STATUS CRYSTALLIZATION A1~ RIGHT­ I'lING EXTREMIST ATTITUDES by GARY BERNAHD RUSH A THESIS Presented to the Department of Sociology and the Graduate School of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy June, 1965 VITA Name of Author: Gary Bernard Rush Place of Birth: Vancouver, British Columbia, Canada Date of Birth: March 27, 1931 Undergraduate Training: University of British Columbia, 1955-1959 Graduate Training: University of Oregon, 1959-1963 Degrees Awarded: Bachelor of Arts, University of British Columbia, October 30, 1959 Grants and Awards: --­ Jonathan Rogers Award, University of British Columbia, 1958-1959 United States Department of Health, Education and Welfare, National Institute of Mental Health pre­ doctoral fellowship grant, 1961-1962 and 1962-1963 Areas of Special Interest: Political Sociology, Value and Belief -- Systems, Social Change Papers Delivered: "Toward a Definition of the Extreme Right," present­ ed at the annual meeting of the Pacific Sociological Association, Portland, Oregon, April 25-27, 1963. tiThe Politics of the Po,\,rerless, II presented at the annual meeting of the Pacific Sociological Assoc­ iation, San Diego, California, March 5-7, 1964. Publications: IIToward a Definition of the Extreme Right," Pacific Sociological Review, 6 (Fall, 1963), pp. 64-73 . ...ll.....­ ," Ul (I) •.-1 ~ ~ ·rl \ •.-1 ~ (}) , ,C t\ ~ " 'i-i ~ (I) Ul '.-1 ~ ::­ '0 ~ : '0 .r; S '0 (j) .c:: III 'r; rl 6 §~ 0 H CJ CJ ~ ell 0 H J::: A ell ~ 0 III rl C\j +'H (j) i=Q 0 E-t TABLE OF CONTENTS Page LIST OF TABLES yiii LIST OF FIGURES x Chapter I . INTRODUCTION 1 Historical Perspective 1 Sociological Perspective 5 Independent and Dependent Variables of the Present Study 7 II. THE STATUS CRYSTALLIZATION MODEL 13 The Concept of Status Crystallization 13 The Status Crystallization Model Used Herein 21 III. TOI-IARD A DEFINITION OF THE EXTREME RIGHT 38 Current Definitions, Explicit or Implicit 38 Toward a Definition of the Extreme Right 50 IV. EMPIRICAL ANALYSIS OF EXTREME RIGHT ATTITtJDES 66 The Cluster Analysis Technique 66 The Cluster Analysis: Hypotheses and Findings 70 The Right-Wing Extremism Score 87 V. STATUS CRYSTALLIZATION AND RIGHT-WING EXTREMISM: DEVELOFMENT OF THE HYPOTHESIS 92 Problems Engendered by Low Status Crystallization 92 The Psychological and Social Consequences of Uncrystallized Status 97 Status Crystallization and Right-Wing Extremism 102 vi L1..JJI~7/ ~ .n_.oU''';07::X='==::~~~~~~~~~~~!!!!!! Chapter Page VI. STATUS CRYSTALLIZATION AND RIGHT-WING EXTREMISM: ANALYSIS OF DATA AND FI1~INGS 107 The Sample 107 Status and Crystallization 112 The Role of Elaboration in Survey Research 117 Testing the Hypothesis 121 VII. SUMMARY, DISCUSSION AND CONCLUSIONS 128 Summary 128 Discussion 131 Conclusions 135 Implications for Future Research 137 BIBLIOGRAPHY 140 Books 140 Journal and Periodical Articles 1!~7 Unpublished Manuscripts 152 APPENDIX A: THE SAMPLE 153 APPENDIX B: CLUSTER ANALYSIS DATA 157 vii __._-­-_ .. LIST OF TABLES Table Page 1. Original Occupational Categories, With Typical Occupations Represented . . . . . . . . .. 24 2. Coding Categories for Source of Income . . . . .. 25 3. Revised Occupational Classification . . . . . .. 29 4. Frequency Distribution of Respondents by Occupation, and Assigned Percentile Scores for Each Occupation Range 31 5. Frequency Distribution of Respondents by Income, and Assigned Percentile Scores for Each Income Range .. 32 6. Frequency Distribution of Respondents by Education, and Assigned Percentile Scores for Each Education Range. 33 7. Frequency Distribution of Status Crystallization Scores 35 8. Frequency Distribution of Scores on the Right-Wing Extremism Scale 90 9. Sample Size by Area, Eugene-Springfield Metropolitan Area Study 108 10. Occupational Status by Mail-Back Response and Non­ Response, and Total Sample 110 11. Income Status by Mail-Back Response and Non-Response, and Total Sample 110 12. Educational Status by Mail-Back Response and Non- Response, and Total Sample 110 11113. Percentage of Right-Wing Extremists by Occupation 14. Percentage of Right-Wing Extremists by Income 111 15. Percentage of Right-Wing Extremists by Education 111 viii Table Page 16. Percentage of Right-Wing Extremists by status 121Crystallization 17. Percentage of Right-Wing Extremists by Low and High Crystallization Groups, Controlled for Status Differences in Occupation, Income and Education 123 18. Original Correlation Matrix: Thirty Variables 159 19. Distribution of Correlations of Each Variable: TtNenty-Nine Variables . . . . . . • • . •. 161 20. Tentative Allocation of Variables to Clusters by the "B" Coefficient Method ....•..... 164 21. Mean Correlation Between Each Variable Iiith Each Cluster 168 22. Final Correlation Matrix: Tvrenty-Nine Clustered Variables 171 ix LIST OF FIGURES PageFigure 1. Final Correlation Profiles of Variables by Clusters .. 82 2. Schematic Representation of Status Crystallization as a Function of Deviations from an Equivalent Status Score Line . . . . . . . . . . . . . .. 115 3. Trial CorreJ_ation Profiles of Variables TtTith Groups .. 169 4. Trial Correlation Profiles of Variables in Tentative Clusters Band E 170 x CHAPl'ER I INTRODUCTION Historical Perspective For more than a decade American Society has been witnessing a political phenomenon which has been variously called The Extreme Right, The Radical Right, The Far Right or The Right Wing. Some writers view this as an entirely new phenomenon, a product of the international tensions which developed after World War II. Many others view it as a resurgence of a political philosophy which lies deep in the tradition of American society. Although an extensive analysis of the history of Right-Wing Extremism in the United States is beyond the scope of the present study, a brief outline of the contemporary development of this phenomenon will help to place it in its historical perspective. Within the past two decades, the Extreme Right has appeared to arise like a Phoenix from the ashes of declining American Socialism. Although Senator McCarthy's pronouncement regarding Communists in the State Department in February, 1950 is certainly one of the more spec­ tacular highlights of Right-Wing activity, the first contemporary appearance of this phenomenon goes back at least five more years to January 3rd, 1945. On this, the opening day of the Seventy-Ninth Congress, the dormant Dies' Special House Committee for the Investi­ gation of Un-American Activities was changed, through the passing of an '1 2 amendment introduced by representative John Rankin, into a permanent committee--the House Committee on Un-American Activities. Three months later, on April 12th, 1945, the New Deal passed into history with the death of Franklin Delano Roosevelt. During the following years, the Extreme Right gained further momentum, both publicly and privately. Publicly, the investigative process became further diversified with the formation, in 1946, of the Senate Permanent Subcommittee on Investigations of the Committee on Government Operations. Privately, one of the first voluntary associ­ ation groups of the Extreme Right, America's Future Incorporated, was founded in New Rochelle, New York. The House Committee on Un-American Activitie s was active in 1947 with the "Holly;.Tood inve stigations" and again in 1948 with the testimony of Elizabeth Bentley and Whittaker Chambers. Their testimony gave rise, through 1948 and 1949, to the perjury trials of Alger Hiss and William Remington. The elections of 1948 signify the last major attempt of the Socialists to gain political recognition and formal power in America. This was the year that the defecting Left Wing of the Democratic Party formed the Progressive Party and nominated Henry Wallace for the presidency, and was also the last year that anyone of the stature of Norman Thomas ran on the Socialist Party ticket. On the international scene, a number of shocks were dealt to the American people during 1949. American faith in China as a great power and friend was destroyed by the defeat of Chiag Kai-shek's Nationalist armies by the Communists. This was also the year that the 3 Soviet Union exploded its first atomic bomb, several years ahead of the date estimated by American scientists. As it was to be revealed later, Soviet technology had received assistance from the West in the person of Dr. Klaus Fuchs. There can be no doubt that the events of 1945 to 1950 helped to establish a climate that made many millions of Americans receptive to McCarthy and McCarthyism. In February, 1950, when he made his startling allegation of Communists in the State Department to a Women's Republican Club in Wheeling, West Virginia, the Senator from vlisconsin was still a "free lance"investigator. Undoubtedly, the tensions created by the Korean War, into which the United States entered in the summer of 1950, helped to sustain the wave of hysteria on which he was able to ride until the abrupt end of his political career in 1954. Al­ though overshadowed by McCarthyism at the time, the year 1950 also marks the inception of the McCarran Senate Subco~mittee on Internal Security. In 1953, McCarthy attained full investigative powers with his appointment as chairman of the Senate Committee on Government Oper­ ations and of its Permanent Subcommittee on Investigations. Paradoxi­ cally, this year also marks the beginning of the end for McCarthy, for it was then that he began his investigations that were to terminate in the disastrous Army-McCarthy hearings. By the mid-fifties many observers heralded a return to normal­ cy in American society. The tension created by the Korean War had been dispelled, Joseph Stalin was dead and it appeared that a better understanding between the East and the West was in the offing. In this 4 country, the first Republican government in twenty years appeared to have survived its initial difficulties. This was the occasion for many analysts to sit back and view the American Right Wing in retrospect, as though it had been a temporary canker that had now been removed. Writ­ ing in this period, Herbert Agar has stated: Because of the President's wise decision L?ot to form a third party in 19537, the American political system has returned, for the time being, to its normal shape. l The Right Wing, however, was not dead. Quietly, it was grow­ ing and spreading, and new groups of supporters for Extreme Right doc­ trine were forming throughout the country. What has turned out to be one of the most sensational of these new groups--the John Birch Society-­ was organized in Indianapolis on December 8th and 9th, 1958, where its founder Robert Welch met with eleven influential men from different parts of the country. The society achieved notoriety early in 1961 when the contents of a "personal letter" of Welch's entitled "The Politician" became publicly known. In this manuscript, which now runs into about 100,000 words, Welch accused some of the nation's leading figures of being either Communist agents, supporters or sympathizers. Among those accused were General George Marshall, President Eisenhower, Milton Eisenhower, Chief Justice Earl Warren, late Secretary of State John Foster Dulles and Alan Dulles, former head of the Central Intelli­ gence Agency. Following up the initial storm of controversy which disclosure of this document created, Welch, himself, in April, 1961, lHerbert Agar, The Price of Power: America Since 1945, Chicago: The University or-Chicago-Press, 1957, p. 162. ---­ 5 publicly urged a congressional investigation of the John Birch Society --a move no doubt intended to silence a growing opposition to that organization. In May of the same year Welch appeared on the television program "Meet The Press"--a rare exception to his no interview policy. Meanwhile, the calls for congressional investigation of the John Birch Society from other quarters were helping to ensure that organization a recognition not afforded to any group since the era of McCarthyism. The purpose of the foregoing brief historical introduction has been twofold. First, an attempt has been made to place the modern development of the Extreme Right in its proper sequence in the events of the past two decades. Second, it has served to emphasize the fact that the Extreme Right is still a viable phenomenon--one which does not need to be studied ex post facto. Sociological Perspective Social scientists who have turned their attention to the phen­ omenon of the Extreme Right seem to fall into four general categories: L There are those who dismiss this phenomenon as a "passing lphase" on the American scene. The ranks of this group are diminishing, since it is evident that this phenomenon, in its present manifestation, has been in existence for nearly two decades and is showing no evidence of decline. lSee, for example, Peter Vierick, "The Revolt Against the Elite," in Daniel Bell (ed.), The New American Right, New York: Criterion Books, 1955, pp. 91-116;"8':'" M. Lipset, tiThe Sources of the 'Radical Right' ," in Bell, op. cit., pp. 166-233. 6 2. There are those who have sought to describe, in more or less impressionistic and literary terms, what the Exteme Right is and how Right-Wing organizations operate. l The ranks of this group are swell­ ing, partly because of a genuine attempt on the part of some writers to increase our knowledge of this phenomenon, and partly because this is now a fashionable subject on which to write. 3. There are those who have made an attempt at explanatory analy­ sis of the Extreme Right. 2 These analysts have been generally more scientific in their approach, but their explanations have not been undertaken in such a way that would allow for prediction. Moreover, the variables which they have used to explain the Extreme Right have not been precisely defined. Examples of such variables are Daniel Bell's "status anxieties" and Talcott Parson's "strains" in the con­ temporary social system. 