is found in every organisa- tion and is highly important in understanding the organisa- tion. Succession is a complex problem, and __ 1R.0. Carlson, "Succession and Performance Among School Superintendents," Agginistrative Science Quarterly, 1961/62, p. 210. 2Gouldner's theory is mainly concerned with organisa- tional changes as a result of succession; therefore, it is not useful to our study. our orientation toward succession is therefore divided into two interrelated though analytically separated parts: Succession from the point of view of the successor's small group system of role- relationship and succession from the point of view of the formal organization as a totality.1 Dealing only with the formal aspects of the organi- sation and emphasizing the relationships between succession and control systems as a basis for his theoretical frame-6 ‘work, Grusky uses the analysis of weber's theory.2 Accor- ding to weber, the three bases for establishing control, which he calls leadership typology, are: 1. Identification with the personal qualities of the leader: charismatic leadership. 2. Shared or common interests between the organisation's members and its leader: traditional leadership. 3. The established pattern of authority: bureaucratic leadership. A historical analysis, Grusky argues, will reveal that American business firms went through the three leader- ship stages suggested by Weber--dbarismatic to traditional to bureaucratic-~and this order does affect the rates of suc- cession. Succession will change as the type of leadership changes. When organisations are led by charismatic leaders, 10. Grusky, "Administrative Succession in Formal Or- ganisations," Soci 1 Forces, XXIII (Dec. 1960), 106, and 'Corporate:-Sise, Bureaucratisation and Managerial Succession," 9 .Aggricgn Qourggl of Sociology, LXVII (1961 261-269. 2R. Bendix,.Ma ber--An tellect 1 Portrait (Garden City, New'Iore, fieufilediy'&.Co., Inc., T955), p. 302. 9 turnover, almost by definition, is Mapossible, as the pro- bability of the immediate appearance of a second such lea- der is low. When organisations move to their traditional, and later to bureaucratic stages, succession increases. Concerned only with the formal aspects of the organi- sation, Grusky considers size to be one of its important characteristics, and therefore studies the relationships between succession and sise as measured by sales volume. Based on his findings, he later concludes that ". . . size of firm was directly related to the organisational stabil- ity as measured by length of life of firm", and these two variables, siae and age, are the major components of bur- eaucratisation. Grusky's basic finding is that "frequency of administrative succession at the top is directly related to sise of firm."1 His explanation of this finding, that succession increaseasas bureaucratisation increases, is that "bureaucratisation (should) serve to nullify, or at least cushion the otherwise disruptive consequences of succession."2 The softening effect of bureaucratisation on succes- sion is possible because personality factors are relatively less important in an organisation where written rules and regulations control the behavior of its members. These mem- bers are accustomed to accepting authority from above without lnecesaarily relating it to any particular individual. 1Grusky, 1960, p. 105. 21bid., p. 260. 10 Grusky's findings and theory are in contradiction to.Michels' theory, and the reason for this seems to be the fact that they deal with two different types of organ- 2. 3. L. 5. 6. are usually heavily loaded. isations. cerned with type 13.1 Organisational Facets 1. How is leadership achieved I!22.A______ Appointed by higher authority. Who offers leader-Higher strata of ship and, later who accounts for the leader? What is re uired of the lee er? Do followers re uire periodi- ca assurance? What motivates4 obedience of followers? Are leeders' decisions challenged? “—— 1 h—n the organisa- tion. Managerial and technical skills. He Hierarchial auth- ority of the or- ganisation. Con- trol by reward and punishment mainly materiai- istic. Rarely by lower strata. No direct relations between leader and follow- ers. Conflicts solved by third party or higher authority. Grusky deals with type A while Michele is con- Izpe B Elected by members of the organisa- tion. Lower strata of the organisation. Political skills. Yes. Established recall system and periodic elections. Legitimacy of the elections. Reward and punishment usually nonsmater- ialistic. By any member of the organisation and direct leader- follower relation- ship. A and B are used here to avoid using names which This typology was influenced by E.E. Raphael's discussion in "The Anderson-warkov Hypo- theses in Local Unions: A Comparative Study," The 53erican Sgciglogical Regigw, XXXII, No. 5 (October, 19 7 . 7 - 7 . 11 This typology explains the contradictions to some extent. we can understand, by identifying two types of leadership, each of which seems to be subject-to different "laws". This study is concerned with organisations of Type B, so its hypotheses will be derived from Michels' theory. However, Grusky's approach is presented here for two reasons: (1) to support our formal approach to the study of organisation and leadership, (2) to illustrate the organisational variables and their treatment. It seems fit now to phrase the problem of this study: To identify the organisational variables which influence leadership turnover in local unions and to evaluate their contribution to our understanding of the phenomenon of turnover e OUTLINES OF THE STUDY The next two chapters include the theoretical and the methodological aspects of the study. Chapters IV, V, and VI include the presentation and analysis of the results. Chapter IV discusses the structural organisational charac- teristics of the local union. Chapter V deals with the phenomenon of turnover per se and Chapter VI is a discus- sion of the relationship between the structural variables and turnover. The last chapter includes a summary of the findings with emphasis on their relation to the organisa- tional theory and their possible contribution to it. In addition, some research propositions are suggested. CHAPTER II THEORIES, CONCEPTS, VARIABLES, AND HYPOTHESES W The discussion in the previous chapter and the pre- sentation of the problem indicate that Michels' theory is the best theoretical framework for this study. Although his theory was originally meant to be an explanation of the internal forces in democratic organisations, especially political parties, it has become an important part of organ- isational theory. The major conclusion.Michels presents is that the "real law” of organisational life is oligarchy. His general argument can be,divided into three parts: 1. The stability of leadership. Once in power, Michele contends, leaders are unlikely to be displaced. Their advantages are numerous: with the increasing sise and complexity of party and union organisations profession- alism invariably develops and the leaders acquire an expertise denied to the mass; the leaders develop political skills while, conversely, the masses are almost universally incompetent in politics; the elected officeholders soon acquire what comes to be regarded as a moral right to the office and is customarily removed only for extraordinary reasons. Moreover, by co-opti the talented and ambitious members the leaders skin of potentially new and intelligent opposition leaders, thereby insuring the continuing impotence of the masses. 2. The need for organisation. In the struggle with their powerful employer adversaries, organisation--effective and hence centralised organisation--is a necessary con- dition for success. Its opposite, disorganisation, represents disunity, ineffective power and invariably the triumph of the employer interests. The imperative 12 13 of successful organisation-oleaders with ex rtise, a professional bureaucracy, centralised aut ority-- combine in.Michels' view, to make government by the masses a technical impossibility. 3. The iron law of Oligarchy. Confluence of the various factors contributing to the stability of leadership and the imperatives of nondemocratic or- ganisation produce.Michels' celebrated 'iron law of oligarchy'. The Tathy of the masses with their need for guidance has ts counterpart in the natural greed of leaders for power--"what was initiated by the need for organisation administration and strategy is comp plated by psychological determinism"--for the leader identifies his personal interests with those of his erganisation. Thus the organisation no longer exists merely as a means to the ends which it was created, but develops in addition its own peculiar ends, notably, 1 the perpetuation in power of the incumbent leadership. Although Michels' theory does not explicitly mention turnover as one of the characteristics of oligarchy, it can be implied. Lipset g§_;; do so when discussingMichels' theory as a basis for their empirical study, Union ngoc- zggy. They mention two major components of democratic or- ganisations as opposed to oligarchies: (1) organised legi- 2 Michels theory timate opposition, (2) turnover in office. has been empirically tested by various students of the Labor Movement and the Lipset g§_gl study, Union Democrac , is probably the most famous one. The results of this study do not support.Michels' Iron Law of Oligarchy, but the Inter- national Typographical Union, the subject of the study, is atypical, being the only union in the American Labor 1cm. Magrath, "Democracy in Overall: The Futile Quest for Union Democracy " Ind tri 1 nd Labor Relations £93114: (July, 1959), p. 5 -5 . 23.x. Lipset, n. Trow, and J. Coleman, Liaise Democracy (Free Press, 1956), p. 3. lb Movement with a clear out two party system. Other studies examined the turnover of leadership in international unions. Taft devotes one chapter to the sub- Ject of elections, and his results indicate low officer turn- over at the International level. The study covers thirty- four internationals during 1900 to 19L8. In this period 2307 officers were elected. In only 18.8 percent of the presidential elections were incumbents challenged. Incume bent officers were challenged in only 23 percent of all the cases.1 These results support Michels theory. Another study was conducted by Brown, who was mainly concerned with intraunion conflicts, and studied leadership turnover as one of its variables. The union under study was the International Ladies Oarment'workers Union, over a per- iod of fifty years. The study concludes: "This signifies that the ggglley the Epics the greater the proportion of top leadership pg; re-elected to office, or, in other words, 3h; ate the t n v ."2 _ It seems that we do have empirical evidence to sup- port Michels' Iron Law of Oligarchy when related to the International level. The next step is to establish a con- ceptual framework by which we will be able to extend our examination to the local level. 1P. Taft, The Structure and Or anisatio of Labor Ugiggg (Cambridge,.Maas: Rarva University Press, 2J.S. Brown, ”Union Sise as a Function of Intraunion Conflict,” Hgggn Relatiggg, II (1956), 75-89. lat «r 15 THE CONCEPTUAL FRAMEWORK 4Michels' analysis of the creation of oligarchies has been elaborated on by others, and various factors and processes have been suggested as a basis for empirical research. Two of these elaborations will be presented here. The first one is Truman's analysis of the "active minority" (his term.for leadership). According to Truman, the following items are considered to be factors which actually contributed to the creation of oligarchy: 1. The mere existence of a formal organisation, which results in the creation of ygriogg fggctigns, the estab- 1ishment of some kind of a hierarch , and the gggggg thag decisiog; by ggde. I 2. The graces; of ggowth of the organisation, which bears a serious effect on the integration of the various units within the growing organisation. 3. 1!; financial stygcture. A. The pgrsoggl chiracteristics of the leaders. 5. The relative pgggyigg of the organisation as related to other organisations. 6. IThe type of follgggrs. Truman's suggestions are basically relevant to politi- l cal organisations, and since he is mainly concerned with political parties, his analysis is directed toward the basic 1D.B. Truman, The Cove Process (New York: Alfred A. Knoff, 1959 . C C s a . - . .- ‘ I '.-. m m s . . ‘ V n - i s a TI 16 problems of these organisations such as financial struc— ture. Some of his conditions, however, can be applied to other organisations, including the local union. Lipset g§_g;,also elaborated on Michels' theory. According to them, the factors which facilitate oligarchy in unions are: l. The development of a negr monopgly of pgger which is possible because of: a) The development of a Burgggcrgtic structure which actually means the increase of power at the top. b) The control over the formal means of cggggpigation by the top. c) The monopoly of the pglitical skills by the top, while these skills are absent among the rank and file. 2. The leadegghip's desire to stay in power. 3. A decrgasg in membeyg' pgrticipgtiog in union activities.1 An analysis of these two elaborations shows that there are two types of concepts which seem to explain the creation of oligarchy: 1. Structural and situational concepts, such as "bureau- cratic structure", "relative prestige", ”communication". 2. Factors of personality and social psychology, such as "the leadership's desire to stay in power", "types of followers". 1Lipset at al, Union Democracy. 