3 lSee, for example, Alan F. Westin, liThe Deadly Parallels: Radical Right and Radical Left,1I Harper's Magazine, vol. 224, no. 1343 (April, 1962), pp. 25-32; Fred J. Cook, "The Ultras: Aims, Affili­ ations and Finances of the Extreme Right, II The Nation, vol. 194, no. 26 (June 30, 1962); Donald Janson and Bernard Eismann, The Far Right, New York: McGraw-Hill Book Company, Inc., 1963; Arnold Forster and Benjamin R. Epstein, Danger on the Right: The Attitudes, Personnel and Influence of the Radical Right and Extreme conservatives, New York: Random House, 1964. --- -­ 2See , for example, Daniel Bell, "Interpretations of American Politics," in Bell, op. cit., pp. 3-32; Richard Hofstadter, "The Pseudo­ Conservative Revolt,W-in Bell, op. cit., pp. 33-55. 3naniel Bell, "Status Politics and New Anxieties: On the 'Radical Right' and Ideologies of the Fifties,1I in The End of Ideology: On the Exhaustion. of Political Ideas in the Fifties:-New York: Collier Books, 1961, pp. 103-123; Talcott Parson~IISocial Strains in America,1I in Daniel Bell (ed.), The ~ American Right, op. ~., pp. 117-140. 7 4. Finally, some attempts have been made to make predictive anal­ yses of the Extreme Right. To the writer's knowledge, the only study published to date that has attempted to isolate operationally the atti­ tudinal characteristics of contemporary Right-Wing Extremism is that of Martin Trow. l Indepepdent ~ Dependent Variables of the Present Study Although psychological variables have been used extensively in predicting social phenomena, the independent variable used in the present study will be sociological in nature. This foll~Ns Emile Durkheim's precedent, articulated in the following principle for investigating the causes of social phenomena: The determining cause of a social fact should be sought among the social facts preceding it and not among the states of the individual consciousness. 2 The independent variable of social class has been used frequently as a frame of reference in studies of political attitudes and political IMartin Trow, "Small Businessmen, Political Tolerance, and Support for McCarthy," American. Journal of Sociology, 64 (November, 1958), pp. 270-281. Trow defines the supporters of McCarthy as reac­ tionary "nineteenth-century liberals" who combine the attitudes of opposition to organized labor and opposition to big business. However, the validity of attaching a "regressive" or "reactionary" label to Right-Wing Extremism is doubtful. This point will be discussed more fully in Chapter III. Moreover, opposition to big business on the part of the Right Wing is questionable, in view of the potential economic support from big business which is sought by Right Wing groups. 2Emile Durkheim, The Rules of Sociological Method, eighth edition, translated by Sarah A. Solovay and John H. Mueller, Glencoe: The Free Press, 1938, p. 110. 8 behavior. l However, rather than using the traditional conception of social stratification as positions on one or more vertical hierarchies a more recent model depicting stratification as a horizontal configura­ tion of status variables will be used. This model, which is regarded by those who have used it as having considerable predictive useful­ ness, is that i-lhich has been referred to variously as "status equilibration,lI "status congruency," "status crystallization" and "status consistency."2 A related model, used more in demo~raphic analyses, is that of "status integration."3 For the purposes of the present study, "status crystallization" or "status consistency" (Vlhich terms will be used synonymously) 'will be defined as the extent to which an individual's rank positions on given status hierarchies are at a comparable level. lSee, for Example, P. F. Lazarsfeld, B. R. Berelson and Hazel Gaudet, The People's Choice, 2nd edition, New York: Columbia University Press, 1943; B. R. Berelson, P. F. Lazarsfeld and W. N. McPhee, Voting, Chicago: University of Chicago Press, 1954; Elihu Katz and P. F. Lazarsfeld, Personal Influence, Glencoe: The Free Press, 1955; S. M. Lipset, Political Man, New York: Doubleday and Company, Inc., 1960. 2Cf. Emile Benoit-Smullyan, "Status, Status Types, and Status Interrelations," American Sociological Revievl, 9 (April, 1944), pp. 151­ 161; Stewart Adams, "Status Congruency as a Variable in Small Group Performance," Social Forces, 32 (October, 1953), pp. 16-22; Gerhard Lenski, IlStatus Crystallization: A Non-Vertical Dimension of Social Status," American Sociological Review, 19 (August, 1954), pp. 405-413; Irwin iol. Goffman, "Status Consistency and Preference for Change in Power Distribution," American Sociological Review, 22 (June, 1957), pp. 275-281; Elton Jackson, "Status Consistency and Symptoms of Stress," American Sociological Review, 27 (August, 1962), pp. 469-480. 3See , for example, Jack P. Gibbs and Walter T. Martin, "A Theory of Status Integration and Its Relationship to Suicide," American Sociological RevievI, 23 (April, 1958) pp. 11.fO-147. 9 Considerable evidence, both theoretical and empirical, exists which suggests that there is a relationship between status inconsistency and political extremism, both left and right. l To this date, however, no test of the relationship between uncrystallized status and Right-Wing Extremism has been published. In view of the strong likelihood that lack of crystallization has a bearing on Right- Wing Extremism, status crystallization has been selected as the in­ dependent variable of the present study. A full discussion of the crystallization model, and the method used to measure this variable will be presented in Chapter II. The problem to which the present study will be addressed is that of predicting the incidence of Right-Wing Extremism with respect to individuals. Although this study will utilize material related to specific Right-Wing organizations for illustrative purposes, it will not be a study of anyone such organization, nor of a group of such organizations. Instead, this study will be concerned with the ideological and attitudinal nature of Right-Wing Extremism. The lSee, for instance, Invin W. Goffman, op. cit.; Elton Jackson, "Status Consistency, Vertical Mobility and Symptoms of Stress," Unpub­ lished Ph.D. Dissertation, University of Michigan, 1960; Gerhard Lenski, 009. cit. and "Social Participation and Status Crystallization," American SOciOIOgical Review, 21 (August, 1956), pp. 458-464; s. M. Lipset, "Social Stratification and 'Right-Wing Extremism1 , " British Journal of Sociology, 10 (December, 1959), pp. 1-38; s. M. Lipset and Reinhard -­ Bendix, liThe Consequences of Social Mobility, "in Social Mobility in Industrial Society, Berkeley and Los Angeles, University of California Press, 1960, pp. 64 ff. For a complete analysis of studies relating status crystallization to political extremism, see Chapter V belOi'I. 10 relationship between attitude, which lies within the individual, and ideology, which goes beyond the individual, is an inseparable one. Attitudes, in fact, may be regarded as indicators of the individual's ideological system. This frame of reference is exemplified by the definitions of these concepts as they will be used in the present study. The term "ideology" will be used as it is defined by Adorno: The term ideology is used in this book, in the way that is cow~on in current literature, to stand for an organization of opinions, attitudes, and values--a way of thinking about man and society. ~\le may speak of an individual's total ideology or of his ideology with respect to different areas of social life: politics, economics, religion, minority groups, and so forth. Ideologies have an existence independent of any individual; and those which exist at a particular time are results both of historical processes and of contemporary social events. l Thus, as the above definition indicates, attitude is a component of ideology. Usage of the term "attitude ll will follow that proposed by Krech and Crutchfield. They define attitudes as: ..• enduring systems of positive or negative evaluations, emotional feelings, and pro or can action tendencies with respect to social objects. 2 In light of the foregoing discussion, the dependent variable of the present stUdy v7ill be that of Right-Hing Extremism, defined as an ideology (specifically, as a political ideology) and operationalized Lr. W. Adorno, et al., The Authoritarian Personality, New York: Harper and BrotherS: 1950,-P: 2. 2David Krech, Richard S. Crutchfield, and Egerton L. Ballachey, Individual in Society: A Textbook of Social Psychology (A major revision of'Theory and Problems of SOcial Psychology by David Krech and Richard S. Crutchfield), New York: McGraw-Hill Book Company, Inc., 1962, p. 139. 11 in terms of attitudinal indicators. Defining and operationalizing this variable will be the topics of Chapters III and IV, respectively. Having defined and operationalized the independent and dependent variables of the present study in Chapters II, III and IV, the balance of the study will be devoted to analyzing the relationship between these variables. Chapters V, VI and VII will concern respec­ tively the theoretical specification of the hypothesis relating status crystallization to Right-Wing Extremism, the methodology involved in testing this hypothesis and the findings of the test, and a concluding discussion of this research and its implications for future studies. Although the format of this research is that of a predictive study, the limited resources for data gathering have restricted this investi­ gation to a secondary analysis of existing data. The source of this data is discussed in Appendix A. Appendix B contains the presentation of the cluster analysis used to identify the attitudinal indicators of the Extreme Right. Because of the limited nature of the data available for this study, those portions of the research concerned with the development of attitudinal indicators of the Extreme Right must be regarded as largely exploratory. In passing, it should be noted that the term which will be used in discussing the phenomenon under study will be "Extreme Right" (and the generic form, "Right-Wing Extremism") rather than the more usual "Radical Right. tl Accepted usage of the term "radical" connotes revolutionary, usually leftist, politics. In view of the fact that the Extreme Right lies considerably to the right of the center of the 12 political spectrum, it is felt that use of the term "Radical Right" would contribute to semantic confusion. Moreover, the term flExtreme Right" is consistent with the notion of political extremism. Since the position that the Extreme Right is a form of extremist political expression will be central to this study, the use of the adjective "extreme" seems more appropriate than that of "radical. 1I 13 CHAPl'ER II THE STATUS CRYSTALLIZATION MODEL The Concept of Status Crystallization Traditionally, social stratification has been conceptualized in terms of a single vertical dimension in which each person in the social system occupies a single position. One of the classical schemes in this tradition is that of Karl Marx, who saw the boundaries of • social classes drawn in terms of relationships to the means of produc­ tion (i.e., social class is economically determined). He envisioned modern capitalist societies as divided into two basic classes along these economic lines--the bourgeoisie and the proletariat. Marx also recognized the existence of other identifiable groups, such as the lumpenproletariat and the intellectuals. These groups, however, were considered by him to be peripheral to the production process, and thus not constituting classes, in his sense of the word. Outstanding among recent stratification models of the uni­ dimensional type is that developed by Warner. l He proposes two differ­ ent ways of measuring social class--Evaluated Participation (E.P.) and an Index of Status Characteristics (loS.C.). The former is a rating lW. Lloyd Warner, et al, Social Class in America: The Evaluation of Status, New York:--Harper and Brothers, 1960. 14 technique used for placing families and individuals in the status system of a community, and is basically a measure of social inter­ action. In this method of rating social class position, the analyst uses data, supplied by informants, on the following six factors: matched agreements of a number of informants, identification I,Iith certain symbols ("symbolic placement"), status reputation, comparison with others whose class position has previously been determined, simple assignment to a class mentioned by informants, and institutional membership. The I.S.C. scale is an index of socio-economic status, and involves totalling a series of weighted ratings on the character­ istics of Occupation, Source of Income, House Type and Dwelling Area. This summation yields a set of scores ranging from 12 to 84. From these, vlarner designates five social classes: Upper, Upper-Middle, Lower-Middle, Upper-Lower and Lower-Lower. Although the factors measured by these two methods of measur­ ing social class are quite different, the two measures are highly intercorrelated. That is, Warner shows that persons who have a high (or low) E.P. rating will also score high (or low) on the I.S.C. Thus, regardless of the number of status characteristics used and the complex­ ities of rating and scoring procedures, Warner's method yields only a single, vertical measure of social stratification. In recent years, the utility of the uni-dimensional approach to social stratification has been challenged by an increasing number of sociologists. Max Weber criticized Marx by pointing out that the econ­ omic concept of "class" vIas only one of at least three "laYs by which 15 groups could be stratified. In his essay, "Class, Status, Party,"l Weber distinguishes between these three dimensions as follows: 1. Class: this is the traditional economic dimension, being stratification according to the relationship of groups to the produc­ tion and acquisition of goods. 2. Status: this concept is within the "social" order in that status groups are stratified according to the principles of their consumption of goods as represented by a certain honor or prestige in following special styles of life. Weber subsumes occupation under the heading of status since social honor is claimed by virtue of the special style of life which may be determined by an occupation. 3. Party: this is stratification according to the acquisition of social pow'er. Since Weber's initial specification of these dimensions of social class, a number of sociologists have contributed to the develop­ ment and implementation of a multi-dimensional model of social strati­ fication variously called llstatus equilibration,"2 "status congruency,"3 lH. H. Gerth and C. Wright Mills (translators and editors), From Max Weber: Essays in Sociology, New York: Oxford University Press-;-I9"5B'":Chapter VII-,-llClass, Status, Party, II pp. 180-195. 2Emile Benoit-Smullyan, "Status, Status Types, and Status Interrelations," American Sociological Review, 9 (April, 1944), pp. 151-161. 3Stuart Adams, "Status Congruency as a Variable in Small Group Performance,tl Social Forces, 32 (October, 1953), pp. 16-22. 16 "status crystallization,"l and "status consistency."2 The theoretical development of this model may be traced through the writings of Emile Benoit-Smullyan, Pitirim Sorokin and Gerhard Lenski. In 1944, Benoit-Smullyan drew Weber's three concepts together under a more general concept of "statu.s."3 Benoit-Smullyan was con­ cerned with the problem of distinguishing between status, which he defined as location on a superiority-inferiority scale with respect to possession of a common characteristic, situs, which he defined as mem­ bership in a social group, and locus, which he defined as a person's socially defined function in an organized group. Of importance to this study is his concept of statu.s, which consisted of three hierarchical scales upon which people could be stratified. These are "economic," "political," and "prestige" statuses, which are respectively equatable with Weber's "class," "party," and "status." Benoit-Smullyan then turns to a description of the interrelations between these various types of status and it is here that the "status equilibrium" hypothesis is first stated. He notes that a person's positions on these three statuses often may be found to be out of line. That is, the wealthiest lGerhard E. Lenski, "Status Crystallization: A Non-Vertical Dimension of Social Status," American Sociological Review, 19 (August, 1954), pp. 405-413. 2Irwin Ii. Goffman, "Status Consistency and Preference for Change in Power Distribution,'.' American Sociological Review, 22 (June, 1957), pp. 275-281; Elton F. Jackson, "Status Consistency and Symptoms of Stress," American Sociological Review, 27 (August, 1962), pp. 469­ 480. 3Benoit-Smul1yan, op. cit. 17 .. are not always the highest in power or prestige (e.g. the nouveaux riches), nor are the most prestigeful always highest in wealth or power (e.g. the "shabby genteel"). He goes on to say, however, that when such dissimilarity does obtain there is a tendency over time for a per- son's rankings on these various status hierarchies to reach a common level. This, then, is the original statement of the status equilibra­ tion concept. Sorokin, in 1947, deals with the same idea when he discusses various stratification strata. l A state of equilibrium would obtain in what Sorokin terms the llinnerly solidary or affine multi-bonded stra­ tum." This would be a stratificational cross-section in which the stratifying bonds were mutually congenial. He raises the point of sta­ tus disequilibration when he states that "the correlation of the affine . strata is never perfect .,,2 He then elaborates on this point: The stratified pyramids of the uni-bonded groups never con­ solidate in such a way that all their strata coincide and create one integral consolidated social pyramid in which all the tops of the uni-bonded pyramids make one integral top, and all the middle and lower strata consolidate into one in­ tegral middle or lower stratum. 3 Aside from the relevance of this statement to the concept of status disequilibration, it raises a further important theoretical point: The pp. lPitirim A. ir Structure and D 289-294. Sorokin, Society, ynamics, New York: Culture, and Personality: Harper and Brothers, 1947, 2Ibid ., p. 292. 3Ibid ., p. 292. 18 equilibration may take place at any crosscutting level for the number of status hierarchies under consideration. That is to say, an individ­ ual's statuses may be equilibrated if they are all at the top level of ~ a number of given hierarchies, if they are all at the bottom level, or at any equivalent in-between point on the hierarchies. Conversely, status disequilibration obtains if one or more of an individual's statuses are divergent from the level held by his other statuses. An important methodological consideration arising from Sorokin's observa­ tion is that all status hierarchies constituting a status crystalliza­ tion model should have equivalent intervals. That is, all vertical status dimensions used in such a model should have an equal, or nearly equal, number of class intervals in order to ensure a valid measure of status crystallization. In a 1958 study of 1>lOrker motivation, Zaleznik, Christensen, Roethlisberger and Homans1 failed to observe this require­ ment, and used a two-interval variable (sex) in conjunction with multi- interval variables (e .g., occupation). This factor may '.;Tell have been contributory to the failure of their hypotheses to be borne out. Lenski, in 195Lf, developed a method for quantifying status crystallization. 2 For his vertical status hierarchies, he used the four statuses of occupation, income, education and ethnicity. These lAo Zaleznik, C. R. Christensen and F. J. Roethlisberger, with the collaboration of George C. Romans, The ~1otivation, Productiv­ ity and Satisfaction of Workers: A Prediction-Study, Boston: Harvard UniveT.Sity Press, 195~ - 2Lenski, ~. cit. 19 were selected because they were considered to be important stratifica­ tion characteristics, and because necessary information related to them could be obtained easily. Education and income were, of course, easily ranked, being quantitative in nature. Using the National Opinion Research Center's study of occupational prestige,l five occupa­ tional prestige levels were defined for the sample under study. The ethnicity scale was constructed by means of rankings by Detroit area students enrolled in introductory sociology courses at the University of Michigan. Having thus established four vertical status hierarchies, ranked on superiority-inferiority continua, Lenski's next problem was to establish common scales for these four hierarchies, so that the relative position of respondents on each of them might be compared. First, a frequency distribution of respondents was established for each of the status hierarchies. From these, cumulative percentile ranges for each hierarchy were computed. A score for each flass interval was then assigned on the basis of the midpoint of the per­ centile range for that intervaL The last step in this procedure was to arrive at a quantitative measure of status crystallization. Lenski reports this step as follows: lNational Opinion Research Center, "Jobs and Occupations: A Popular Evaluation," Opinion News, 9 (September 1, 1947), pp. 3-13. 20 This ~uantitative measure of status crystallizati0ri7 was accomplished by taking the square root of the sum of the squared deviations from the mean of the four hierarchy scores of the individual L?ccupaticn, income, education and ethniciti7 and subtracting·the resulting figure from one hundred. The more highly consistent or crystallized an individual's status, the more nearly his crystallization score approached one hundred; the less consistent or crys­ tallized his status, the more nearly his crystallization score approached zero.1 In algebraic terms, the square root of the sum of squared deviations from a mean may be expressed as J,[ d.1. or jf-x". This the reader i-Jill recognize as the formula for Standard Leviation (11":: J~':lt") without the denominator in it. The use of squared deviations from the mean in computing status crystallization scores has the effect of maximizing score dispersion. Lenski was seeking maximum dispersion in order to increase the discriminatory power of his status crystallization meas­ ure. This observation is consistent with Lenski's reporting that he used squared deviations from the mean in order to emphasize the effect of larger deviations and to minimize the effect of smaller ones. He considered this necessary because no great importance could be attached to small deviations in the status hierarchies--a conlequence of the cr\~de techniques used in quantifying intervals on these hierarchies. Subtracting the computed scores from one hundred was done for reasons of semantic clarH;y. That is, such a procedure would give responde nts lLenski, 2.£. cit., pp. L.o7-408. 21 with a highly crystallized status a numerically higher score than those with a poorly crystallized status. l The Status Crystallization Model Used Herein For the purposes of the present study, status crystallization will be defined as the extent to which an individual's rank positions on given status hierarchies are at a comparable level. 2 The crystal­ lization model used in this study will closely parallel that of Lenski, as discussed earlier. The statuses to be used are occupation, income and education. These were chosen because of two factors. First, it is generally considered that these statuses are important ones with , respect to social stratification.3 Second, the necessary information related to these statuses is available, and relatively easy to quantify. Although Lenski used ethnicity as a fourth status characteristic, it will not be used in this study. First, information on ethnicity is lFor a detailed illustration of the application of Lenski's status crystallization measure to the present study, see infra, p. 2Goffman, op. cit., p.275, defines status consistency as It ••• the extent to-Which an individual occupies ranks on relevant status dimensions that are defined as comparable in shared expecta­ tions." Jackson, op. cit., p. 469, defines the concept as " •.. the degree to which an-rndividual's rank positions on important societal status hierarchies are at a comparable level." 3see , for example, August B. Hollingshead, Elmtown's Youth: The Impact of Social Classes on Adolescents, New York: John· Wiley and Sons, 1949;--A. B. Hollingshead and F. Redlich, Social Class and Mental Illness, New York: John Wiley and Sons, 1959; W. Lloyd Warne~et al, Social Class in America, op. cit.; James West, Plainville, U.S.A~ New York: COlUmbia UniversitY-Press, 1945. ----­ 22 not available in the data used for this study. Second, one would not expect a wide range of ethnic groups to be represented in the Eugene- Springfield area. This region does not have a highly diversified productive base, as is the case in the Detroit area. Finally, ranking ethnicity remains a subjective procedure, and is thus open to consider­ able doubt as to reliability. Lenski, after a review of literature on social status, pointed out that there is considerable evidence to the effect that the family may be considered as a status unit, and that the status charac­ teristics of the family head are the chief determinants of status for other dependent family members. l To the extent that it is possible with the data at hand, status information for head of the family will therefore be used in the present study. In the sample under study, basic interview schedule and mail-back data were available for 741 respondents. Of these, 310 were ~le and 431 were female. Of the latter, 206 out of the 289 who were married listed no occupation other than "House1'iife or Unemployed." Since occupation of spouse was avail­ able for this sample, using the occupation of the head of the household precluded the elimination of the bulk of the "married female" group. Income was given as "family income,1t the use of which is consistent with the concept of treating the family as a status unit. Education, however, was given as that of the respondent himself, and no education data for spouse is available for the It married female" group. However, lLenski, op. cit., p. 407. 23 lack of such educational data in the present study does not seriously violate the principle of treating the family as a status unit. Al­ though the occupation of a family head does determine the "prestige" of the family, and total family income determines the economic "style of life" of the family, education may be regarded more as a "personal" status characteristic of the individual, and not a determinant of family"status." If, however, considerable discrepancy exists between the educational levels for males and females in any sample under study, the differential use of male and female education within families will invalidate comparisons of family units. This problem is not raised in the present study, however, since the median education for male re­ spondents is 12.6 years, and the median education for female respond­ ents is 12.5 years. In the data on which the present study is based, occupation was originally coded according to census categories, as shown in Table 1, page 24. In the final occupational classification used in this study, the categories of "Farm owners or Farm managers" and "Farm worker or Farm foremen" were eliminated, since no respondents were recorded in them. In the original sample of 741 respondents, 74 were coded as "Retired or Student." Since this represents a fairly high proportion of the sample, and since impressionistic evidence regarding support for the Extreme Right suggests that retired persons might be particularly prone to extremist political attitudes, it was decided that the category of "Retired or Student" would be retained for special treatment, in an attempt to separate the retired persons from it. 24 TABLE 1. ORIGINAL OCCUPATIONAL CATEGORIES, WITH TYPICAL Occupational Category Technical and Professional Farm Owners or Farm Managers Business and Managerial, Proprietors Clerical and Sales Craftsmen Machine Operators Armed Forces (enlisted) , Service Workers Laborers Farm Workers or Farm Foremen Retired or Student Housewife or Unemployed OCCUPATIONS REPRESENTED Typical Occupations Represented Clergyman, engineer (graduate engineer or equivalent, e.g., civil, chemical, aero­ nautical, mechanical), physician, dentist