17 In the early stages of development of the organi- sational theory, structural concepts were used as basic elements to explain organisations (for example, Weber's model of bureaucracy). The major weakness of these early theories was their total neglect of the informal aspects of the organisation. Other theories emerged which were more concerned with individuals and groups within the organisation (for example, those of.M.P. Follet, the "Human Relations School", and others). These theories, with some modifications, still prevail and were supported during the last period by numerous empirical studies, star- ting‘with the famous "Hawthorne Studies". Although the formal structure of the organisation was considered a fac- tor, its importance seemed to be secondary to that of socio- logical and psychological factors. The theories now seem to put less emphasis on the .conceptualisation and research of the structural and formal aspects as compared with the socio-psychological aspects of organisations. Therefore, our knowledge of structural and formal aspects is substantially less than that of the socio- psychological aspects. Our approach in this study is formal. However, as this approach is less developed than the socio-psychological, and because of the limitations of the study, we may find our- selves unable to explain some of the results. In such cases, explanations of a socio-psychological nature are suggested. 18 Based on the above discussion and on additional literature which is included in the bibliography, four formal concepts are suggested as the core of the concept- ual framework of this study and the source of its hypo- theses. These concepts are Formalisation, Complexity, Attractiveness, and External Elements (or Environment). Each of the above concepts can be operationalised by different variables. The two crucial considerations in selecting the variables for each concept are: (1) to what extent does the variable "represent" the concept?; and (2) is information about the variable obtainable? The final decision about the ”representation" of a variable depends on the researcher's Judgment. Accessibility of information about a variable depends on two conditions: the existence of the information and its availability (which, in turn, is dependent on factors such as money, time, reputation, and the expertise of the investigator). This study, being no exception, is also subject to these two conditions. .IBE CONCEPTS THE VARIABLES AND THE Foyggliggtion The concept of formalisation can be defined in.more than one way and on various levels. The following two quo- tations will serve as the source of our variables: The formalisation process continues as an organisation gets older and larger. . . Patterns of behavior l9 stabilise; individuals settle into characteristic roles; standard operating procedures are established. The most significant characteristic of this develop- ment is that it represents organisational learning-- the learning of a formal organisational structure, expressive of the problems which must be solved and a rationale for solving them. The formal structure provides the necessary framework within which labor is divided and specialised, responsibilities are delegated, routine communications are systematised, and inducements are allocated.1 A shorter, but fairly similar statement can be found in Tsouderos: ~ _ The process of formalisation is defined as a sequen- tial, stage by stage develo ent of organisational activity over time: a stan ardisation of social relationshi and finally as an increasing bureaucrati- sation of e organisation. ' . These two definitions have one major characteristic in common: both describe organisational processes during a given period of time and define them as "formalisetion". These processes, like "stabilisation of behavior”, "alloca- tion of responsibilities", "establishment of standard opera- ting procedures", and ”routinisation of the communication channels” are hardly variables which can be observed in an organisation, and further operationalisation is necessary. This operationalisation can be done in various ways. One method is to identify a specific variable which can be observed and represents a specific process. Here, the -¥ 1W.H. Starbuck, "Organisational Growth and Develop- ment" in J.G. March (ed.) Handbook of Or anis tion (Rand Maclally &,Co., 1965), p. A75. 2J.E. Tsouderos "Organisational Change in Terms of Series of Selected variables," American Sociological Review, 11 (April 1955), footnote 3, 2 . 20 existence of written operating procedures will be the obser- ved variable for the process of establishment of standard _operating procedures. ‘ Another possibility is to identify the basic charac- teristics of the organisation which can be observed and/or measured and which are related to the process of formalise- tion. This study, because of its approach and structure and the nature of its data sources, will use the following characteristics as operationalised variables of the process of formalisation: (1) age; (2) sise as measured by annual income and number of members. Starbuck's quotation.mentions age specifically as a factor influencing formalisation. The second quotation (Thouderos) considers age implicitly, by relating the pro- ceas of formalisation to the time dimension. The sise of the organisation is also explicitly mentioned in the first quotation as a factor influencing the formalisation process. The literature, which considers sise an indicator of formali- sation, discusses how it should be measured. In business organisations for instance, some of the common measures are sales volume, number of employees, total assets. The two significant measures of the sise of the local seem to be membership and annual income. Relationships between.membership sise and formalise- tionwwere theorised by Coplow whose basic idea is that when 21 the number of members increases there is a growing need for the institutionalisation of interactions, which, in turn, increases with the increase inmembership.1 It seems that when Tsouderos speaks about "stan- dardisation of social relationship" he is partially referring to this concept. The three variables suggested (age, sise of member- ship and sise of income) can be considered as static variables as they reflect the situation of the organisa- tion in a given point of time. However, the phenomenon of turnover is dynamic and can be measured only for a pagigg of ting. In order to add the dynamic dimension to the static variables of formalisation, two variables will be added: growth rates of membership and of income. These variables will indicate changes in the organisation during a specific period of time. Based on the above discussion, and in accordance with.Michels' analysis, we suggest this overall hypothe- sis: As formalisation increases, stability of leadership increases. This hypothesis is the equivalent of one that can be tested in our study: As formalisation increases, turnover of leadership decreases. The more specific hypotheses of the concept of for- malisation are directly inferred from the above discussion lT. Coplow, "0r anisational Sise," Administrative Science Quarterlz,—I, fi957),h8A-505. 22 and the specifications of the variables: Fl1 - The older the_ipcal. the lower its leadership turnover. F2 - The bigger the legal in size or in total income,_th§ lower its leadership turnover. F3 - The higher the rates_of growth in sipe of membership or total incomeI the higher itspiggdership_turnover. Complpxitz Hall and Bass in their research define complexity as:’ “The degree of internal segmentation--the number of separate parts of the organisation as reflected by the division of labor, number of hierarchial levels and the spatial dispersion of the organization."2 As complexity increases, more skills are required to lead the organiza- tion. The control of the incumbent leader increases too, as they are familiar with the various organisational com- ponents and activities. At the same time, the disadvan- tages of any apposition to the existing leadership increase. The selection of variables which represent this con- cept is especially difficult for type B organisations. variables which can be used in type A organisations, such as the number of hierarchial levels, are meaningless in the local union. A close look at the locals included in this study suggests the following variables as a possible measure 1Letters and numbers will designate a specific hypo- thesis and will serve as a reference in following chapters. 2R.H. Hall and J.E. Bass, "Organisational Size, Com— plexity and Formalization," American Sociolo ical Review, (XXII, No. 6 (December, 1967 , 90 . In) 23 of complexity: 1. The number of units in the local. This number (which represents the number of different companies or plants in which local members are employed) is an indication of heterogenity, diversity of viewpoints, and demands. 2. Total resources and facilities which can be used by the local. These include total resources of the local (cash on hand as well as property), and the existence of a local newspaper. It is suggested that the more a local owns (compared to other locals) the more complex it is, since more skills are involved in running it. 3. Total money expenditure for personnel employed by the local. The assumption underlying this measurement is that the more personnel the local employs, the more activities it offers, and therefore it becomes more complex. The total payments to local union personnel is considered an indication of its complexity.1 The overall hypothesis is: as complexity increases, leadership turnover decreases. The specific hypotheses are: - e r the n r of loca an t the h purppver of its leadership. Q2_:_ngals which_gontrol more resources_(own their own hall and offices and publish their own newspaper)_will have 1The third variable, and to some extent the second, raises the question of the relationship between complexity and sise. Hall and Hess conclude that these relationships are very weak. 25 lower rates of turnover of leadership. lowpr the tpppover of its leadership. Aptractiveness As stated previously, only the formal aspects of this concept are to be examined here. Under this condi- tion, the suggested definition of the concept will be: attractiveness is the force which stimulates potential leaders to seek leadership positions and causes leaders in office to stay there. The general assumption is that the more attractive the position the less will its turnover be. Attractive positions will cause many new candidates to come pate for it. On the other hand, this attractiveness also means that the incumbent officer'will do his best to stay in power. Usually, the incumbent officer has advantages over new candidates, and his chances of staying in power are high, especially if there are many candidates, and assuming that his performance in the past satisfied the demands of the members. ’ The formal.measurement of attractiveness to be con- sidered is the materialistic compensation to the leader- ship. At least one more important variable of attractive- ness is already included aaong variables suggested earlier. Payne1 shows that an organisation's sise and prestige are *R. Payne, "An Approach to the Study of Relative Prestige of Formal Organisation," Social F rces, XXIII, (1951:) . adv-2&7. 25 highly correlated. It is logical to assume a relationship between pres- tige and attractiveness, therefore, it can be argued that the larger organisations have more prestige. Following the same reasoning, the leadership turnover of the more prestigous organisations will be lower than smaller or less prestigous groups. The specific hypothesis suggested is: Al - The hi or the com iower its turnover. External Elepents There is very little doubt that the environment in which an organisation operates is of great importance as a ficter influencing processes within the organisation. Howe ever, the direct relationship and influence are still un- known. An interesting illustration of an attempt to analyse this factor is Munro's article.1 In this article, Munro studies some external forces of the local union, and examr ines their influence on the processes within the local.‘ As the concept of external elements is very broad, only part of its factors will be studied in this research. It is assumed that the external influences occur on various levels. The lowest level is the immediate environment with which the A 1J.L. Munro, "External Factors of Control in the v.3.a. Labor Union Locals " o nal of Industrial Relat ons (Australia). 1: (l9oe).mT“‘~135-1 5. ' ' ' " 26 organisation under study heavily interacts. The second level is the larger community of which the organisation is a part. The third and broadest is society as a whole. The first level is interorganisational relations. The variable suggested for study of this level is the pglppnyfiwith‘which the local has a labor agreement. The relations between the local and the company have direct bearing on the standing of the leadership and determine the members evaluation of their leadership. This evalua- tion is supposedly an important factor influencing turn- over. However, our knowledge at this time does not permit us to predict any specific relationship.1 The hypothesis we suggest therefore is: El - Diifprent ipcaip copppcted with difierppt cpmpgpiep l diffe t t or rate . The second level suggested for study here is the come munity. In this case the suggested variable is the sise of php_ppppppipz in terms of its population, adjusted by two additional factors: (1) the percentage of blue collar wor- kers in the particular communitYS (2) the membership sise of the local. The sise of the community seems to be an impor- tant factor, as it influences the prestige of the local union leadership. In a small community, where local union leaders 1The only detailed treatment available is found in F.H. Harbison and R. Dubin, Patte of Unio Mans ement Relat o a On ted Automobi } General Motors, Chicago: Science Researc Association, 1§A7). 27 are an important part of the community leadership, the prestige and the attractiveness of the union leadership increases, and, based on the discussion of attractiveness, its turnover decreases. Based on the above presentation we hypothesize that: E - l comm tie t e t over of local uni n ipadership giii pe lpger. The third level is the society as a whole. It is very difficult to study the influence of this factor as it has many components and lacks an operational definition. we suggest testing the influence of two general factors which seem closely related to the influence of the ”whole society". One factor, in its most operationalised expression, is the influence of the element of time on the turnover of officers. 'Actually, we will ask whether changes in standard of living, economic activities and value systems cause dif- ferences in rates of leadershiplturnover in different time periods. Namely,'whether in time t1 we will find different rates of turnover than in time t2. ' The second factor is mainly an attempt to relate leadership turnover, with the element of time, and the negotiation of the labor contract. The question is whether 'we will find different rates of leadership turnover in years of negotiations as compared to years in which no nego- tiations took place. We assume that the negotiations re- fleet economic and other related issues beyond those of 28 wages and other materialistic compensations.l Based on the above discussion, we suggest the follow- ing two hypotheses: E a - There e differs ces rates of leaders tpgppver ip the differept zears covered by this study. B b - More leadershi t over will be found in ears of ippor contract negotiations and immediately after. 1It was suggested to me in one of the interviews which.were conducted as a preparation for the study, that the most decisive factor influencing turnover is the economic factor, namely, the economic issues the local has to cope ‘with and the success of its leadership to satisfy the members in this respect. To the contrary, we could not find any sup- port of this argument in the literature. Sayles sld Strauss in their Local Union-- ts - ce in the Industri 1 Plant (New York: Harper, 3, revise l9 , argue t at " t ough we have listed a number of situations where economic issues play an important part in union elections, it should be emphasised that these are more or less exceptions". (p. 97) The examina- tion immediately above relates the turnover to the total labor contract, including the economic and the noneconomic issues. CHAPTER III DATA AND METEODOLOG! Research and theory are now re—emphasising the for- mal aspects of organisational behavior as opposed to the past tendency of emphasising the socio-psychological as- pects and studying individual and group behavior. we have come to realise that better understanding of organisations rests on consideration of units of analysis such as the structure of the organisation, the interrelations of the positions, communication networks. This method does not neglect individuals and groups, but regards themtas "given" in.a formal study, or as "products" of the organisational characteristics in other types of studies. Accepting this trend as valid and meaningful, the approach of this study will be of a formal type. This ap- proach does not deny the fact that turnover is a behavioral phenomenon, but suggests that in this study its formal as- pects will be emphasised. Hage's article, "Axiomatic Theory of Organisation", explains the advantages of studying the formal characteristics of organisations: 1. These characteristics can both differentiate between organisations with similar objectives and indicate simp ilarities between organisations with different objectives. 29 30 2. They are not time specific or cultural bound. 3. Because of the above two, they are useful in studying organisational evolutions.1 These points, especially the first and second, are relevant to this study which deals with the differentiations between organisations with similar objects over a period of time. The decision to emphasise the formal aspects dic- tates the type of data and the methods of collecting it, but does not exclude the use of techniques generally used in the socio-psychological type of study. Another consideration of the study is related to the type of the dependent variables. The major variable is the phenomenon of turnover; therefore the study's time span is important. ‘He have selected a nine year period (1959-1967 inclusive) as this is the longest period with adequate available data . 1959 was the first year in which the local unions had to file their annual reports under the "Labor Manage- ment Reporting and Disclosure Act", and 1967 was the last year for which information was available for this study. This period covers five elections and although this time span is long enough to lead to meaningful conclusions, it is not long enough to identify patterns of change. A stratified (by sise of membership and type of 1J. Hage, "Axiomatic Theory of Organisation," idministgatiop Science Qpprterly, X, No. 3 (December, 19 5 g 2 -320e 31 local) random sample of 100 United Auto workers (UAW) locals was drawn from the population of 275 U.A.W. locals in Michigan.1 SOURCES OF THE DATA Many information sources have been used in data collection relating to this study. Mbst of the data is of a secondary nature. An important characteristic of much of the data is that it was obtained from sources open to the public. These include annual reports and U.A.W. material filed by locals according to the above mentioned Act.2 This characteristic makes the replication of the study relatively simple; an important advantage in a field ‘where practically nothing has been previously researched. Other sources are more unique to the study, especi- ally those which are related to the U.A.W. (excluding the publications for the public). Many researchers have stu- died the U.A.W. because of its unique characteristics as ‘well as the ready aid and enthusiastic help of its leader- ship. In this case, as in.aany others, the cooperation was much more than any researcher could expect. Appppi Reppzps of the ipcai Upiopp Every local must file this report with the 0.3. .Department of Labor as specified in.the Labor Management 1For detailed description and reasons for selecting 'the U.A.W. see Appendix A. 2See. 205 of the Disclosure Act states that all re- ports and documents filed shall be public information. 32 Reporting and Disclosure Act of 1959. These reports fur- nish the information about turnover and most of the inde- pendent variables. Gozpppppptai Spurcep Some items of information were collected from other Governmental sources: a) Census information which is used in the analysis of community influence. b) Other publications of the 0.3. Department of Labor. Q,A,w, Conyention Prpceedipgs This source was used for three major purposes: a) To establish the population's distribution by sise and type as a basis for the sampling process. b) To furnish information about the turnover and election of regional directors. c) To achieve some degree of insight into the rela- tionship between the locals and the International. Of spe- cial interest and importance are the conventions of 1951 and 1953 (Previous to the period of the study) when changes in the term of service of the local officers were discussed. 0,5,w, fipipted Materiai These include the International Constitutions of the last nine regular and special conventions, local by-laws, labor agreements, educational publications (such as 33 "Election guide", "The President's Job") and other descrip- tive material of the U.A.W.1 U,§,W, Compilation fpr phis Stpgy Some information was compiled especially for this study by the U.A.W. staff. This includes information about the age, type, and size of the 100 locals at the beginning of 1968 and information about the turnover of regional dir- actors and staff members of U.A.W. International Headquar- tors e Ippprvien with U, MW, Off icigll Interviews were conducted for four different pur- poses and therefore used different techniques. 1. Preliminary interviews which were aimed to familiarise the researcher with the problem, to explore available infor- mation, and to foster a free exchange of ideas and methods to be applied. These interviews were carried out mainly in U.A.W; Headquarters and included the staff of the several departments. The interviews took the form of general dis- cussions, without any particular structure. They ranged from.a half-day conference with the peeple in the Research Department to a quick phone call for clarification of a par- ticular point. 2. Later we decided to interview a small sub-sample of local presidents to find some more "from the field" explanations of 1Formoro details see Bibliography section, U.A.W. Documents. 3A the statistical findings. In order to construct the ques- tionnaire we conducted additional interviews in the field and in Headquarters. These were mainly aimed at determin- ing what information would be feasible for the local offi- cers to supply. 3. Four interviews were carried out as a pro-test of the interviews to be conducted with the sub-sample. A. The presidents of eight locals were interviewed, repre- senting the locals of the four highest and four lowest turn- over.1 The average time of the interview was an hour and a half using an Open-ended questionnaire. UALITI OF THE D T The annual reports of the local unions furnished :mest of the data used in.this study. Because the original purpose of these reports is unrelated to the purpose of the study, the reports have a high degree of objectivity. Such objective information is necessary for this study's .formal approach. Three reasons make the accuracy of the data highly reliable: :l. The law specifies high fines for violating its provi- sions (up to $10,000 fine'and/or one year in prison). 2. The U.A.W. is well-known for its integrity. 3. There is no particular reason for the local to manipu- late its reports. 1Formore details about the sub-sample see Appendix A. And Appendix C. 35 As the reports are required by law and as the U.A.W. has a well-organised "red-tape" system, only 1.7 percent (16 out of 900) of the reports were missing, and only about 3 percent caused any problems. The sequence of the reports helped clarify most of these problems. The quality of the other printed materials, the Governmental and various U.A.W. sources, seems to be satis- factory. Basically none are directly related to the sub- Ject of the study, nor do they have any particular reason for any bias. It is difficult to evaluate the interviews; however, with the exception.of one aspect (the attitude toward the International), the respondents were quite candid and answered all questions frankly and to the best of their knowledge. It is important to mention however, that their knowledge with respect to some issues is questionable as it required a very good memory for details. METHODOLOGY The methodology of this study is selected to achieve .four goals: 1. To describe the phenomenon of turnover,and to compare it to other available descriptions. To describe some of the structural characteristics of the local union. This goal will be achieved by using instruments such as regular tables, and frequency distribution. 2. To establish functional relationships between the 36 variables by using I? and correlations. 3. To test certain hypotheses about the functional rela- tionships. This will be achieved by using signifi- cance levels for the above measures. h. To describe the contribution of the various indepen- dent veriables to the explanation of the dependent variable. Through computers it is possible to apply sophisticated testing methods, as specified above. A question arises regarding the type of statisti- cal methods to be used-~perametric or non-parametric. As the nature of the data is still unknown, and as no clear cut guidelines are available, both parametric and non-para- metric statistics1 will be employed. This, in turn, calls for special caution in the interpretation of the statis- tics to be presented. 1For a discussion of this point see Siegel Egg- Rgflnetric Satistics, (1965), Chapter 3, pp. 1845. CHAPTER IV E STRQCTURAL VQIAQLES 01" EB LOCALS SIZE OF LOCALS ED RATES OF GROWTH is the membership sise of the local, measured by number of votes in the conventions, was one of the two criteria for sampling, the. sise distribution of the sample is an accurate reflection of the sise distribution of the UJJI. locals in Michigan with one exception: the eleven largest locals are included in the sample.1 Table IV-l illustrates distribution of the 100 locals according to the two measures of sise used in this study--total annual income and membership. The smallest local (when an average of the 9 years has been used), has about 25 members while the largest has more than 30,000.2 As for the sise of locals in terms of income, the smallest income is about $1,000 per year, while the largest is over two million dollars average annual 1See Appendix A for the detailed sampling procedure and the comparison of the population and the samp e. 2The mean sise of membership was about 2200 and the standard deviation L152. All the above figures were calcu- lated by dividing the local per capita taxes each local paid to the International by the per capita for each member, which is $36.00 yearly. 37 kl 38 TABLE IV-1.--Distribution of sample locals by income and membership (annual average). Income Group Number of Membership Number of in 1000 Dollars Locals Group Locals 1 - 100 72 25 - 600 50 101 - 200 a 601 - 2500 26 201 - 300 5 2501 - 5500 8 301 - 500 6 5501 - 8500 h 501 - 1000 10 l 8501 - 10000 6 1001 and up 3 10,001 and up 6 Total 100 Total 100 income ($2,059,000.00).1 The simple correlation between the size of member- ship and the income is R=.99.2 This high correlation is a result of the fact that the major source of local union income is its members' dues. Because the two measures are highly correlated, the calculations later on in the study 1The mean income of the locals is $168, 000. 00 and the standard deviation $321, 057. 00. 21f not otherwise specified, the correlation figures hereafter are Kendall Rank Order Correlations Coefficient (Tau) . For technical presentation, see S. Siegel ‘flgg- parametric Statistics for the Behavioral Science New York: Mc raw- , pp. 21 -22 . our measures of correla- tions were used: 1. Simple correlation.and partial correlation. 2. .Multiple correlation, as part of the various least s uares calculation. This calculation includes r- tgal correlations and significance levels for a correlations coefficients. 3. Spearman.Rank Order Correlation (Rho) including one tail significance level. .L. Kendall Mnk Order Correlation Coefficient (Tau) one tail significance level, partial correlations without significance levels. The basic measures of correlations (R, Rho, Tau) are usually close and in the same direction. However, in cases where they differ a detailed discussion will be presented. 