professor, teacher (elementary, high school), technician or semi-professional (M.S., M.A., or higher academic degree or its equivalent), lawyer Self explanatory Business official (credit man, buyer, college bursar), government official (federal, state and local officials with professional status, also commissioned officers of the armed services), managers (production manager, supervisor in a factory, but not a foreman), proprietor (self-employed businessman, contractor, distributor) Clerical worker (file clerk, general office worker, typist, bookkeeper), salesman (sales agent, insurance salesman, realtor), sales­ person (retail, local or small district merchandise salesman) Carpenter, plumber, electrician, machinist Factory machine worker, bus driver, switchman Member of the armed forces (enlisted), occupations not elsewhere classified, service worker (policeman, fireman, letter carrier, waiter,barber, COOk, usher, attendant) (except farm) Self explanatory 25 The procedure used will be discussed later in this chapter. The cat­ egory of "Housewife or Unemployed" also received special treatment. Married females in this category were classified on the final occupa­ tional scale according to the occupation of spouse. The rationale for this procedure has already been discussed. For males and "unattached" females (single, widowed, divorced or separated) who ~ of retirement age, the assumption has been made that "unemployed," for the purposes of occupational classification, is roughly equivalent to "retired." Therefore, persons in the "Housewife or Unemployedtl group, who are not classifiable by spouses' occupation, will be treated in the same manner as those who are in the "Retired or Studenttl group. In addition to age, sex and marital status, one further vari­ able was used to assist in classifying respondents, particularly those in the two categories discussed in the preceding para.graph, by occupa­ tion. This variable is "source of income,tI which also enters into the classification of divorced and separated females. The coding procedure for this factor is presented in Table 2 below. TABLE 2. CODING CATEGORIES FOR SOURCE OF INCOME 1. Husband's salary or wages from employer 2. Wife's salary or wages from employer 3. Income from own business or profession 4. Rentals of property to others 5. Annuities or pensions (including G.I. Bill) 6. Interest or dividends on investments 7. Farming 8. Welfare or social security payments 9. Other income 10. Don't Know 11. No Answer 26 Based on the foregoing discussion of relevant variables, a revised system for classifying respondents by occupation was developed. In the following outline of procedures, the categories of "Retired or Studentll and "Housewife or Unemployed" were excluded in each case, and treated separately. 1. Male; all marital statuses: With the exception of the categories noted above, all male respondents were classified according to their occupation as given on the schedule. 2. Female; single and widowed: Being considered as unattached to a household unit, females in these two categories were classified according to their occupation as given on the schedule. 3. Female~ married: For this group, the occupation of spouse was used in making the revised classification. As indicated earlier, this functions to classify the family according to the occupation of its head. 4. Female; separated and divorced: In those cases where the main source of support for the respondent was from the former spouse, and where an occupation for such spouse was given, it was used for the occupational code. For self­ supported females in these two groups, their own occupation, as given on the schedule, was used. With respect to .the categories of URetired or Student" and "Housewife or Unemployed," the following technique of classification 27 was used. Those who were of possible retired age (taken as fifty-five and over, which ensures inclusion of the "voluntarily" retired), and for whom a source of income was indicated, were retained. The remain­ ing cases were discarded as "not classifiable." This residual category thus contains all those who are either students, unemployable, or tem­ porarily unemployed. Assuming that one's occupation determines, in large measure, his source of income, the remaining respondents in these two categories (jointly classified as "retired") were placed in the occupational classification according to their main source of income. On the basis of this variable, three categories of retired persons may be discerned: 1. Retired on independent income (income from own business or profession, interest on dividends or investments, or rentals of property to others). 2. Retired on annuities or pensions. 3. Retired on welfare or Social Security. Males and females in the retired group were treated according to the system of classification outlined in procedures 1 through 4 above, substituting "source of income" for "occupation." This system of classification serves to retain the retired respondents in the sample, while making it possible to classify them meaningfully with respect to occupation (i.e., source of income). In order to place these three retired categories in the occupational hierarchy, certain assumptions regarding the meaning of "source of income" must be made. First, it may be considered that an independent source of income (i.e., income from one's own business or 28 profession, interest or dividends on investments, or rentals of prop­ erty to others) warrants ranking close to the top categories of salaried individuals, and above those whose main source of income is from wages. Second, it is assumed that those who are retired on annuities or pensions spent most of their occupational career in those middle or upper blue and white collar occupations that either provided a pension plan, or made it economically possible to set aside a retire­ ment income in the form of life insurance, etc. Therefore, this cat­ egory should be placed somewhere near the middle range of occupations. In order to minimize the possible error of ranking persons in these two first retired categories too highly on the occupational scale, these categories have been inserted at the bottom of the general groups to which they seem to belong. This placement will be indicated in Table 3. Third, it is assumed that those who derive their main income from welfare or social security should rank at the bottom of the occu­ pational hierarchy. The revised occupational classification, arrived at on the basis of the foregoing discussion, is presented in Table 3 on page 29 below. This ranking of occupational groups closely parallels that prepared by Alba Edwards for the U.S. Bureau of the Census,l with the exceptions that "Clerical" and "Sales" are combined, and that all ser­ vice workers are classified together. lAlphabetical Index of Occupations and Industries, prepared by Alba M. Edwards, Bureau of the Census, U.S~epartment of Commerce, Washington, D.C.: U.S. Government Printing Office, 1940. 29 TABLE 3. REVISED OCCUPATIONAL CLASSIFICATION Occupation Number Technical and Professional 85 Business and Managerial, Proprietors 87 Retired on independent income 34 Clerical and Sales 101 Craftsmen 80 Machine Opera.tors 106 Retired on annuities or pensions 29 Armed Forces (enlisted), Service Workers 80 Laborers 65 Retired on welfare, social security 28 Not Classifiable 46 Total 741 In keeping with the previously discussed requisite of an equal number of class intervals in each vertical status hierarchy, income and education were also classified on a ten-interval range. Thus, the income hierarchy is merely a ten-interval range from the lowest to the highest incomes represented in the sample. Similarly, the educational hierarchy is a ten-interval range from the lowest to the highest years of schooling completed by the respondents in the sample. Only one category, that of "Trade or Business School," is not expressed in terms of years of schooling completed. This type of training is, for the most part, related to skilled blue-collar 30 occupations and secretarial and stenographic white-collar occupations. Since it may be regarded that trade and business schools impart train­ ing in specialized skills--training not generally available in the public school system--this educational category has been placed between ItCompleted high school" and "Some college." Tables 4, 5 and 6 present the scores assigned to the rank­ intervals on each of the three status hierarchies of occupation, income and education. The method used is that proposed by Lenski,l and is based on a frequency distribution of respondents in each of the status hierarchies. From these frequency distributions, cumulative percentile ranges are computed. Scores for each interval are then assigned on the basis of the midpoint of the percentile range for each interval. In addition to the forty-six cases lost in the occupational classifica­ tion, twenty-two more cases lacked information on either income or education. Therefore, a sample of 673 respondents remains for which complete occupation, income and education data is available. Using the status scores computed in Tables 4, 5 and 6, status crystallization scores ,,[ere next computed for each respondent in the sample, according to the following formula: 3Status Crystallization Score = 100 - j"i. d , where d represents the deviations from the mean for each of his three status hierarchy (percentile) scores, and 100 the score inversion factor. This formula expresses, in notational form, the calculation lLenski, op. cit., p. 407 31 TABLE 4. FREQUENCY DISTRIBUTION OF RESPONDENTS BY OCCUPATION, * AND ASSIGNED PERCENTILE SCORES FOR EACH. OCCUPATION RANGE Occupation Number Cumulative Assigned Percentile Score Range Technical, Professional 82 87.9 - 100.0 94 Business and Managerial, Proprietors 84 75.4 - 87.8 82 Retired, independent income 31 70.8 - 75.3 73 Clerical and Sales 100 56.0 - 70.7 63 Craftsmen 80 44.1 - 55.9 50 Machine Operators 105 28.5 - 44.0 36 Retired, annuities or pensions 25 24.8 - 28.4 27 Armed Forces (enlisted), Service Workers 78 13.2 - 24.7 19 Laborers 64 3.7 - 13.1 8 Retired, welfare or social security 24 0.0 - 3.6 2 Total 673 * Occupation of family head for married and other attached individuals; respondent's own occupation for single individuals. 32 TABLE 5. FREQUENCY DISTRIBUTION OF RESPONDENTS BY INCOME,* AND ASSIGNED PERCENTILE SCORES FOR EACH INCOME RANGE Annual Income Number Cumulative Assigned Percentile Score Range $15000 and over 17 97.6 - 100.0 99 10000 - 14999 48 90.4 - 97.5 94 7000 - 9999 106 74.7 - 90.3 83 6000 - 6999 95 60.6 - 74.6 68 5000 - 5999 127 41. 7 - 60.5 51 4000 - 4999 101 26.7 - 41.6 34 3000 - 3999 68 16.6 - 26.6 22 2000 - 2999 42 10.4 - 16.5 13 1000 - 1999 49 3.1 - 10.3 7 less than 1000 20 0.0 - 3.0 2 - Total 673 * Income of total family for married and other attached individuals; respondent's own income for single individuals. 33 TABLE 6. FREQUENCY DISTRIBUTION OF RESPONDENTS BY EDUCATION,* AND ASSIGNED PERCENTILE SCORES FOR EACH EDUCATION RANGE Education Number Cumulative Assigned Percentile Score Range Graduate work 46 93.3 - 100.0 97 Completed undergraduate college 68 83.2 - 93.2 88 Some College 105 67.6 - 83.1 75 Trade or Business School 47 60.6 - 67.5 64 Completed High School 182 33.5 - 60.5 47 11 years of schooling 39 27.7 - 33.4 31 9 - 10 years of schooling 71 17.2 - 27.6 22 8 years of schooling 88 4.1 - 17.1 11 7 years of schooling 14 2.0 - 4.0 3 1 - 6 years of schooling 13 0.0 - 1.9 1 Total 673 * Education of respondent as reported on interview schedule. 34 --.-----.- ..._--'---,.,-;-, -.- --­ of status crystallization scores developed by Lenski. Before inver­ sian, these scores are similar to a standard deviation of percentiles. l The following hypothetical example will illustrate the calculation procedure. Assume that a respondent has occupation, income and educa­ tion scores (from Tables 4, 5 and 6 respectively) of 36, 68 and 47. The mean for these three scores is 50.33. Deviations from the mean are thus -14.33, 17.67 and -3.33. The respondent's status crystalliza­ tion score is: Status Crystallization Score =100 _ j! ~2 J 2 2 t = 100 - (-14.33) + (17.67) + (-3.33) .. 100 - )528.60 = 100 - 22.99 .. 77.01 The process of calculating 673 such scores was expedited by the use of the IBM 1620 Computer in the University of Oregon Statistical Lab and Computing Center. 2 The distribution of status crystallization scores obtained for the sample under study is presented in Table 7, page 35. Recent research using the status crystallization model has indicated that certain variables are closely associated only with highly uncrystallized status. These findings appear consistent with Parsons' observations that certain llmechanisms of adjustment" exist lFor a discussion of this point, see supra, p. 20. 2Appreciation is expressed to the University of Oregon for financial support of this computing work, and to Ronald Sikes (senior in Economics) and Terry Byers (graduate in Mathematics) for programming and computing assistance. 35 TABLE 7. FREQUENCY DISTRIBUTION OF STATUS CRYSTA~LIZATION SCORES* Score Frequenc.l Score -­ Frequency 32 1 66 8 33 - 67 20 34 1 68 9 35 1 69 2 36 - 70 16 37 3 71 25 38 - 72 28 39 - 73 2 40 3 74 13 41 3 75 12 42 3 76 12 43 - 77 33 44 L~5 9 4 78 79 8 16 46 5 80 23 47 2 81 9 48 7 82 12 49 7 83 10 50 9 84 18 51 5 85 13 52 4 86 13 53 54 15 2 87 88 9 16 55 10 89 33 56 6 90 25 57 58 8 8 91 92 12 10 59 60 10 5 93 94 4 21 61 11 95 16 62 - 96 8 63 15 97 8 64 19 98 9 65 22 99 - 2 Total 673 * A high score indicates "crystallized" status; a low score indicates "uncrystallized" status. 36 in American society that tend to reduce the effects of status discrep­ ancies. l Thus, it may be expected that only pronounced status dis­ crepancies, which are not as greatly conpensated for by these mechanisms as mild discrepancies, will result in given symptomatic reactions. Lenski, in his study of political liberalism, found that variable to be most closely associated with status crystallization when he compared the least crystallized quarter of his population (actually, the least crystallized twenty-eight per cent) with the most crystallized three-quarters. 