39 will use only the size of membership.1 Which sise measure should be employed when studying organisations is an important question as sometimes dif- ferent measures will yield different results. Some sug- gestions for size measures of organisations are number of employees, volume of sales, value of assets, membership,- income. The above correlation suggests that when studying sub-units of the same organisationnwith the same financial structure, the two measures of size--income and membership-- are close that either is sufficient. The relationship between the sise of the organiza- tion and its rate of growth is a question'with which many students of organisational theory are concerned. Simon summarises the relations between sise and growth for busi- ness organisations by saying: "It is generally agreed that there is little or no relations between sise of firm and expected percentage of growth."2 1After discussing the concepts in the theoretical chapter, we suggested variables to be studied and analysed as representation of these concepts. The data reveals, how- ever, that sometimes the correlations between variables ‘within the concept are so high that, with the exception of the multiple correlation calculation, it is possible to analyse the relationship of turnover by using only one variable. As a rule we use only one variable in an analy- sis when the simple correlation between.any two variables of the same concept is R (.70. Following this rule we eliminate one sise variable: income. Nb differences in.any results were found when the gt? of income was used and sise of membership was elimina- 2R. Simon "Comment-~Firm Sise and Rate of Growth," J’ l f Po t c 1 Bo , LXXII (Feb. 196k), 81-83. LO The findings of our research are in the same direc- tion. The correlation between the size of membership and the rate of membership~growth‘was found to be Tau=.09, 9509.1 The analysis of the overall relationship between sise and rate of growth is heavily supported by the com- parison between single-unit and amalgamated locals.2 The amalgamated locals are significantly3 larger than the sin- gle-unit locals, however, the one way analysis of variance has shown that the type of the local has no bearing on its rate of growth (P=.718). A tentative explanation can be found in the close relationship in sise between.the busi- ness firm.and its local. The sise and the growth rates of the local are related to those of the firmnwith'which the local has an agreement. This kind of explanation.assumes a maximum.unionisation of employees. This reasoning however does not apply to amalgamated locals which are encouraged to organise as many plants as possible. The factor which seems to prevent the growth of 1The correlation between the rate of growth of mem- bership and the rate of growth of income is R=.80 and follow- ing our rule we decided to use only one measure of rate of growth. 2Amalgamated locals are those composed of more than 0:: unit, when each unit works for a different company or p at. 3Hereafterywhen the term significance is used with- out any number, it means BC .05. This is not the level of acceptance of hypotheses of this study. As the meaning of the significance level is not only statistical, discussion in each case will consider the significance level along'with other concepts. #1 amalgamated locals beyond a certain sise, in terms of units, is their limited resources. As one president put it: "we should erganise more, but we do not have the time."1 AGE OF THE LOCALS AND OTHER VARIA§L§§ The 100 sample locals range in age from ten to thirty- feur. The average age is 2b.9 years and the standard devia- tion is 6.A. The correlates of the age of the local are sumrised in Table Iv-z. This table mount.» a positive relationship between age and sise of the local. This seems to be the major relationship, based on the fact that, when the sise variable is constant, the correlation with pay- ments to employees and officers as well as with total re- sources drops substantially. No relationship‘wes found between the age of the local and the rate of growth or with the community index. It was found that the amalgmmated locals‘with an average age of 27.1 years, are significantly older (P<:.03) than the single-unit locals (2h years). Older locals tend significantly (P‘<2005) to have their own union hall and offices. The average age of lecals with facilities is 28.2 years compared with 22.8 years for those without facilities. The older locals also tend significantly (P<.005) to pub- lish their own newspapers. The average age of locals with a newspaper is 27.1 years as compared to 23.3 years of those without a newspaper. 1Interview'with an amalgamated local president. (‘2 TABLE IV-2.--The Correlates of age with other variables of __ l _ _ 3h; stair. _ _ .1 Level of Membershipb variables‘ Correlation Significance constant Membership .33 P<:.005 Membership Growth Rates (-) .OOL P = .L8 Payment to Employees .31 P<:.005 .07 Payments to ' ' Officers .36 P<:.005 .18 Community IMO! (’) e05 P as e25 (" e12 Total - Resources .38 P< .005 .20 *-¥ ‘Fer explanation of the variables and their measures, . see Appendix B. bAs the sampling distribution of Kendall Partial Rank Corrilation is still unknown, no tests of significance are new'possible. :5; TYPE OF LOCAL The type of local was the second criterion for the selection of the sample. Table IVAB gives the distribution of amalgamated locals in the sample and in Michigan. bSome information about the amalgamated locals has already been presented. They are larger in membership, older, and they do not differ from.single-unit locals in rate of growth (P - .718). Same additional information about the amalgamated locals is included in Table IVAA. Amalgamated locals tend #3 TABLE IV-3.-- Amalgamated and single-unit locals in the sample in Michigan.‘ In the Sample In Michigan Type ef Local Number of locals Number of locals % Amalgamated 28 66 2A locals Single Unit ' 72 209 76 __lonalas_, ._. ___. __._.__ Total 100 275 100 'For a detailed table see Appendix A. to have their own.fecilities in.aere cases than single-unit locals (P<;.001). The amalgamated locals tend to have more newspapers (P .05) and their payments to employees and officers are higher (P<:.05) than those of the single-unit locals. In terms of resources, the amalgamated locals average about $A23,000.00 while the single-units average only $55,700.00 (P<.005). Ne relationship has been found between.the type of the lecal, compeny*with which it is associated (P<:.95). and the community index (P=.11). THE C AC R The community index1 gives information about the relationship of the sise of the local,the sise of the com- munity, and the proportion of blue-collar workers in this 1See Appendix B for detailed discussion of the con- structien of this index. AL TABLE IV¥A.--Relationship between the type of the local and facilities, newspaper and payments to employees and officers. __A _— ZAQILIIIEQ Single Unit Amalga- All Statistical locals mated Samples Tests ___ _____ __ locheL___ .0 __. Locals with 2 Facilities 16 18 3A X = 16.98 Locals‘without Facilities 52 8 60 Df = 2 Locals with Facili- ties Ac uired dur- ‘ ing stu y l. 2 6 P<.OOl Total Number of wvvfi locals 72 28 100 NEWSPAPER Y s_;. __ .2 a Locals Publishing X b.979 Newspaper 26 17 #3 Df = l Locals Without Newspapers 1.6 11 57 P (.05 Total Number of locals 72 28 100 PAYMENTS _ ‘ — Payments in $1000 per year: to em lays-8 1A.8 63.9 28.6 All P<:.05 to al officers 5.8 17.1 8.9 (Parametric to President 2.3 6.7 3.5 One way to Vice President 0,7 3.3 1.h Analysis to Recording Secretary 0.7 2.1 $ 111 of . to Financial Secretary 2:1 5,0 $ 2.9 Variance) #5 community. Some of the correlations of the community index are presented in Table IVAS, and some were discus- sed earlier. TABLE IV-5.--Correlations between community index and other variables. A‘w‘ _. - Tau Level of Viriables Correlation Significance Membership .19 P<;.005 Mbmbership GrMh "’ e 06 P =e 20 Payment to Employees . 17 P (.005 Payment to Officers .13 P =.02 Resources . 16 P <. 005 com NY ‘ The distribution of the 100 locals among the various companies is as follows: Twenty-nine are locals of the "Big Three"1 (sixteen of General motors, ten of Chrysler, and three of Ford). The other seventy-one are locals of other companies and amalgamated locals. Table 1756 contains some comparison of the "Others" and the "Big Three" grouped together and separately. A 1A separation and an analysis was done of the "others" compared to the Big Three locals. The Big Three is considered a special group, as can be seen from publications such as "New'from U.A.W." and "Solidarity House." #6 TABLE IV-6.--"Big Three" locals as compared to others.‘ .—k w _‘_— Locals Others Big Pc G.M. Chrysler Ford Pd 'b Three (16) (10) (3) M2221! _ 71)....(29) ._ _ _ Income (in $1000.) $100. $331.1 .001 $A18.8 $258.7 $133.3 .005 Payments to Employees (in $1000.) $ 17.5 $ 55.6 .01 $ 65.1 $ #0.? $ 55.3 .065 Payments to Officers (in $1000.) $ 5.0 $ 18.5 .005 $ 17.h $ 19.6 $ 20.6 .005 Total Resources (in $1000.) $1A7.5 $186.A .70 $217.5 $16A.5 $ 93.3 .98 Age 21.A 26.7 .07 26.8 26.6 27 .35 Community Index 23.3 21.8 .91 25.2 18.1 16.A .98 Rate of Growth of Alembership lOO.A 131.2 .09 135.8 1A6.6 55 .11 .Using parametric One Way Analysis of Variance bNumber of Locals in the sample. °"Others" and Big Three compared as a group. cl"Others" and each of the Big Three are compared. A comparison of "Others" to the "Big Three" shows that ‘the Big Three locals have more income (an.average of $33t,100.00 as compared to $100,000.00 or P<.005). Their payments to employees and officers are higher (P<. 01). They are older, (P=.07) but do not differ significantly in terms A7 of their total resources (P=.70) and their community index (P=.91). When we compare each of the Big Three separately, ‘we find that General motors (GM)locals lead with the high- est income and largest resources, largest payments to employees, and lowest payments to officers. The total resources, the age, and the community index are different but not significantly so, and, as already mentioned, the rate of growth is not significantly different (P-.ll). Big Three locals tend to have more facilities than "Other" (72.A percent compared to 18.3 percent, 12 test yielded a value of 29,96h, and P<:.001). Newspapers are also more common in the Big Three locals: 79.3 percent of them have a newspaper as compared with 28.2 percent of .the others (12=21,9; P<:;001). There is no association between the type of the local (amalgamated or single-unit) and the company (R<390). It seems from these findings that no one general conclusion can be drawn about the relations between the company and other structural variables of the study. Jnggpgts of the Locals to Employees and Officers As presented in Table IV-A, payments, both to employees and to officers, differ significantly between single unit locals to amalgamated. All the calculations regarding these payments are subject to two limitations. The first is a statistical limitation which is caused by L8 the nature of the distribution of these variables, as pre- sented in Table IV-7. The second limitation is a result of the formal nature of the study. The actual payments are not a direct reflection of the amount of work, as it is well known that in small locals, and sometimes even in large ones, part of the administrative and organisational work is on a voluntary basis. #9 TABLE IV-7.--Distribution of locals according to their yearly payments to officers and employees. *A— * __‘ — Payment to: $O-$500. $501r$1,000. $1001. Total number and up of locals All four 3A A 62 100 President 38 12 50 100 Vice President 66 17 17 100 Recording Secretary 69 1A 17 100 Financial Secretary Al 23 36 100 Other . Employees 0 AA 56 100 .Includes all payments to employees and officers other than the four under study. Another indication of the diversity of payments is the comparison between the average yearly payments of all locals and the highest payment made by a For all four officers: For President: For For Vice President: , Recording Secretary:Average, 1, Average, 8, Average, 3, Average, For Financial Secretary:Average, 2, lFor other employees: Average,$28,600 000; 500; 100; 900; single local. highest, 56,000. highest, 20,000. highest, 15,000. highest, 1A,OOO. highest, 1A,OOO. highest, 5A,600. CHAPTER V PATTERNS OF ELECTIONS, IQRNOVER AND INNER MO§ILI 0? CAL 0 R8 The phenomenon of turnover and the mobility of lea- dership will be analysed in this chapter. In an attempt to describe, compare, and analyse this phenomenon, it will be independently of the other variables. It is difficult to conclude or prove any causal relationship from the data, and it is equally as difficult to explain some of the results. However, based on general knowledge, insight of various people in the field, and material collected in interviews, we can offer some tenta- tive and speculative explanations. COMgAAISOE AND DISCUSSION 0: TURNOVER The study deals with the top leadership of one hundred U.A.W. locals during the period 1959 - 1967.1 ##— 1For purposes of this study, top leadership is cone sidered President, Vice-President, Recording and Financial Secretaries. The deviation from this pattern of top leader- ship was found to be of no significance. Although the con- stitution allows for more than one Vice President, only three of the test cases had a Second Vice-President. One of these had elected a Third Vice-President for a short time. Two other locals had elected a Second Vice-President in the past, and in two cases, the leadership had once included three Vice Presidents. In only one case had the Vice Presi- dential position been permanently vacant. The office of Treasurer was omitted as it existed only in three locals. The remaining ninety-seven locals followed the option given to them in Section 9, Article 37 of the International U.A.W. Constitution: "At the discretion of the local union the offices of Financial Secretary and Treasurer may be Combined". 