2 He gave as his reason for establishing this cutting point the fact that a "natural break" could be observed at this point. Kenkel, in a replication of Lenski's study, cut his sample into the two crystallization categories at the median. 3 He reported finding no difference between the two crystallization groups with respect to the variable under study. In replying to Kenkel's commentary, Lenski stated that he deliberately made a comparison between the most crystallized three-quarters of his sample and the least crystallized one-quarter, and concluded as follows: lTalcott Parsons, "A revised Analytical Approach to the Study of Social Stratification," in Reinhard Bendix and Seymour Martin Lipset (eds.), Class, Status and Power, Glencoe: The Free Press, 1953, pp. 92-128. The two mechanisms, discussed on page 122, are the relatively wide range of facilities available to the public without specific status implications, and the degree of insulation that exists between different status groups, thus preventing excessive direct contact. 2Lenski, op. cit. 3William F. Kenkel, "The Relationship Between Status Consist­ ency and Politico-Economic Attitudes," American Sociological Review, 21 (June, 1956), pp. 365-368. 37 The evidence observed in the original study indicated that liberal tendencies seemed to result only from marked or pronounced inconsistencies of status. l In a more recent study, Jackson reports the following finding: Thus, in a gross comparison of consistency LCrystallizatio~7 categories, only the least consistent fifth of the sample appeared to be different from the consistent group in LPsycho­ somati~7 symptom level. 2 In light of the foregoing evidence, 101'; and high crystalliza­ tion categories for this study will be established by dividing the sample at the first quartile, taking as a base the lowest crystalliza­ tion score. This is accomplished by dividing the frequency distribu­ tion given in Table 7 between scares 64 and 65. The low crystallization group thus consists of 176 cases (the sum of frequencies from score 32 to score 64), which represents twenty-six per cent of the total sample of 673 cases. The remaining three-quarters of this sample will there­ fore be regarded as having high status crystallization. In this chapter, the independent variable of the present study--status crystallization--has been defined and operationalized. The next chapter will be devoted to arriving at a definition of the dependent variable of Right-Wing Extremism. lGerhard Lenski, "Col11'TI.ent on Kenkel's Communication," A'TI.erican Sociological Review, 21 (June, 1956), pp. 268-369, quoted at p. 369. 2Elton F. Jackson, "status Consistency, Vertical Mobility and Symptoms of Stress," unpublished Ph.D. dissertation, University of Michigan, 1960, pp. 118-119. 38 CHAPl'ER III TOWARD A DEFINITION OF THE EXTREME RIGHT Current Definitions, Explicit ~ Implicit One of the most persistent impressions one gets when reading the considerable body of literature that has accumulated on the phenom­ enon of the Extreme Right is that there seems to be a tacit assumption on the part of the writers concerned that everyone knows exactly what is being discussed. Such an assumption may be sufficient for impres­ sionistic observations about the Extreme Right, but if any intensive and scientific study is to be made of this phenomenon, a clear defini­ tion of what is to be studied would be in order. No study can claim to be "scientific" unless a.n explicit definition on the subject to be studied is given. This logical necessity of a definition has been pointed out by Cohen and Nagel: Logically, definitions aim to lay bare the principal features or structure of a concept, partly in order to make it definite, to delimit it from other concepts, and partly in order to make possible a systema.tic exploration of the subject matter with which it deals,l ThUS, it would appear that the first concern of this study should be to investigate what definitions of the Extreme Right are currently in use, lMorris Cohen and Ernest Nagel, An Introduction to Logic and Scientific Method, New York: Harcourt, Brace and Company,-r9~p-.-­ 231-232. 39 with a view to developing a definition capable of being expressed in operational terms. In Chapter I, it was pointed out that the definition of the Extreme Right to be used in this study would have to satisfy two gen­ eral conditions--it should focus on the ideological aspect of the Extreme Right, and it should be capable of being operationalized in terms of attitudinal indicators. All of the definitions which will be considered in this section seem to satisfy the first of these condi­ tions. Some explicitly contain the terms "ideology" or "movement," the latter of which, in the impressionistic sense of the word, connotes an ideological organiza~ion of beliefs, values, opinions, feelings, etc. about some social object. Others imply the concept of ideology in the use of words such as "dissent," "discontent," "conviction," and so forth. However, all of these definitions contain shortcomings with respect to the definiens, not the least of which is the inclusion of terms too vague to be expressed in operational terms. This emphasis on operationalism is not meant to imply, of course, that operational definitions are the only kind that have a place in sociological re­ search. In any investigation of social phenomena, one has to begin somewhere, and certainly a logical starting point is an ad hoc classi­ fication of what is to be studied. This kind of definition facilitates the meaningful organization of the researcher's observations, and helps him to limit his analysis to the specific problem at hand. However, such an approach to the definitional question is not sufficient for 40 more systematic investigations requiring explicit indicators of the subject under study. Thus, although the definitions reviewed in this section serve to focus attention on certain characteristics of the Extreme Right, they all suffer shortcomings l-Jhich invalidate them for the particular research purposes of the present stuqy. These definitions may be grouped according to three weaknesses: phenomenological misinterpreta­ tions, lack of equivalence between the definiendum and the definiens, and vague or obscure referents. l 1. Phenomenological misinterpretations in the following defini­ tions arise from two sources--either their authors failed to take into account certain properties of the Extreme Right in their analyses, or events subsequent to the formulation of the definitions have rendered them invalid. The first definition to be considered in this section is the following statement regarding Right-Wing groups: What all these groups are at heart. is the same old isolationist, Anglophobe, Germanophile revolt of radical Populist lunatic-fringers against the eastern, educated, Anglicized elite. 2 First, while it is true that certain points advocated by Right-Wing groups are isolationist in nature (e.g., withdrawal from the United lThe second and third of these points are discussed in Cohen and Nagel, ~., pp. 238-241. The authors also draw attention to the problems of circular definitions and definitions expressed in negative terms. However, neither of these two problems are evidenced in the definitions considered in this section. 2Peter Viereck, liThe Revolt Against the Elite,1I in Daniel Bell (ed.), ~~ America.n Right, New York: Criterion Books, 1955, p. 95. 41 Nations), this concept seems inconsistent with the aggressive militancy of the Extreme Right about foreign affairs. This militancy goes beyond the position of a "world-wide Monroe Doctrine," and would have the United States be the aggressor nation in war against the Communist bloc. Second, whereas American Populism of the late 1880's was basically a revolutionary agrarian movement, directed largely against urban interest groups, it is doubtful that "populist" can adequately define the Extreme Right of today. Although considerable support for the Extreme Right comes from rural areas of the South and the Mid-West, it also draws a great deal of support from the urban upper classes (e.g., business executives, entrepreneurs, and the D.A.R.). This problem of defining the Extreme Right in terms of oppo­ sition to the upper classes also characterizes Parsons' definition of McCarthyism, which may be taken as an example of Extreme Right expres­ sian. In Parsons' terms, McCarthyism was: . . . both a movement supported by certain vested-interest elements and a popular revolt against the upper classes. l As will be demonstrated later in this chapter, a major characteristic of the Extreme Right is its opposition to certain socialistic trends in modern society. The bureaucratic control and management and the higher tax rates that are concomitant with this modern trend would conceivably tend to alienate the upper classes, particularly the "ragged aristoc­ racy,1l from the main stream of American politics, at least from an lTalcott Parsons, IlSocial Strains in America," in Bell, ibid., p. 136. 42 ":"'":"-=--- .~~.~ .~-~~:-=~-_::-=-.,:-=._~.:::::=-._.;'-:.:~:::.:_._:..::-::-=-:::.~:=~:..­ economic point of view. Thus, there would appear to be greater justi­ fication for placing the old, elite classes on the same side of the political fence as the Extreme Right, rather than in opposition. This nature of Extreme Right support is documented by Janson and Eismann, who observe: First to produce a sizeable following for the Far Right were the Southern and Southwestern bastions of religious fundamentalism and political conservatism and the upper­ class suburbs of urban areas. l A final point regarding Parsons' definition is his reference to "vested­ interest elements." This concept would be difficult to interpret in operational terms, since Parsons does not indicate what these elements are, nor the nature of their interests. 2 The last definition to be considered under this heading is the only explicitly-stated one to be reviewed, and was arrived at through an extensive survey of the literature of, and pertaining to the Extreme Right: In summary, the American Right Wing may be said to include all those who share the conviction that the relationship of government to the individual should be severely limited.3 Although this definition is accurate as far as it goes, it only par- IDonald Janson and Bernard Eismann, The Far Right, New York: McGraw-Hill Book Company, Inc., 1963, p. 7. - -- --­ 2Although this point actually belongs in the section on "vague or obscure referents," it was included here to avoid repetition and confusion. 3Ralph E. Ellsworth and Sarah M. Harris, The American Right Wing: A Report to t14 Fund for the Republic, Washington, D.C.: Public A'ffairs-Press, 19b27P.4I':" - ­ 43 tially characterizes the phenomenon of the Extreme Right. Its limita­ tion becomes evident when we observe that extremist opposition to government is contradicted by the fact that the Extreme Right tolerates or even encourages "big government l1 in matters of national security or national defense. Moreover, this definition would also characterize left-wing anarchists. 2. The rule regarding equivalence between the definiendum and the definiens is expressed by Cohen and Nagel as follows: A definition must give the essence of that which is to be defined. The definiens must be equivalent to the defini­ endum--it must be applicable to everything of which the definiendum can be predicated, and applicable to nothing else. l Of the definitions reviewed, two do not appear to meet this require­ ment. The first is given by Hofstadter, who defines the Extreme Right as a: . . • dissent . • • LWhic~7 can most accurately be called pseudo­ conservative ... because its exponents, although they believe themselves to be conservative and usually employ the rhetoric of conservatism, show signs of serious and restless dissatis­ faction with American life, traditions and institutions. 2 Hofstadter's net would appear to be cast too wide to catch only the Extreme Right. Believing oneself to be a conservative, and being dis­ satisfied with American life, traditions and institutions would certainly characterize the political attitude of many conservatives who are not Right-Wing Extremists. However, before summarily rejecting the concept of I1pseudo-conservative," a closer investigation of who lCohen and Nagel, op. cit., p. 238. 2Richard Hofstadter, "The PseUdo-Conservative Revolt," in Bell, op. cit., pp. 34-35. ---_._,:-~.,--~ '---,•.._-_.-~..:":";""~-=:.-===--::~~:;:.::...::.:.:-~_.- 44 the pseudo-conservative is should be made. The concept is taken from Adorno, who defined the pseudo-conservative as follows: The pseudo-conservative is a man who, in the name of up­ holding traditional American values and institutions and defending them against more or less fictitious dangers, consciously or unconsciously aims at their abolition. l This definition, like so many of a psychological nature, falls far short of one which can be operationalized in sociological terms. What are "traditional American values and institutions" and "more or less fictitious dangers"? Do all those who uphold these values and institu­ tions and defend them against dangers, fictitious or otherwise, aim at their abolition? Conversely, do those who aim at their abolition uphold and defend them? Too much latitude is given by the phrases "more or lesst! and lbonsciously or unconsciously." Hofstadter himself asks, and answers, the question of who is a pseudo-conservative: Who is the pseudo-conservative and what does he want? It is impossible to identify him by class, for the pseudo­ conservative impulse can be found in practically all classes in society, although its power probably rests largely upon its appeal to the less educated members of the middle class. The ideology of the pseudo-conservative can be characterized but not defined, because the pseudo-conservative tends to be more than ordinarily incoherent about politics. 2 Thus, aside from the basic problem of lack of equivalence in Hofstadter's definition, it would appear that the problems inherent in operationalizing the concept of "pseudo-conservative" itself further rule against the use of this definition in the present research. lIbid., p. 35. The original source is Adorno, et al., The Authoritarian-Personality, New York: Harper and Brothers:-1950, p:-676. 