5O 51 According to the U.A.W. constitution, regular elections should be held every two years.1 In some cases we find spgcAa; elections--elections held on other than a regular election year. The special elections were held to ful- fill a constitutional provision stating that vacancies be promptly filled, with the exception of the Presidency, which is automatically filled by the Vice President. When vacancies occur within 60 days before the regular elec- tions, the executive board appoints temporary officers.2 fig! ggch turnover? The U.A.W. constitution requires locals to elect officers every two years. When.this provision was adop- ted, 80 percent of all U.A.U. locals decided to have elections in 1959 and 20 percent in 1960.3 Of the 100 locals in the sample, seventy-two are of the first group and had fix; elections during the period under study. The remaining twenty-eight held four elections. However, because the reports provide sufficient data, we can treat After analysing the position of Treasurer in the three 10- cals mentioned above, the office which most closely resem- bled that of Financial Secretary was considered the one under study. 10.1.11. Constitution 1968, Article 37, Section 11., p. 7A. All references to U.A.W. Constitution are to the 1968 Constitution unless otherwise specified. No major changes which might influence the analysis in this study have occurred since 1959. 2The Executive Board consists of the above four officers and any other executive officer of the local. Usually these include three trustees, Sergeant at arms, and the Guide. 3Personal communication‘with U.A.W. officials. 52 all 100 locals as having held five elections. Table V-l describes the distribution of locals by the number of officers. The means and accumulated numbers indicate differences in turnover among the four officers. Two groups can be identified--the President and the Vice President in one, and the Secretaries in the other. TABLE Vh1.--Distribution of the hundred locals by the num- ber of officers. No. of Mean No. Officers of offi- Office Total 1 2 3 A 5 6 7 0 cars President 31A 8 25 33 19 10 A 1 - 3.1A Accumulation 8 33 66 85 95 99 100 Vice President. 330 9 18 29 25 12 A 2 1 3.30 Accumulation 9 27 56 81 93 97 99 100 Recording Secretary 2A3 2A 36 21 13 A 2 2.A3 . Accumulation 2A 60 81 9A 98 100 Financial Secrets 216 36 29 23 8 3 l 2.16 Accumula ion 36 65 88 96 99 100 Total No. of 7- 9- 11- 13- 15- 20- Officers A 6 8 10 12 1A 16 23 ‘ _ ' Total No. Of Locals 1103 2 A 16 23 20 25 8 2 11.03 ‘ ‘ _.—._ 'This row of figures includes thirteen Vice Presidents elected after the previous Vice President moved automatically to the position of President. See Table V-8 for details. 53 Table V-2 presents the officers and the individuals who served. The total number of officers in a local is the number of people elected during the nine year period. If an individual served as President, than as Recording Secretary and then again.as President, he was counted three times for the turnover calculations but only once for the individuals' figures. If he succeeded himself as President, he was counted only once in both the turnover and individual calculations.1 The figures reveal that the presidency more than any other, is the office to which more individuals are re- elected. The lowest rates of reelection are for the Vice Presidency. Looking at the whole group, we find more than 10 percent of all officers (125 out of 1103) served in more than one office or in the same office more than once. we have compared the various rates of turnover among officers but we have yet to answer the question, "How much turnover?". The problem lies with the standard of compar- ison. Lacking another sample for comparison, we can com- pare our turnover rates to a theoretical model. This model consists of the calculations of the highest and the lowest possible numbers of officers and the demarkation line between "low" and "high" turnover. Having one hundred 1Although Table V-2 furnishes some information which can be useful, the closeness of these two turnover indexes, as the correlations suggest, show that the individual num- ber cannot be used for further meaningful calculations. 5A .moo.Vn ado .pweo on» ma moaoeoeamwo eem n .hoopm one no moaned one do encouumo no cornea on: eamsouoaomm mo noossm Hence one omenonoon menace some on eonsmau one .eoummo came on» ma she» rose non essence so Haas .ooaosoonmoo one nape» ea: «a .ome some memo once e>non moo Hmsoa>aoma o om. emu: con no.0 Hm mo.~ mom 4a.m 4am unconnonm o.namso«>«omH meoo m.em«q enooammo mdmsoubaomH one: m.mHmso one: oneoauho ooaumo decoy one amuse cap-x mo Hence no Hosea :«oaan mo .02 emeoamuo Hones Mo mode unease «0 * nodal mneoau mo .oz Heaps no modpeaomnoo tamauea the Ho demos Haves .ueoomnoa no meoem nu.~1> mqm exonv «a co m a o o o o euheuneenm nmeoduuo meeeolmneoauuo men-o mmeoauno memeo muooamuo meneo eoauho no no mo me we «o no meossz neonsz teams: mo seems: no no as: deems enouooeam a mouuomam a .ooCuo em... .3 anemone? Hegem- uo monogram-no; 3m: 63 PATTERNS 0;: RE;ELEC TIONS The material presented thus far has not revealed what happened to those elected in special elections. We can hypothesize that the majority of them will be re- elected when regular elections are held. The rationale is based on two points: (1) if one agrees to run in special elections, he is ready to serve, and will be ready to run again in the coming regular elections; (2) the in- cumbent officer has an advantage over new candidates. Table V-7 presents the data related to this ques- tion. About 35 percent of all officers are not re—elec- ted (36 out of 108). The highest figure is for Presi- dents-~50 percent of those elected in special elections, ‘when.for some reason the vice president did not move to the office automatically, were not reeelected (6 out of 12). The second highest is the vice presidency: 36 per- cent were not re-elected. Twenty-seven percent (7 out of 26) of the recording secretaries and 26 percent (6 out of 23) financial secretaries were not re-elected. The previous hypothesis concerning the incum- bents' advantages is hardly supported by these results. ‘we can only speculate that some candidates receive much greater support for special elections than for regular elections. It is clear, however, that this point needs (further investigation; and the given data is not suffi- cient to draw conclusions. 1)). (it! I"... i. 6A_ TABLE V-7.--What happened to those officers who were elec- ted in special elections, by office. Re-elected in the following Not a Officer ‘ Total elections re-elected Unknown All Four 108 68 36 President 12b 6 6 0 Vice President A? 28 17 2 Recording Secretary 26 19 7 0 Financial Secretary 23 15 6 2 A ______—_ .The special elections that took place in 1967 are included in this study; however, in these cases, the results were unknown at the time of data collection. bThese 12 cases represent situations when the Vice President did not move automatically to he presidency. As mentioned earlier, when the president is unable to function, the vice president moves automatically into the vacancy. Table V-8 shows what happens in these cases. It is interesting to note that while 50 percent of the presidents elected in special elections were not re4iected (Table V-7), 55.6 percent (Table V-8) of the Vice Presi- dents were not re--elected.1 Although this is not statis- ‘tically significant, it indicates a trend. The office 1We do not know however how many were not re-elected becauae they had decided not to run again. 65 TABLE V-8.--The results of the vice presidents' automatic movement to the presidency. Total auto- New Vice New Vice matic move- President President ments. Elected Not Elected —; A Vice Presidents who moved up and were re-elected in the following elections. 7 6 1 Vice Presidents who moved up and were not re-elected. 10 7 3 Unknown, 1 1 0 Total 18b 1A A 'See footnote in Table V-7. bThis total of 18 relates to 17 locals: i.e. in one local it happened the automatic movement occurred twice and neither was re-elected President. of the Vice President has two functions: (1) to assist the President in the performance of his duties; (2) to take over when the President is incapable of performing his dut- ies. ‘We found in our interviewlthat the assistance to the President is usually limited to replacing him when he is temporarily absent. The question therefore arises: to what extent do voters, in choosing one vice president over another, take into consideration the fact that his potential duty is to replace the president. ‘We do not have enough information to explore this question in depth, but the data suggests 66 that quite a few vice presidents, after moving to the presi- dency automatically, do not gain the confidence of the local members and are not re-elected.1 PATTERNS 05:;NNER MO§;L;IZ The term inner mobility will be used here to describe a situation when an individual has been elected to differ- ent offices, one after the other, or to the same office but not consecutively. Some indication about this phenomenon can be found in Table V-2, in the column of the differences between the total number of officers and the number of individuals occupying these offices. In this chapter we will discuss the differences between the total number of officers and the total number of individuals. In sixty-four locals out of 100 some inner mobility amourred: in twenty-nine of these locals, one individual 'was involved, usually holding two different offices during the nine year period; in twenty-three cases, two individuals ‘were involved; in eleven.cases, three individuals exper- ienced this mobility; and in one local, there were four individuals. . The patterns of movement are presented in Table V-9. Anong the four officers studied here, the vice presiden- tia1.office is the major source of new presidents:2 more 1Again'w'e assume that they ran for office. 2The numbers in the Table do not include automatic movement of the vice president. 67 TABLE V-9.--Inner mobility patterns.a __ —— _— Current Presi- Vice Recording Financial Office dent Pres. Secretary Secretary Total Presieus Office __i President(P.) - l 0 2 3 ~ (5) (6) . (3) (1L) Vice President (V.P.) 23 - l 2 26 (27) (2) (2) (31) Recording Secretary (R.S.) 6 5 - L 15 (13) (5) (h) (22) Financial Secretary (F.S.) h L 1 - 9 (h) (k) (1) (9) Total 33 10 2 8 53 (kt) (1h) (9) (9) (76) _* __ —_ 'The numbers in parentheses are the cases of total inner mobility, when the three and four steps of mobility are included. There are 11 cases of three steps of mobi ity, four of them are R.S.-P.-R.S., one of each of the following: ROSe-PQ-VOPO’ VQPO-P.-ROSO-. Vapo-Po-VOPO’ VOPO-‘utmtic PO- P.-V.P.-R.S., P.-R.S.-P., and one case of four steps, -V.P.- automatic P.-V.P.-P. than.60 percent of new presidents came from this source. The contributions from the offices of recording and financial secretaries was much lower (30 percent and less than 1 per- cent respectively). very often the term mobility means "upward mobility". “It is important for us to understand the significance of ‘the term as it relates to the local union. The eight presi- dents who were interviewed answered questions about the 68 relative importance of the various positions.1 Table V-lO contains the analysis of their answers. The re- sults suggest that the vice president and the financial secretary are equally important mainly because the an- swers of Local G's president deviate from the basic pat- tern. It seems that the ranking order of: President, Vice President, Financial Secretary, Recording Secretary, reflect their relative importance fairly well. TABLE V-10.--Analysis of presidents' answers about the relative importance of the four offices.‘ ‘A _ ‘_—__ - h; _ Presi- Vice Priii- Recor- Finan- __» Local dggt_g __dent __digg Sec, cial Sec. A h 3 2 1 High B h 3 l 2 Turn- C h 1 2 3 %L D 4 L.._1_ L_. Total 16 10 6 8 E k 3 1 2 Low F k 1 3 2 Turns 0 3 l 2 L ML 3 LL—Js 4~ -4—— Total 15 8 7 10 .Total Rating 31 18 13 18 __ —_ # 'Presidents were asked to grade each office in terms of its importance. A descending scale: A 3, 2, 1 was used with h representing the most important office. 1See Appendix C for a detailed description of the questionnaire. 69 On the basis of this order of importance, the fig- ures in Table V—9 were rearranged in Table V-ll so that the upper left side represents upward mobility while the lower right side represents the downward mobility. The results indicate that most (forty-six) mobility cases with- in the local are upward, as compared with only seven cases of downward mobility. When we add the three and four steps of mobility the upward cases are fifty-seven, and the down- *ward are nineteen. These results enable us to examine the framework suggested as an explanation of the higher turnover rates of president and vice president as compared to the recording and financial secretary. Assuming the existence of these two groups, we can hypothesize greater mobility ‘within each group then from one group to another. TABLE V%11.--Rearrangement of mobility patterns to examine upward and downward mobility.‘ ‘— *4— Current Presi- Iice Recording Financial Office dent Pres. Secretary Secretary Previous Office Recording 6 5 h - Secretary (13) (5) (t) Financial A L - 1 Secretlry (h) (h) (1) Vice 23 - 2 1 President (27) (2) (2) President - l 2 O- (5) (3) (6) 'Numbers in parentheses represent total cases of inner mobility, as in Table V-9. 70 Table V-12 tests this hypothesis which cannot be proven statistically. The reason for this can be that the technical skills required for the offices of Finan- cial Secretary and Recording Secretary are not inter- changeable, while, on the other hand, the possession of the same technical skills and the experience acquired by serving in these two positions can be of value when.mov- ing to the political-type positions of President and Vice President. TABLE Vr12.--Mobility between groups of offices.