2Ibid., pp. 35-36. 45 The other view of the Extreme Right which falls into this category is one given by the Marxist historian Aptheker in his discus­ sion of McCarthyism: LMcCarthYiS~ •.. is the name of an ultra-reactionary, a fascist, political movement whose main stock-in-trade is anti-Communism .•• McCarthyism is American fascism. l Although "anti-Communism" is a recurrent theme in the expressions of Extreme Right groups, it is not sufficient as the central criterion for defining the Extreme Right. There are many individuals and groups in this country who are anti-Communist, even militantly so, who are not Right-Wing Extremists. The pitfall of defining the Extreme Right in terms of "anti-Communism" is that of "mistaking the symptom for the disease. fl Notwithstanding the fact that anti-Communi sm is one of the many attitudinal characteristics of the Extreme Right, the label of "Communist ll is also a convenient and culturally approved brush with which to tar many other elements in the society, in order to legitimate the extremist's opposition to them. An additional point in Aptheker's definition may be referred to here, although it actually relates to the problem of phenomenological misinterpretation. The point is that Aptheker's definition of McCarthyism as American fascism might be difficult to document. To be sure, McCarthy's demagoguery was reminis­ cent of Hitler's, and much of McCarthy's overt following came from those who had supported fascist and semifascist movements in the thirties and forties. However, as Richard Rovere points out, there are crucial lHerbert Aptheker, The Era of McCarthyism, New York: Marzani and Munsell, Inc., 1962, pp. I4b, 152-.­ 46 points at which contrast between Hitler and McCarthy is more striking than comparison: Hitler had a program for the coming millennium; McCarthy had no program for tomorrow morning. Hitler's aim was to win control of the machinery of state; it is still arguable as to whether McCarthy was up to anything of quite this magnitude. He never encouraged direct action by his followers; he did not organize uniformed groups or even raggle-taggle street fighters. Politically, he never tried to organize outside the existing party structure, and there are reasons for supposing that he never intended to do so.l Commenting on McCarthy's political philosophy, Rovere writes: He was not, for example, totalitarian in any significant sense, or even reactionary. These terms apply mainly to the social and economic order, and the social and economic order didn't interest him in the slightest. If he was anything at all in the realm of ideas, principles, doctrines, he was a species of nihilist; he was an essentialy destruc­ tive force, a revolutionist without any revolutionary Vision, a rebel without a cause. 2 Thus, at least to the extent that McCarthyism may be regarded as an example of Extreme Right expression, the uses of the terms "anti- Communist" and"fascistll would appear inappropriate in defining the phenomenon under study. 3. Four additional definitions may be considered under the head­ ing of "vague or obscure referents. 11 By this is meant the introduction in the definiens of substantive examples which are not supported by reference to a reality situation, or the use of vague and general concepts which are not clearly defined or established in prior or lRichard H. Rovere, Senator Joe McCarthy, Cleveland, Ohio: The World Publishing Company (Meridian~oks), 1960, p. 19. 2Ibi~., p. 8. 47 subsequent discussion. The first definition to be considered in this section is one advanced by Daniel Bell. He describes the Extreme Right as an example of: • • • protest movements . • • LO!7 new divisions LWithin the societl7 created by the status anxieties of new middle class groups • •• .1 Thus, Bell seems to be defining the Extreme Right in terms of "status politics." This concept is not uncommon in political sociology, and has been advanced by writers such as Bell, Richard Hofstadter, and S. M. Lipset. Lipset has defined the term as follows: Status politics ..• refers to political movements whose appeal is to the not uncommon resentments of indi­ viduals or groups who desire to maintain or improve their social status. 2 This concept may have considerable utility as an "orienting" concept for the present investigation, since it would appear to be relevant to the concept of "status crystallization." A major problem in Bell's definition, however, is created by the use of the terms "new divisions" and "new middle class groups." The nature of these new divisions within society is not made clear, nor is there any evidence given to indicate that they really exist. If such heuristic constructs are to have utility in scientific research, there must be some way of "grounding them in reality." Since this has not been done in Bell's discussion, his definition would not appear suitable for the purposes of the lDaniel Bell, "Interpretations of American Politics," in Bell, op. cit., pp. 5, 29. 2S. M. Lipset, "The Sources of the 'Radical Right' ," in Bell, op. cit., p. 168. 48 present investigation. Moreover, Bell's statement does not comply with the logical conditions of a definition, since he confounds defini­ tion with hypothesis. That is, the statement, "created by the status anxieties of new middle class groups," seems to be an implicit hypothesis. Riesman and Glazer view Right-Wing Extremism as an expression of the discontent of what they call the "discontented classes."1 This discontent centers in the new middle class which has evolved as a result of fifteen years of prosperity. This definition suffers from the same shortcoming as Daniel Bell's. That is, no elucidation of what the "discontented classes" are is given, nor is any evidence given to indicate that such classes really exist. Lipset, in his paper in the Bell volume, makes no attempt at a formal definition of the Extreme Right. However, he does state: These L!ight-win~7 groups are but one more manifestation of American political and moral activity, much like the popular atte~ts to ban liquor, gambling, or immorality in comic strips. This definition could have value in studying Right-Wing Extremism, since it suggests that attention should be focused on the moralistic elements in this type of political behavior. However, its application would require a prior definition of the vague term, "political and moral activity," which is not given in Lipset's discussion. IDavid Riesman and Nathan Glazer, liThe Intellectuals and the Discontented classes," in Bell, op. cit., p. 6L 2S. M. Lipset, op. cit., p. 181. 49 In another paper, published elsewhere, Lipset defines the Extreme Right, specifically McCarthyism, as an: • • • irrational protest ideology ,[espoused by discontenteg ••• declining 'liberal' classes living in declining areas. l In addition to the questionable utility of the term "irrational" (Le., against what criteria is the rationality of the Extreme Right to be judged?), the reference to "declining 'liberal' classes" poses the same problem as the previous definitions considered in this section. That is, there is little or no evidence given to indicate that these social groups actually do exist and can be clearly identified. In summarizing this review of the literature on definitions of the Extreme Right, the following points may be made. 1. Very few explicit definitions of the Extreme Right have been made to date. 2. A number of implicit definitions may be drawn out of the existing literature on the Extreme Right. 3. Many of these definitions have a potentially valuable "orienting" function, in that they direct attention to certain sig­ nificant characteristics of the Extreme Right. 4. For a number of reasons, however, none of the definitions considered, in their present form, adequately define the Extreme Right for the purposes of the present investigation. Either because of IS. M. Lipset, "Social Stratification and 'Right-Wing Extremism' ," British Journal of Sociology, 10 (December, 1959), p. 28. -----------------------------~--~--=--=~====".... ....._--------------------------------------------------------­..... .­ 50 ----------------------------~-,---.- ---"""!II! phenomenological misinterpretations, lack of equivalence between the definiendum and the definiens, or the use of vague or obscure refer­ ents, none of the definitions considered could be readily expressed in operational terms. In view of these considerations, the next task in this chap­ ter must be to define the Extreme Right in terms that will permit representation by operational indicators. Toward ~ Definition ~ the Extreme Right A basic assumption of the present investigation is that the Extreme Right is essentially political in nature. The fact that this country has a two-party political system tends to obscure the possibil­ ity that a variety of political attitudes can and, in fact, do exist. In countries having a multi-party system, such as France, institution­ alized channels of expression for various types of political attitudes are manifest. Consequently, a wide range of political attitude is laid open to investigation by an equally wide range of political behavior. In the American political system, on the other hand, splin­ ter parties, operating through alternative channels of political ex­ pression, have not been successful in gaining direct political power, and are either short-lived or doomed to obscurity. The necessity for the two dominant political parties to compete for the support of a heterogeneous electorate tends to reduce the difference between them, and institutionalized channels of extremist political expression are forestalled from developing within the existing political framework, ---- -_.=====~=~=----------- ---r 51 since they 'Houle. be antithetical to the necessarily "middle of the road" posture of both parties. l The militancy and millenarianism of the Extreme Right reflect the "crises of legitimacy" to "Thich this element in the American political system is exposed. Frustrated in gaining access to formal and institutionalized political power, the Extreme Right has pinned its hope on a policy of militant and uni­ lateral action designed to bring about desired changes in the political system in a not-tao-distant millennium. 2 Further evidence of the political nature of the Extreme Right lies in the fact that most of the changes .Thich extremist groups advo­ cate are designed to be implemented primarily through the political framework of society. In addition, the Extreme Right seeks alignment with institutionalized conservative political forces. This is evi­ denced, for example, in Right-Wing support for Senators Goldwater, Tower and Eastland. IDaniel Bell, in "Interpretations of American Politics," op. cit., pp. 4-5, discusses this situation as follO'l{s: ­ -- "Perhaps the most decisive fact about politics in the United States is the two-party system. Each party is like some huge bazaar, with hundreds of hucksters clamoring for attention. But while life within the bazaars flows freely and licenses are easy to obtain, all trading has to be conducted within the tents; the ones who hawk their wares outside are doomed to few sales. This fact gains meaning 'when vle consider one of the striking facts about American life: America has thrown up countless social movements, but fevT political parties; in contradiction to European political life, few of the social movements have been able to transform themselves into political parties." 2This problem of inability to gain formal political power and recognition is discussed in S. M. Lipset, Political Man, Garden City, He'd York: Doubleday and Company, Inc., 196n, Chapter III, "Social Conflict, Legitimacy, and Democracy," pp. 77-96. --- .__. ----­ 52 In sum, then, the Extreme Right may be regarded as an as yet non-institutionalized political ideology. The use of the term "ideol­ ogy" here is in the sense in which it is defined in Chapte:r; I--an organization of attitudes, opinions and values about society. In the literature on social movements, ideology is usually distinguished from the specific program of change that a group advocates, as, for example, in the following statement from Turner and Killian: At the outset we may distinguish the particular program of change that a movement advocates from the conception of society through which it justifies that program. The latter we shall call the ideology of the movement . . . .1 This conception of ideology will thus permit a study of the attitudinal characteristics of the Extreme Right without going into a detailed analysis of the content of the program of the Right-Wing movement. The advantage of this position becomes apparent when we consider that the Extreme Right is composed of ~ groups advocating many different (and sometimes opposing) programs of change. Aside from the very vagueness of most of these programs, there exists no one effort to inaugurate changes that is universal to the Extreme Right per ~. Thus, studying the Extreme Right under the full rubric of a Il social movement" analysis would necessitate analyzing the organization and programs of the totality of groups constituting the Extreme Right--a monumental task vThen one considers how many there are. lRalph M. Turner and Lewis M. Killian, Collective Behavior, Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1957, p. 331. 53 The next phase of this investigation will be to establish, by the process of "intuitive induction,"l a definition, capable of being operationalized, which will delimit the Extreme Right from other phenomena. To this end, we shall attempt to discern a number of the outstanding characteristics of the Extreme Right. The most extensive survey of such characteristics published to date is that of Ellsworth and Harris. 