‘ Current Pres. and Recording Sec. Office Vice Pres. and Financial Sec. . Total Previous (Office President and 2h 5 29 Vice President (32) (13) (1.5) Recording Secretary 19 5 2L and Financial (26) (5) (31) Secretary_ 1, 10 53 ““1 453) 113) (76) 12 = .1106; P<.80; (x2) = 1.653).; (P)<.20. mobilit;hlumbers in parentheses include three and four steps SUMMARY The findings presented in this chapter should be used for two purposes. The first is to determine their relation to our theoretical framework. The second purpose is to analyze special elections and mobility in order to determine 71 their impact on turnover. Turnover was defined as leadership changes, regard- less of whether these occurred through regular elections, specialelections or automatic movement. Although special elections took place in fifty-eight out of one hundred locals, their impact on turnover in terms of their contri- bution to the number of changes is relatively small (less than 10 percent of the total changes occurred as a result of special elections). Eighteen changes were as results of automatic movement, and these plus the special election changes total 11.4 percent of the officers who came into power by other than regular elections. Furthermore, as the data shows, about 65 percent of those elected in special elections were re-elected in the regular elec- tions; therefore, the actual contribution of special elec- tions and automatic movement to the total turnover is even smaller (only thirty-six cases). The amount of mobility is not directly related to the amount of turnover because, according to the defini- tion used in this study, the same :rate of turnover can have different rates of mobility. However, we must con- sider mobility in our analysis, otherwise the relation- ships with which we are concerned will include an addi- tional dimension which we do not intend to study. Our analysis shows 125 cases of mobility. This number is not high enough to justify the modification of 72 our analysis of turnover because of it. We can conclude that although the mechanisms and organizational processes of special elections and mobil- ity are important, their impact on the rate of turnover in this study does not require modification of our treat- ment of turnover and its structural variables. CHAPTER VI ._§TRUCTURAL VARIABLES AND TURNOVER This chapter includes the testing of our hypotheses and concepts drawn in Chapter II. Each of the variables will be related to the turnover of the different offices, and then the relationship of each concept to the phenom- enon of turnover will be discussed. FORMALIZATION Table VI—l contains the correlations between the formalization variables and the turnover of different officers.1 Only the measures of membership size and rate of growth, as found and explained in Chapter IV, are pre- sented. Hypotheses F1, the older the local the lower the turnover, is supported by the findings.2 The negative 1The correlation to be used, unless otherwise specified, is Kendall Rank Order Correlation (Tau). 2The question of how high the correlation and sig- nificance should be, in order to support a hypotheses is a difficult one. In the textbooks, any correlation lower than .50 is termed a "weak correlation". However, any correlation, when its one-tail significance is better than the standard .05 level, is considered as an indication of relationship. Empirical studies of our type usually yield correlations of about -.30 to .30, and even those are rare. We suggest therefore to accept a correlation as supporting 73 7h TABLE VI-l.--Correlations of turnover and formalization variables.a Variables: ize of Rate of Growth of embership Age Membership TAU Rb Offic_rs ‘_ All four '015 -028 008 022 (.01) (.005)(.11) ”881d0nt ’e 11 -e 2“ .01. 003 ' (.05) (.005)(.29) Vice "’ e 03 "' e 22 e 11 a 20 President (.32) (.005)(.06) Recording -.10 -.O8 .10 .18 Secretary (.08) (.13) (.06) Financial Secretary -.1h -.10 -.01 .12 l (.02) (.07) (.16) __ ; __'__ _ .— 8The numbers in parentheses are the One Tail Significance Level. bSimple Correlations correlation indicates that the older the local, the less turnover it will experience. This is true when we con- sider the total number of officers, the President and the Vice President. Both correlation and significance drop substantially for the Recording and Financial Secretaries. These findings support our previous suggestion of the division of the local leadership into two groups. We will see in the forthcoming discussion that various ‘ a hypothesis when it is higher than .10 and its one-tail significance is .05 or better. This guideline is subjec- ted to the same reasoning stated on page 40, footnote 3. 75 variables influence the various offices differently.l As none of the theoretical frameworks used in this study differentiate among different types of lea- ders, explanations of such differences will be post hoc. Age is considered a major factor in the creation of organizational and behavioral patterns. As the organi- zation grows older, many of its facets are institution- alized and, thus, the leaders can avoid unnecessary con- flicts and can engage in activities which will help them to be re-elected. Following this reasoning, we can ex- pect age to influence turnover. However, the technical leaders whose patterns of activities are established at the organization's creation, are less influenced by age, as compared to the political leaders. We classified the Financial Secretary and the Recording Secretary as technical officers in our dis- cussion in Chapter V, and the lower influence in their case is therefore understandable. Hypothesis F2: the bigger the local, the lower the turnover, is statistically supported and we conclude that with the exception of the Vice Presidency, the lar- ger locals have lower turnover. The office of the 1Correlation, as a rule, does not imply causal rela- tionship. However, under certain conditions, when it is possible to assume only Qne diregtion of relationships, as in the case of age and turnover, we will use the term influence, In other cases, when the relation cannot be interpreted as being one-way, we will use the term relate. 76 Financial Secretary seems to be more clearly related to the variable of size then the other three offices: the job is directly related to the "amount of money" involved which, being almost equivalent to the income of the local, is directly related to the size of the membership as dis- cussed in Chapter IV. As the local grows, the functions of the Financial Secretary become more and more complex and require higher technical skills. These skills are rare among local members, and the one and natural way to acquire them is by experience. Therefore, the longer a Financial Secretary stays in office the better will be his performance. The voting member, well aware of this fact, tends to re-elect him. It was also established (in Chapter V) that the stability of the Financial Sec- retary is the highest among the four offices. Hypothesis F3 suggests direct relationship between growth rate and turnover. Membership growth infers the influx of new members causing instability and new prob— lems, and, in these cases we assume a higher leadership turnover. Acceptance or rejection of this hypothesis depends on the statistics used.1 The non—parametric cor- relation suggests a low relation for the Vice President and the Recording Secretary and no relation for the others. With the exception of the President, the parametric statis- tics support the hypothesis. 1When only Tau correlation is presented, it implies thatlthe other statistical methods yield about the same resu ts. 77 The local presidents interviewed in this study all argued strongly that the growth rate is related to the turnover rate. We therefore accept the results of the parametric statistics and the hypothesis that with the exception of the President, higher officer turnover is related to higher rates of growth. Based on the socio-psychological approach we can speculate that presis dential turnover is not related to the growth rate while turnover rates of the other offices is related to it. The influx of new peOple introduces new ideas and atti- tudes into the local, but for these new peOple, the president represents the union and the benefits for which they joined. So, although they attempt to bring about changes, they still want the president to remain in power. As we have shown, age, size, and growth with some exceptions seem to be significant factors when related to turnover. However, the analysis thus far has yielded only the direct relationship between each separate variable and turnover. In other words, it has shown how good each variable is as a predictor of turnover. It is important to know how the concept of formali- 1 zation is related to turnover, and how much of its 1The importance of this analysis can be illustra- ted by studying Grusky's theory ("Corporate Size") and its criticism by G. Gordon and 3. Becker, "Organizational Size and Organizational Succession-~A Re-Examination," American Journal of §901010gY, LXX (196A), 212-223. Grusky tested His Hypotheses concerning the relationship between size and succession by using X2 on tables containing informa- tion about organizational size and succession. A table 78 variability can be accounted for by all the variables together.1 In this study, a variety of Least Squares2 pro- grams was used to find these relationships. The purpose of these calculations was to assess the contribution of all the variables and their relative importance accor- ding to the statistics of Multiple Correlation. Our study contains fourteen variables which can be used in the Least Square programs.3 Table VI-2 contains the in- formation about the relationship between turnover and these variables. was constructed for the positions of Board Chairman, Pres- ident, Treasurer, Comptroller and Secretary. The signi- ficance levels of his Ofindings corresponding to these positions are: W<§ P<:. 25; P<:.15; P<:’35; P<:.05. His second sample oes not suggest much better results, and all of them are far below any standard level of acceptance. He argues, however, that it was the multiple comparison and the replication which provided a basis of confidence in the existence of a relationship between size and succession (1961, p. 263, footnote 9), but does not explain how it was practically done. Gordon and Becker, in a replication of Grusky's study using Multiple Correlation Anal sis, prove that the size factor (both sales and employees) explains less than 1 percent of the phenomenon of succession. 1A close estimate of the contribution of each vari- able to tge variability of turnover can be obtained by the R2 or Tau Only correlations about .32 yield an R2 which accounts for more than 10 percent of the variability. None of our findings is that high. 2See statistical and computer reference section. 3These fourteen variables are: Membership, Income, Age, Membership Growth Rate, Income Growth Rate, Number of Units in the Local, Total Payments to Employees and Offi- cers not under study, Total Payments to the Four Officers under study, Payments to each Officer, Community Index, and 79 TABLE VI-2.--0verall statistics--turnover and the fourteen structural variables. Statistics Multiple Coefficient Level of Correlation of Signifi- Coefficienta Determination cance foicers__ _g R R2 P All Four .58 .3h .001 President .hl .17 .253 Vice President ‘ .57 .33 .001 Recording ( Secretary .A2 .18 .231 . Financial Secretary .h3 .18 .193 —*;_ — E 8The Multiple Correlation Coefficient does not indi- cate Cthgirection of the relationship, as it is computated R = . Column 2 in the Table--R2 or the Coefficient of Determination1--reveals that these fourteen independent variables account for 17 percent to 3b percent of the ; Total Resources of the Local. For a detailed description of the variables see Appendix B. ‘When considering the total effect of the independent variables of this study, 'we should remember that there are four other variables which are not included in the Least Squares calculations: Type of the Local, Facilities, Newspaper, Company. 1R2--the coefficient of determination can be de- fined as "the proportion of the sum of the Square devia- tion from the mean of the dependent variable accounted for by the independent variable". Michigan State Univer- sity-gCalculations of Least Squares. Description No. 7, pa "' ,e 8O variability of turnover, depending on the office. We will see however, that a substantial portion of the variability can be explained by fewer variables, because then relative contributions differ greatly. Using the same statistical method, but considering 1 we find the results only five variables of formalization, presented in Table VI-B. This table reveals that 9 per- cent to 17 percent of the variation of turnover for TABLE VI-3.--Multiple correlation of formalization vari- ables and turnover.8 __—_ — A — _ -_.—; _ Statistics Multiple Coefficient Correlation of Significance Coefficient Determination Level Officers R R2 P All Four .h2 .17 .005 President .31 .10 .08 Vice President .35 .12 .03 Recording Secretary .30 .09 .11 Financial Secretary —1 .20 .04 .56 aThe formalization variables in the equation include size (membership and income); age; rates of growth (membership and income). 1Income size and rate of growth are included in this calculation although omitted elsewhere because a variable, unless its correlation with other variables is 1.00, still contributes something to the explanation, and therefore should be included in calculations. 81 President, Vice President and Recording Secretary can be explained by variables of formalization, when grouped together. The percentage drops to A percent in the case of the Financial Secretary. The theoretical framework of this study implied that the concept of formalization would account for a substantial portion of turnover. We expected a higher contribution than we found. The question is whether we can identify a factor which might decrease the potential contribution of the concept. It seems that such a fac- tor, which was found relevant in the case of the U.A.W. and was labeled "instant bureaucracy" could also be rele- vant to other organizations with the same patterns of relations between headquarters and the sub-units. This pattern of relations is directed toward fac- ilitating the adjustment of new locals to their organi- zational environment, and enabling their leaders to per- form their functions efficiently. The International's mechanisms for implementation of this instant bureau- cracy are: 1. The International Constitution.which includes detail- ed specifications for newly established local as to structure, leadership, financial structure, and rela- tions with other organizations. 