2 To those which they have identified, I have added a number drawn from my own surveys of available literature.3 The initial investigation took the form of a content analysis of this material to determine what the Extreme Right advocated or supported and what they lBasically, this process consists of making empirical observations of a phenomenon, in order to arrive at a basic knowledge of it. For a discussion of this form of reasoning, see Morris R. Cohen and Ernest Nagel, op. cit., pp. 273-275. 2Ralph E. Ellsworth and Sarah M. Harris, ~. ~. 3A comprehensive list of more than 1800 Right-Wing Groups is given in The First National Directory of "Rightist" Groups, Publica­ tions and Some Individuals in the United States, fourth edition, Los AngeleS:- Alert Americ~Association, 1961. The major sources investigated in the present survey are speeches made by acknowledged Extreme Right spokesmen; pamphlets published by America's Future, Inc., New Rochelle, New York; pamphlets published by Bible Recordings, Inc., Baltimore, Maryland; transcripts of testimony given before the House Committee on Un-American Activities; radio broadcasts and newsletters of The 20t~Century Reformation Hour, Rev. Carl McIntire, Director, Collingswood, New Jersey; ilDawn,~monthly newspaper published by Independence Foundation, Inc., Portland, Indiana; the "Blue Booklf and monthly bulletins published by The John Birch Society, Inc., Belmont, Mass.; Frank J. Donner, The Un:Americans, New York-:-­ Ballantine Books, 1961; Gene Grove, Inside the John Birch Society, Greenwich, Conn.: Fawcett Publications, 1961; Telford Taylor, Grand Inquest, New York: Ballantine Books, 1961; Richard Vahan, The Truth About The John Birch. Society, New York: McFadden Books, 1962; Donald Janson and~nard Eismann, The Far Right, op. cit. 54 opposed or rejected. As might be expected, a considerable overlapping of the attitudes expressed by different groups and individuals occurred. However, some twenty-seven relatively distinct attitudes, mostly nega­ tivistic, appeared manifest. A preliminary inspection of these attitudes suggested that they fell under four general headings. l These hypothesized categories, and the attitudes constituting them, are as follows: T. Attitudes regarding government: Opposition to strong central government. 2 Belief in strong government and leaders, but at the local level. Dissatisfaction with the United States Supreme Court. Opposition to the Federal Reserve System. Conviction that there is corruption in government. General distrust of the federal government. Opposition to increased government spending, higher taxes. Opposition to metropolitan government. Opposition to urban renewal. II. Attitudes regarding international relations: Opposition to foreign entanglements. Dedication to an ltAmerica Firstlt approach. Opposition to the United Nations. Opposition to foreign aid, Point Four Programs, NATO, etc. IThese hypothesized groupings will be tested by means of a cluster analysis, to be presented in Chapter IV and Appendix B. 2It must be borne in mind that this opposition is selective. The Right-Wing Extremist1s conception of the function of the state is not unlike that of the classical liberal, who saw the state as a pro­ tector of property and a preserver of order, much like a night watchman, rather than as an entity which imposed positive obligations upon individuals. Thus, the extremist's major opposition to the federal government is in those areas where he is told what he should do (e.g., desegregation). On the other hand, Right-Wing support is given to strong government in matters pertaining to security (e.g., congression­ al investigating committees). For an interesting discussion of this conception of the state, see Harry K. Girvetz, From Wealth to Welfare: The Evolution of Liberalism, Stanford, California: Stanfor~University Press, 1950, p~ 68-78. ---------- 55 III. Attitudes regarding modern social principles: Opposition to modern education. Opposition to racial integration. Suspicion of international collectivism (e.g., the Common Market). Militant anti-Communism. Political cynicism. l Opposition to "social gospel" Protestantism. IV. Attitudes regarding modern social structure and operation: 2 Opposition to socialized medicine. Opposition to collective bargaining. Support of "right to work" proposals. Support of It free enterprise." Opposition to "full employment." Opposition to the "welfare state." Opposition to federal aid to health and education. Suspicion of modern "progressive" innovations.3 The main theme running through these attitudes appears to be a general opposition to certain forms of "collectivism." This view is strengthened by an inspection of statements by some of the leaders of Extreme Right groups. Robert Welch, for example, devotes all of section two of his ~~ to the "cancer of collectivism,,,4 Leslie Fleming, lThis attitude refers primarily to the attitudes and motiva­ tions of politicians and to the operation of the political system in general rather than to any specific political issues. 2A distinction should be drawn between this category and the preceding one regarding "modern social principles." The latter refers primarily to a generalized attitudinal framework through which the Right-Wing Extremist regards his so·ciety. "Modern social structure and operation," on the other hand, refers more to specific programmatic policies, particularly those of the contemporary liberal state. 3E.g., fluoridation, psychiatry, pastoral counselling, mental health programs, mental hospitals, etc. 4Robert Welch, The Blue Book of the John Birch Society, sixth printing, 1961, Section Two, ff But Let'sLook Deeper • . . ,n pp. 41-55. - .._~-- - .. - ..."- -­ 56 --.- -~-.... --~-_.. _._~------- ---­ President of the Oregon John Birch Society, states that the program of the Society is tl ... individuality opposing collectivism,tll and R. K. Scott, President of America's Future, Incorporated, writes of his organization's ••• unbending opposition to all forms of collectiv­11 ism including the communist conspiracy.tl2 The kinds of collectivism opposed by the Extreme Right are primarily Communism and the type of contemporary liberalism practiced in the United States. With a few notable exceptions (such as the Tennessee Valley Authority and the Bonneville Power Administration), contemporary American liberalism does not tend toward governmental ownership and management of the means for the production and distribution of goods. Rather, the governmental role in the economy tends more toward increasing control and regulation, while relying, for the most part, on a market system economy and its corollaries of private property and private profit. In addition to this control or regulatory aspect, Extreme Right opposition also extends to an underlying philosophy of contempo­ rary American liberalism--one which may be called the tlameliorative" or welfare function of the state. This function refers not only to an lExcerpted from the transcript of a speech delivered by Leslie Fleming, President of the Oregon John Birch Society, at Clark College, Vancouver, Washington, February 26, 1962. Published in Pace: The Emerald Features Supplement, University of Oregon, Thursday, May 17, 1962, under the following affirmation: "This is the views (sic) of myself and the John Birch Society. LESLIE FLEMING (signed), Box 3174, Eugene, Oregon, May 5, 1962." 2Quoted from a personal communication received from R. K. Scott, President of America's Future, Incorporated, 542 Main Street, New Rochelle, New York, March 12, 1963. 57 acceptance of responsibility on the part of the state for the basic well-being of its members, but also to systematic planning aimed at providing fuller lives for all members of the state. Extreme Right antagonism to this kind of welfare function, or "quasi-paternalism," reflects not only a practical opposition to taxation--a requisite for the financing of welfare and aid programs--but also a belief that man will not be motivated to work unless he suffers deprivation. This belief has its roots in the classical liberal view of the psychological nature of man. l According to this view, man is motivated by self- interest ("egoismlt ) or pleasure ("hedonismlt ). From this, it follows that purposive activity must somehow be induced--unless some enticement is offered by way of pleasure or advantage, man will remain apathetic and inert ("quietismll ). However, man is also basically rational (II intellectualism" ). Given a choice of alternatives, reason will balance the quality of pleasure or pain involved, and conduct will follow that course of action which carries the greatest pleasure or the least pain. In order to motivate man to purposive activity, the pains of deprivation must therefore exceed those of work. If, on the other hand, public welfare and assistance programs permit man to con­ surne without toiling, the pleasures thus derived will be so great as to encourage perpetual indolence. lFor a full discussion of the nature of this view and its consequences for classical liberal thought, see Harry K. Girvetz, op. cit., esp. pp. 7-27 and 28-42. 'T 58 It is against the collectivistic or socialistic principles of contemporary American liberalism that the main force of Extreme Right expression is directed. Since a great deal of the policy of the present liberal government is based on these principles, it is understandable that the federal government should be a primary target of the Extreme Right. However, tendencies toward socialism at any governmental level are decried. At the state level, public (i.e., governmental) management of state supported institutions (particularly educational institutions) is closely scrutinized. At this level, how­ ever, extremists must walk a very thin line, since IlStates' Rights" is a battle cry against the extension of federal government power. Socialistic tendencies at the local government level (such as urban renewal--a joint federal-local project--and metropolitan government) are also opposed by the Extreme Right, primarily for tactical reasons. This is understandable in light of the extremist conviction, articulated by spokesmen such as Carl McIntire, Billy Hargis, Fred Schwartz and W. Cleon Skousen, that the fight against collectivism must begin at the local or Ilgrass roots" level. At the international level, the greatest source of concern to the Extreme Right is the United Nations, and America's participation in it. In the eyes of the extremist, that body is collectivism incarnate, and practically every group on the Extreme Right has expressed vehement opposition to it. American aid to foreign countries is also protested, being regarded in the same light as social welfare on the domestic scene. -- .__._-.--~."-.-.-._... ~ ...~'~ ._. --"""!!!!I 59 The militant anti-Communism of the Extreme Right can also be understood under this frame of reference, since the Right Wing con­ siders that Communism is either the polar extension of socialism, or, what is worse, that socialism is merely Communism by another name. This approach to the anti-Communism of the Extreme Right also explains the bias against intellectuals, liberals, modernists and progressives. Since these are the main groups advocating socialism, it is understand­ able that they be labelled as "Communist," and treated as such. By the same token, it is natural that most of the modern philosophies and programs advocated by these liberal groups be suspect in the eyes of the Extreme Right. Thus, "progressive education," a modern innovation, is regarded by the extremists as a means through which the socialistic, liberal elements in the society are attempting to indoctrinate the younger generation with their insidious beliefs. Higher education, primarily in the liberal arts and the social sciences, is also a target of the Extreme Right. Similarly, modern innovations such as pastoral counselling, mental health programs, psychiatry, mental hospitals, and even fluoridation are all regarded as part of a moder~collectivistic conspiracy. What the extremist proposes as an alternative to this collectivism is a political state in which the individual is the final arbiter of truth. A consistent and militant theme in the agitation of the Extreme Right is for the individual to awake, to arouse himself from the apathy to which he has been led, to be aware of the dangers which surround him, and to take action against those 60 -----:::.~ --:~ .~-_:-::=:-_~~~:_=~.;...-:-.-=-::....=----= ----_._-------------. --~--- dangers, as well as against the leaders who have brought him to this pass. The notion of an enslaved people dominated by corrupt leaders is a recurrent one in Extreme Right expression. The Right-Wing Extremist views his world in terms of a simplistic, black-and-white dichotomy. In his eyes, an impersonal, rationalized, complex, techni­ cal, bureaucratic "society" is destroying the simple virtue of man as an "individual. "1 Given this picture of the extremist "world view," the philos­ ophy underlying the Extreme Right may be discerned. The essence of this philosophy lies in the Right-Wing Extremist's atomistic, as opposed to organicist, view of the relationsip between man and society. "Atom­ ism" refers to the conception that the whole is nothing more than the sum of its parts. "Organicism," on the other hand, refers to the con­ ception of the whole as something more than the sum of its constituent parts--a reality sui generis. Applied to society, "atomism" is the view that individuals make society; "organicism" is the view that society makes individuals. 2 Girvetz has summarized the atomistic philosophy as follows: lThis extremist "world view" has been recognized by several of the contributors, in their 1962 articles, to Daniel Bell (ed.), The Radical Right (The New American Right expanded and updated), Garden City, New York: Doubleday and Company, Inc., 1963. See, for example, Daniel Bell, "The Dispossessed," Pl'. 1-38; David Riesman, "The Intel­ lectuals and the Discontented CJ.asses: Some Further Reflections," Pl'. 115-134; Talcott Parsons, "Social Strains in America: A Post­ script," Pl'. 193-199. 