2. The requirement that the local write its own bylaws, or adopt the relevant chapter from the Constitution, 82 and follow the "master bylaws" pattern.1 A. A close follow-up of all the local's activities by a representative of the International. These mechanisms,and the intensity of their appli- cation influence the development of phenomenon such as stabilization of behavioral patterns, establishment of standard operating procedures, division of labor, rou- tinization of communication and allocation of responsi- bilities and resources. The impact of these mechanisms minimizes the influence of variables such as size and age on the development of the same phenomena. This is a manifestation of the formalization concept, as al- ready discussed in the theoretical chapter. In other words, we can argue that in organiza- tions where the relations between headquarters and sub- units are not subject to the mechanisms of "instant bur- eaucracy," size, age and other variables of formaliza- tion will account for a high proportion of turnover. Most facets of instant bureaucracy are related to the technical aspects of leading the local union. This accounts for the even smaller contribution of formaliza- tion to the explanation of the turnover of the technical officers--the Recording Secretary and the Financial Secre- tary. “ 1The bylaws of the eight locals which were inter- viewed were closely scrutinized, and a "master example" is recognizable in them. 83 We can now conclude our discussion of the variables and concept of formalization. With few exceptions, all three hypotheses, F1, F2, and F3, are supported by the data. Examination of the variables' contribution to our understanding of turnover rate differences reveals that, although each variable by itself can account for only a small portion of the variation, all the five variables together account for A percent to 17 percent of the vari- ation. COMPLEXIT1_ Hypothesis Cl states that as the number of units in the local increases, turnover decreases. This is suppor- ted by Column 1 of Table VI-hl, except for the Recording Secretary, as well as by the results presented in Table VI-5.2 Based on these results, we can accept hypothesis CI. The hypothesis is derived from Michels' theory which suggests that when organizational complexity _—‘ 1The variables of complexity will be analyzed by correlation and multiple correlation. Some variables which cannot be analyzed by these statistics, because they are not in a continuous order, will be analyzed by One way Analysis of Variance. 2The "number of units" includes the dichotomy between single-unit and amalgamated-locals, but contains more infor- mation about the number of units. However, the correlation calculations might be biased by the distribution of locals by units: 72 are single-unit locals; l6 locals have 2 units; 2 locals have 3, 2 have 5.2 have 9, and 6 locals have from 10 to 99 units. 8h increases the stability of its leadersincreases too. A specific explanation of this complexity variable seems to emerge from the nature of the political struggle in the amalgamated local as compared to the single-unit local. The politics in the amalgamated local takes place at two different levels-~the unit level, where each unit has its own leadership and the local level where the leadership of the entire local is elected. When we examined the intensity of the political struggle in the amalgamated local, we found the unit level struggle much stronger and more meaningful to the members than the struggle at the amalgamated local level.1 One indication of this phenomenon can be found in election results. Our interviews and analysis of election results show that the total votes cast for officers of the units is higher than the total votes cast for officers of the local union. The number of votes can serve as an index of apathy, and we can conclude that members are more apa- thetic toward election of officers in the amalgamated lor cal than toward the election of officers in their immed- iate unit. According to Michels, apathy is related to turnover and we should expect to find lower turnover where there is higher apathy: and this is our finding-~lower turnover in the amalgamated local. __ —— 1Interviews and analysis of election date in about 5 amalgamated locals were used in this examination. 85 TABLE VI-A.--Correlations of turngver and complexity variables. variables No. of units Total Total Payments in the Local Resources to Employees Officers ____g_g .__ __ 1. All Four -.21 -.17 -.12 (.005) (.005) (.Oh) PreSident -028 “011+ .e15 (.005) (.02) (.01) Vice -.16 -.08 -.02 Pr951dent (e01) (013) (oi-#0) Recording “.02 -006 'eOl} Secretary (.37) (.18) (.26) Financial -.15 -.13 -.08 Secretary (.01) (.03) (.12) 8For an explanation of this variable see foot- note 2 page 83. The numbers in parentheses are the sig- nificance level of the Tau correlations coefficient. Other factors contributing to the stability of the amalgamated local leadership are the advantages of the incumbent officer, which are much greater in the case of the amalgamated. Because of the geographical diSpersion of the units, the incumbent officer is fully exposed to the members, while this exposure is very costly, and sometimes impossible for a new candidate.1 Hypothesis CZ deals with the resources of the local. 1This discussion excludes the Recording Secretary, as the type of local does not influence him. 86 TABLE VI-5.--Average number of officers and type of local. “ ‘ Type of Single Unit Amalgamated Level of Signi- Local Local Local ficance Pa Officezs ___ __ ___ All Four 11.5 9.8 .01 President 3.3 2.6 .01 Vice President 3.A 2.9 .10 Recording Secretary 2.A 2.h .85 - Financial Secretary 2.3 1.9 .10 _ ‘_ “A. h T" 8These are results of parametric one way analysis of variance. The three variables are: total resources, facilities, a newspaper. Table VI-h, Table VI-6 and Table VI-7 present the findings. The total resources variable is negatively correlated with turnover, which means that more resources are related with less turnover, when dealing with all four officers, the president, and the vice-president. Facilities (Table VI-6) have no relationship whatsoever to turnover. With the exception of the vice president, newspapers seem to be related to turnover (Table VI-7) and their publica- tion is related to smaller turnover. The findings related to hypothesis C2 are in accor- dance with the analysis and explanation suggested by Lipset (et a1) and discussed in our theoretical framework: the 87 TABLE VI-6.--Average number of officers and the existence of facilities.8 With Without Level of Facilities Facilities Significance P Officers _g _l___ All Four 10.7 11.3 .46 President 2.9 3.3 .5h Vice President 3.h 3.3 .53 Recording Secretary 3.h 2.5 .93 Financial . Secretary 1 2.0 2.3 .A1 .__-_._ 8A third group of 6 locals acquired facilities dur- ing the period of the study, and were not included in this analysis. TABLE VI-7.--Average number of officers and the existence of a newspaper. ‘ With Without Level of Newspaper Newspaper Significance P Officers _____ _i 4_ All Four : 10.4 11.5 .07 President 2.9 3.3 .16 Vice President l 3.h 3.2 .A7 Recording Secretary 3.1 2.6 .OA Financial Secretary 1.9 2.A .OA 88 control of resources and communication networks makes it easier for the incumbent officer to stay in power and more difficult for a new candidate to enter and win the race. The exception of the facilities variable is a re- sult of its definition in this study. The definition ignored the differences of value of the facilities--and amount of control the local has over them. We assume that a definition and analysis which will include these aspects will reveal a relationship between facilities and turn- over. The hypothesis dealing with the relationship be- tween resources and turnover includes three variables, and as we do not have statistical methods to weigh them, its acceptance is based on our earlier discussion. Hypothesis C3 states that higher payments to the employees of the local are related to lower local leader- ship turnover. The test results of this hypotheses are presented in Table VI—A column 3 and except for the presi- dent and total officers do not support the hypothesis. The explanation may lie in the distribution of the vari- ables as shown in Table IV-7. The analysis of the complexity variables, when grouped together, by multiple correlation, is presented in Table VI-8. As this table suggests, the contribution of the complexity concept to the statistical explanation 89 TABLE VI-8.--Multiple correlation of complexity variables and turnover.8 Statistics Multiple Coefficient Significance Correlation of Level Coefficient Determination Officers R __ R2 R__ ,_ All Four .26 .07 .07 President .25 .06 .10 Vice President .22 .05 .20 Recording Secretary .15 .02 .52 Financial Secretary .12 .02 .68 8The complexity variables in the equation include: number of units; total payments to employees; total resources of the local. The other two variables, facili- ties and newspaper, are not included because they are not in a continuous order and could not be analyzed in multiple correlation. of turnover is lower than that of formalization.l As in the case of formalization, we expected a higher contribution of the concept than was actually found. The reason for the lower contribution seems to be the ex- clusion of two basic measures of complexity from our analy- sis. The first one is the demographic complexity. A care- ful comparison of the eight interviewed locals shows very 1The addition of the contribution of the facilities and newspaper, which are not included in the calculations because of statistical reasons, does not seem to change the results substantially. 9O few items which differentiate between locals with high turnover from those with low turnover. However, one clear factor does emerge in all four locals with high turn- over: the heterogeneity of its members. In two locals, it is the age distribution of the members,1 another local has a majority of female workers, and the fourth has a tremendous influx of high school stu- dents into the local during the summer months. The second measure, the number of issues faced by the local, is highly related to the measure of demogra- phic complexity. The existence of many issues implies the high degree of segmentation. The existence of a higher number of issues in an organization will increase the probability of the existence of many different groups within it. We now can conclude our discussion of the variables and concept of complexity. Two of the hypotheses, C1 and 02 were accepted, and 63 was rejected. We found that the contribution of the concept to the statistical explana- tion of turnover is substantially lower than that of for- malization, due to the exclusion of two important measures, and, again the offices of the Recording Secretary and the Financial Secretary are much less influenced by this concept. 1The importance of members' ages was recognized early in the study, and Professor Daniel Kruger strongly suggested its inclusion. However, it was impossible to collect data relating to this subject. 0n the same sub- ject see the washi ton Post, July 6, 1969, "Apathy Splits Labor's Ranks". 91 ,ATIBAQIIVENESS Hypothesis A1 states that higher payments to offi- cers are related to lower turnover. The findings presen- ted in Table VI-9 support the hypothesis. However, there is some information which should be introduced and weigh- ed before evaluating the variables and concept of attrac- tiveness. T First, the statistical analysis is subject to the limitation of the distribution of these variables (see Table IV-8.) Second, our interviews with locals' presidents reveal two basic Opinions. All agreed that the material compensations of leadership, whether for a full time job or for "lost-time", is in no way proportionate to the amount of time and effort invested. Table IV-8 clearly indicates the level of compensation to local leaders, and supports the above opinion. The interviewers also emphasized that the real attractiveness of leadership offices rests in its non- materialistic compensation. Our study is directed toward "explaining" organi- zational turnover through the use of variables which is a limiting factor. This limitation is more clearly empha— sized in the case of attractiveness. The partiality of materialistic compensation as a measure of attractiveness has already been discussed. Statistically, as indicated 92 1 2 of Table VI-9 its contribution to the ex- in Column Tau planation of turnover is very low. Although the hypothesis Al was supported statis- tically, the previous discussion shows that the concept of attractiveness needs the modification and elaboration which measures of non-materialistic compensation add. TABLE VI-9.--Turnover and payments to officers. Payments to Officers Tau Level of 2 Offiggrs Cpgrelation _§ignificancggP Tau All Four "519 e005 e015: President -.21 .005 .04 Vice President -.11 .05 .01 Recording Secretary -.12 .04 .01 Financial Secretary -.18 .005 .03 EXTERNAL ELEMENTS The influence of environment or external elements, will be examined on three different levels. The first is 1Multiple correlation cannot be used in this case, as we assume no relation between the payment to a parti- cular officer and the turnover of others. This assumption is supported by the pattern of mobility discussed in Chapter V. We will therefore use Tau2 which is the coefficient of determination and which will show the contribution of a specific officer's payment to the turnover rate of the same 0 ficer. 93 the immediate environment, the company with which the local has a labor agreement. Hypothesis E1 states that the com- pany is related to turnover and the findings are presented by using two grouping systems in Table VI-lO. The first grouping and analysis consists of the Big Three, when com- pared with all others.1 The second is all locals other than the Big Three when compared separately to each of the Big Three, and the Big Three compared among themselves. As Table VI-lO indicates, no significant relations whatso- ever exist between the company and turnover and, therefore, we reject trs hypothesis. Hypothesis E2 states the relations between the second level of the environment (the community) and turn- over: the smaller the community factor, the lower will be the turnover. In our study, this factor is measured by the community index, which is the community population ad- justed by the number of blue collar workers and the size of the local. Table VI-ll presents the correlates of the community index and the turnover of officers. The find- ings do not support our hypothesis; therefore, we reject hypothesis E2. Hypothesis E3(a) states the relations between rates of turnover and different years (the time dimension). The results are presented in Table VI-12 and VI-13. The #5 1For the rationale of this grouping, see footnote 1 Po - 9A a TABLE VI-lO.--Average number of officers by the company. Company Others All Big P General Chrysler Ford P Three Motors Officers g__ All Four 11.2 11.0 .72 12.2 9.9 10.3 .32 President 3.2 3.1 .88 3.3 3.2 1.7 .2A Vice President 3.2 3.6 .15 A.0 3.1 3.3 .19 Recording - Secretary 2.A 2.A .93 2.6 2.1 2.7 .80 Financial Secretary 2.2 2.1 (.61 2.3 1.5 2.7 .2A 8One Way Analysis of Variance (parametric). TABLE VI-11.--Correlations of turnover and community index. Level of Tau Significance Officers Correlation P All Four (-).03 .31 President .02 .36 Vice President (-).0A .29 Secording Secretary (-).07 .15 Financial Secretary (-).06 .18 95 TABLE VI-12.--Numbers of new elected officers by the year: I (Regular Elections Only). Total 1961 1963 1965 1967 Qificers .1 ___ __ All Four ASA 105 117 11k 118 President 146 33 AA 35 3A Vice President 1A7 31 36' A5 35 Recording Secretary 89 20 20 17 32 Financial Secretary 72 21 17 17 17 1(2) = 10.541 Df. = 9 P <:.50 TABLE VI-13.--Number of new elected officers by the year: II (Regular Elections Only). (Total 1960 1962 1964 1966 Officers ___ All Four 113 « 27 32 24 30 President 36 10 7 10 9 Vice President I 33 . 5 10 8 10 Recording Secretary 23 5 10 3 5 Financial Secretary 21 7 5 3 6 12 = 7.865 Df. = 9 P<<:.70 96 assumption is that changes in the environment are reflec- ted in the flow of time. For example, changes in unem- ployment rates are related to time and place. Based on this assumption, we calculated the tables so that they represent not only all the officers who were replaced in each regular election, but also, each officer separately.1 The results of both tables indicate that no rela- tionship exists, and we therefore reject hypothesis E3(a). Hypothesis E3(b) states the relationship between negotiation years and rates of turnover. Table VI-lh examines these relations by comparing the number of turn- over cases to the number of non-turnover cases in either the contract expiration year or one year later.2 This hypothesis, too, is rejected. Although the statistical findings reject the hypo- theses, we do not feel that the original assumption con- cerning the relationship between external elements--envi- ronment--and turnover should be disregarded. The fact which seems to account for these findings is the sample itself: U.A.W. locals in Michiga . We have already dis- cussed some of the characteristics of the relationships 1In this case we had to treat locals with four or five elections separately, and therefore we constructed two tables. Special elections were also calculated sep- arately (they are not shown here) and their P's are to ' of Net anal and I tore t os‘_ Local U-io« ’1965). twas o .. ter, in ’o', that t ee o to 0’01ocals in the sample ceased to exist. 2See U.A.W. constitution, 1959, page 17 article 8 Section 7. This provision.haa not been changed since then. 118 mm «a 4.0m com on or mom can use Hana» o a con o o.oo a u u.« o +Onoon +Hoa .o wooed o m 8H an m :3 o m 4 .3” :0QO 00.73 . m Q3“ o no e.am on em on so o.o« on gonad on-««.4 are." a o n.nm m." 5.3 «H mm Hz: an 30mm 2..» .m mom m «N «.mn um and 1:” on «Sm no :03 «IN .N e an e.an an atom «a no o.e~ on .meae . was me one» mo.oz sauna ameonem Hench cam no ems» . smoonem 01:: .n .H “an: we ewes «6.02 .neo we 335m umeonem Haves use: .oz .5.»an: 5 383 54.3 82.322 l1 ode-em one moausdaomunéué H33. (IJAV “\\H1u 119 TABLE A-2.--Comparison of the 8 interviewed locals. Vsriable Average No. Average Average Number of of Officers Age Sise Amalgamated Member- Lanals Group ship High Turnover 18.75 26.2 230 0 (A) Low Turnover 5 4 29.2 1350 2 (A) Total 11.8 27.7 790 2 SuozSaoplo for the ioolyois of Nogotiotiooo A study of the relationship between turnover and the time of negotiations was conducted on.a sample of twenty-nine locals. These locals were selected only be- cause of the availability of their labor agreements during the nine years under study. The distribution of this sub- sample was found to be very close to the distribution of the main sample in one important respect-dwhile 29 percent of the locals in the main sample are Big Three locals, 27.A percent of the sub-sample are Big Three locals. APEENDII B DATA COLLECTION: SOURCES AND VARIABLES This appendix is a technical description of the variables and how and where the information on each of them'was obtained. SOURCES OF iNFORMAIIO! AND VARLQBLES The following code letters will be used in this description: 3 - CP- information obtained from "Labor Management Reporting and Disclosure Act of 1959." According to the law, each local union (with some exceptions which do not affect this study) must file an annual report. The informa- tion in the report includes financial data, notifi- cation of bylaws, and changes of officers. Information obtained from the U.A.W. - Conventions Proceedings. Information obtained from interviews with U.A.W; officials. U.A.W. - Information obtained through special compilation for this study, in the U.A.W. Headquarters by dif- ferent departments. 120 121 Typo of Locai A dichotomic variable - single unit local or anal- gamated local, a local which includes more than one unit. Source of information: CP, U.A.W. W The existence of the union hall and/or offices; The information*was obtained from.R and was organised in three categories: ‘ 1. The local owns a building. 2. The local does not own a building. 3. A building was either built or bought during the period of study. Note that the value of the facilities is not con- sidered here. This aspect is included in the variable total resources of the local. Imam The existence of a local newspaper. This infor- mation was obtained from the Labor Archives, Mayne State University. QQIIIEI The company with which the local has a labor agree- ment. 1. Chrysler. 2. Ford. 3. GM. A. Others and amalgamated locals. If a majority of locals in an amalgamated local are 1,.Er 0’w‘.’§ , BI 122 in the same company, this local was counted as one of this company. The four categories above were combined together for some calculation: 1. Big Three companies. 2. Others. Sources of information - CP and U.A.W. W1 Membership figures were based on the per capita payments of the local to the Internationa. (using an average of the nine years under study: AV). Source - R. nceme The total income of the local (AV.). Source of ififflntion " Re _—_- ‘— 1The exact meaning of membership sise is still quea- tiemable, even after the passing of the Reporting and Disclosure Act. A review of the available sources and their reliability, where international unions are concerned, can be found ianarcus (1962). However, this review does not include the annual reports as a source. Estey, in his article sounds very pessimistic: "BLMR (the agency which was responsible for these reports prior to its merger in 1963 with the office of welfare and Pension Plans) data on unions structure is inaccurate, incomplete and internally inconsistent." Later -however in 19 6, Ferguson and Stober suggested that the LMRDA data is the best available, better than using direct union figures and better than convention information. The data is available faster and suffers from less bias than other sources. They suggest some techniques to overcome most of the common deficiencies of the data. ,‘IIIY Aelnllt‘ils.’ 123 Ass The age of the local beginning with the date in which it was first chartered by the U.A.W. International. Information was supplied by the U.A.W. except in three cases where the direct inquiry in the locals yielded the information. M. rs , te f Growt Membership of Union X in 1967/ Membership of Union X in 1959 All unions' membership—in_l967/ All unions' membership in 1959. Rate of Growth r The formula was suggested to me by Professor Morcus. For detailed explanation see Marcus (1962). P. 62. Source of data for calculations is R. iocemo Rage of Gyogop The same as above but with income figures instead of membership. Hail The number of units in each amalgamated local. Single unit locals were given the value of one for various oaloulationa.’ Source of information or, U.A.W. iggygonts to employees and officers other than the i Hgsr 1'.qu The total amount of money, salaries and other pay- ments made by the local during the year to all employees and officers such as chairman of committees and recreational 12h directors. The figure used for calculations was the storage for the nine year period. Source of informa- tion - R. The major weakness of the total payment as an tindicator of complexity is the fact that in many locals, especially small ones, much of the work is on.a volun- tary basis. Eoyooots to Officers Total amount of money, salaries and other payments that the local paid to the four officers under study. In cases of’more than one Vice President.only the amount paid to the first one was included (AV). Source of in- formation - R. limi- '. 1.; Pr“.~ -tVe’ 's ~ , Recon As above, for each of the officers separately, (AV). Source - R. Wits The index is a one figure description of the rela- tionship between the sise of the local, the total popu- latien of its city or town, and the percentage of blue collar employees in the city or town. The Index Number -* YIST‘E}E" X 100 R AM - The sise of the local in terms of membership (as presented above). 125 P - Population of the city or town, based on the 1960 census. R - "Non-worker-worker ratio" is the ratio of the persons not in the labor force to those in the labor force in the same location as above, plus one. B - Percentage of blue collar workers in this location. In few cases (six) where the population was smal- ler.than.1000, R and B were not available. In these cases an.average was calculated between the highest and lowest possible index. The highest index number indicates the largest sise local as related to the sise of the population and the blue collar population in this community. Source of information: Census information for 1960, and R. Togo; Rosooyces of tho iocoi This figure represents the total assets of the local--cash, securities, fixed assets, (AV). Source- R. v of Off Five measures of turnover are used herein: (1) the total number of officers; (2) the presidents§'(3) the vice-presidents; (A) the recording secretaries; (5) the financial secretaries. The measurement is the actual number of officers during the period of the study. .l V’ .l ElishIwo-l ll 1. 2. 3. A. 5. 6. 7. 8. 9. 10. APPENDIX C INTERVIEW GUIDELINES Number, name, type of local. 'With which company do you have an.agreement? When was local established? D¢ you have bylaws? How many members do you.have now? How many did you have in 1960? For how long have you been serving as officer? Do you intend to run again? To the same position? Why? Why did you run in the first place? Were you the only candidate? Do you have a caucus in the local? Who opposed you? ‘Why? Who supported you? Why? How do you like this office? What are the rewards (advantages) of elected officers? (Laterask especially about materialistic compensation, seniority rights, prestige) What are the "costs" (disadvantages) of being an offi- cer . Why do other people try to be elected? Do you usually have enough candidates for the various offices you have to fill? What happened to previous officers, during, for instance, the last ten years? (try to count all of them). Where are they now? Are they active members of the local? Do they have influence on what goes on in the local? People say, though I do not agree with it, that the local officer is no more than a senior clerk executing the constitution of the U.A.W. What do you think? Can you describe your responsibilities, tasks and duties? 126 11. 12. 13. 1h. 15. 16. 17. 18. 19. 20. 21. 127 Can you do the same for the other officers in the local? Can you arrange the four officers--President, Vice President, Recording Secretary and Financial Sec- retary according to their importance in the local? According to the prestige of the office? Do you use outside consultant (not U. A.‘W. people) for your activities? Do u think that the functions of the office are inf uenced b the person who holds it? (count the four offices) What, in your opinion, are the factors which influe ence turnover of officers in the local? It seems that turnover at the International level is different than in the Local. Can you explain it? Do you think that there are many differences in turnover between the four offices in the local? Where do you think you will find mere change? (count the four offices). How can you explain that? In.your opinion,. what are the chances of local officers to gate job in the International? What Job do you have in mind? What can the officer do in order to get it? Do you think that the regional director and/or the international representative influence changes of leadership in the local? How can they do it? In your opinion, do the following factors influence turnover of officers-~generally? In your local? a. Age of the local ‘ b. Type of the local- single unit or amalgamated. Sise of membership d. Sise of income. e. A slower or faster rate of growth. f. The company with which the local has a labor agreement. g. The type and sise of the community. h. The percentage of male and female members. Are you a member and/or officer in other organisa- tions? I I p'\ 93 "'il’iimfflfiufiflm‘nfin[gliflujffliifihlmmflifififs