2Theories of society based on these polarities have character­ ized almost the entire history of social thought. Although their roots go back into classical philosophy, the development of atomism and -----'--_.,_._----,~-- --- --------_.. - -"­ I I 61 J!I The classical liberals, in accord with their atomistic outlook, regarded social institutions as the handiwork of pre­ existing individuals whose characteristic mental and emotional endowments antedate the social arrangements into which these individuals enter. Even rights are often regarded as natural, that is to say, as antedating the state. Social arrangements affect individual human nature only superficially. They are additative and artificial, and their importance is largely negative, an importance which consists mostly of removing obstacles which might prevent individuals from achieving com­ plete self-expression. The relationship between individuals and society is an external one; the individual with his various propensities and faculties is given, and society is an arrange­ ment of convenience, whereby faculties operate more effectively and propensities are more likely to find fruition. To repeat social institutions are created by the fiat of self-contained individuals; they are instruments, even expedients, which the individual can employ or discard without fundamentally altering his own nature. l This quotation, although somewhat extenSive, is included in its entire­ ty because of its central importance to the thesis which is being developed in these pages. This atomistic philosophy, together with the classical psychological view of the nature of man discussed earlier, constitute the basic identifying characteristic of the Extreme organiclsm may be conveniently traced from their Eighteenth Century formulations. The former is evidenced in the writings of men such as Hobbes (1588-1679), Rousseau (1712-1778), Burke (1729-1797), Benth~ (1748-1832) and, in its Nineteenth Century specification, Spencer (1820-1903). Organicism may be traced through the writings of Locke (1632-1704), Hume (1711-1776), Blackstone (1723-1780), DeMaistre (1754­ 1821), DeBonald (1754-1840), Comte (1798-1857) and Durkheim (1855-1917). The empirical validity of such polarities is an academic problem, and one which need not concern us here. What is essential is the fact that the Right-Wing Extremist does view his society in these terms. That he does so has important implicatIOnS for the studY-of his attitudes and behavior, for, as W. T. Thomas has pointed out in his concept of It definition of the situation," what men believe to be true are true in their consequences. IGirvetz, op. ~., p. 23. 62 Right--a characteristic which will be referred to herein as "individu­ alism." Individualism is defined by Webster as: A theory or policy having primary regard for individual rights, specifically one maintaining the independence of individual initiative, action and interests ..•.1 This individualism is not unlike that discussed by Max Weber in his essay on the Protestant Ethic. 2 Specifically, it is epitomized in the Ascetic Protestant attitude that the success or failure of every man, both in this world and in the next, rests on his own individual initi­ ative. This is the attitude which finds expression in the Right-Wing Extremist's opposition to "assistance" programs and his abhorrance of the maxim, "from each according to his ability, to each according to his need." It is from the expression of this attitude by Right-Wing Extremists that many writers deduce the notion of "regression" as a dominant characteristic of the Extreme Right--a ,desire to return to the nineteenth century. In maintaining this principle of individualism in opposition to what has been earlier described under the heading of "collectivism," the Right-Wing Extremist is opposing a system which would not only take away the fruits of one's labors (through taxation) to distribute it to those who did nothing to earn it (through social welfare), but would also restrict (through government control) one's prerogative to act in his own best interest in all matters. lWebster's New International Dictionary of the English Language, second edition, unabridged, Springfield,JMassachusetts: G. and C. Merriam Company, 1935, p. 1268. ~ax Weber, The Protestant Ethic and the Spirit of Capitalism, translated by Talcott Parsons, New York: Charles Scribner'S Sons, 1958. 63 The form of individualism which is advocated by the Extreme Right is selective in nature. First, independence of initiative, action and interests would not be extended to all individuals or groups. Minority groups would be excluded, as would those individuals who did not share the Extreme Right's opposition to collectivism (such as mod­ ernists and liberals). Second, the infringement of individualism would be tolerated in those situations where intervention and/or con­ trol may be regarded as furthering the aims and interests of the Extreme Right. This explains the Extreme Right's Janus approach to the federal government. For example, while federal intervention in school desegregation is regarded as an infringement of States' Rights (and is thus Vigorously opposed), the granting of federal defence con­ tracts benefits the economic situation of the state, and the federal control accompanying such a form of intervention is overlooked. More­ over, strong government control is advocated by the Extreme Right in matters pertaining to the "police function" of the state--the protec­ tion of property, the preservation of order and the maintenance of security, both internal and external. ThUS, as indicated earlier, the Extreme Right supports the F.B.I. and congressional investigating com­ mittees. Toleration of collectivism is also extended to corporate "big business,tI which group the Extreme Right courts as a source of economic support. It may be said that the Extreme Right opposes collectivism by proposing individualism, except in those areas where collectivism can be rationalized as furthering extremist interests. 64 The foregoing discussion of the nature of the Extreme Right may now be summarized, preparatory to establishing the definition of this phenomenon for the purposes of the present study. Throughout this discussion, the concepts of "collectivism" and Ifindividualismlf have been used extensively. Unfortunately, no one word can completely express the complexities of a phenomenon such as the Extreme Right. However, it is hoped that the particular contextual meaning of these terms, as they are applied to this study of the Extreme Right, has been made clear in the preceding pages. By Ifcollectivism" we mean primarily the lIquasi-paternalisticlf welfare functions of twentieth century American socialism, and, on the international scene, Communism. Thus, Right-Wing opposition is directed primarily against government, whose function it is to administer aid and welfare programs. This opposition to government is not universal, however, since the Extreme Right toler­ ates and even supports certain functions of the state (such as the lI police functionlf and the maintenance of the lI market-place economy"). Corporate business, which, in the classical Iflaissez-faire" sense, may be regarded as detrimental to the individualism of entrepreneurial business, is also spared extensive Right-Wing attack, primarily because it is a potential source of economic support for Extreme Right groups. It has been proposed that the Extreme Right opposes this form of collectivism by advocating "individualism. 1f This latter term, as it is used in this investigation, also has a certain restricted meaning. This principle of individualism has its basis in the classical Ifatom_ istic ll view of society--the view that social institutions are subordi­ · -'~---._--~_._- ._---~----_._---- - --_. 65 nate to the individual. Thus, the Extreme Right tolerates those functions of the state that protect the interests and conditions of the individual. The individualism advocated by the Extreme Right is limited, however, since the right to act in one's own best interests would not be extended to all individuals. Those who would subvert what the Right-Wing Extremist regards as "the greatest good," (e.g., liber­ als and leftists), as well as the non-productive and shiftless members of society (e.g., most minority group members), would not be afforded the privileges of individualism. Based on the foregoing discussion, it is proposed that the Extreme Right be defined as follows: The Extreme Right is a militant and millenarian political ide­ ology, espoused by numerous Right-Wing groups and individuals, which maintains as an ideal the principle of "limited individu­ alism"; this principle being articulated as opposition to tlcollectivis~l in government, international relations, modern social principles, and modern social structure and operation. By focusing on the ideological aspects of the Extreme Right, rather than on individual groups of this persuasion, it should be possible to delineate more sharply the attitudinal indicators of this phenomenon. If, as it has been hypothesized in this chapter, the Extreme Right is an ideology related to the elements of "government," "inter­ national relations," modern social principles" and "modern social structure and operation," then attitudinal indicators used to measure this phenomenon should cluster into groups definable by these four dimensions. To this end, a cluster analysis of available attitudinal indicators of the Extreme Right, as well as an extremism score devel­ oped from these indicators, will be presented in the next chapter. 66 CHAPrER IV EMPIRICAL ANALYSIS OF EXTREME RIGHT ATTITUDES The Cluster Analysis Technique The technique of cluster analysis is one of several forms of correlation analysis developed primarily during the 1930's, all of which were based on the concept of "factor analysis" developed by Spearman in 1904,1 At this time, Spearman proposed a "tvlO-factor" theory, which held that all human abilities are basically dependent upon a factor of general mental ability (the Qfactor) and a number of specific ability factors operant in task-oriented situations (the! factor). In later decades, the hlO-factor theory gave way to Ilmultiple-factor" theories, such as those of Hotelling2 and Thurstone.3 Basically, factor analysis is a method of analyzing the intercorrelations between a number of related observations (or lCharles Spearman, IlGeneral Intelligence Objectively Deter­ mined and Measured, II American Journal of Psychology, 15 (1904), pp. 201-293. 2Harold Hotelling, "Analysis of a Complex of Statistical Variables into Principal Components," Journal of Educational Psychology, 24 (September and October, 1933), pp. 417-441 and 498-520. 3L. L. Thurstone, Multiple Factor Analysis: A Development and Expansion of the Vectors of Mind, Chicago: The University of Chicago Press,1947:" - -­ 67 variables) in order to assess whether, and to what extent, these observations are determined by a smaller number of basic "factors." The assumptions underlying factor analysis have been stated by Thurstone as follows: Factor analysis assumes that a variety of phenomena within a domain are related and that they are determined, at least in part, by a relatively small number of functional unities or factors. l Two considerations regarding this statement should be noted. First, the assumption that certain phenomena (or variables) are related does not imply or require they they are in fact related. The identification of any subsystems of variables which possibly cluster together is, in any case, one of the primary purposes of factor analysis. Second, if there are any underlying factors determining these phenomena, then it may be expected that variables with high "saturations" or "loadings" on anyone factor (!.~., variables which are highly associated with that factor) will be highly intercorrelated. This facilitates a second purpose of factor analysis--namely the selection, by means of analyzing the relationships between clustered variables, of a parsimoni­ ous measure of the dimension represented by these variables. A major advantage which recent multiple factor analysis methods have over simpler techniques is that they permit the assignment of different portions of the variation of individual variables to different factors, thus making possible a precise definition of the structure of each factor. By comparison, cluster analysis is a much lIbid.; p. 56. 68 simpler method of correlation analysis which seeks to sort variables into mutually exclusive systems or "clusters." Although one of the purposes of factor analysis is also the delineation of systems of variables, in cluster analysis membership in each cluster is confined to only those variables which are strongly related to that system. In factor analysis, as already indicated, different portions of the variation of each variable may be assigned to different factors, so that underlying factors may II cut acrossll a number of clusters. In this sense, it may be said that cluster analysis is a preliminary step to factor analysis. That is, cluster analysis is useful for isolating mutually exclusive clusters of variables, whereas factor analysis techniques go on to analyze both the inter- and the intra­ cluster relationships between variables in order to discover underlying factors. Cluster analysis, as such, does not make possible the isola­ tion of factors accounting for the variation of elements in different systems (~.~., factors which II cut acrossll different systems). Never­ theless, it is possible for the researcher, through a process of logical inference, to interpret the underlying element common to each system of variables isolated by the cluster analysis. In passing, it should be noted that there is one case where both cluster analysis and factor analysis will yield the same results. That is, when the vari­ ables under analysis are "factor pure rr (when each variable is related to only one factor), each cluster of variables, and the factors deter­ mining them, will be mutually exclusive. It may be appreciated, 69 however, that the probability of this occurrence in social science investigation is very remote. The cluster analysis method to be used in the present study is that of "correlation profile analysis," developed by Tryon. l This method has been chosen because it does not involve the complicated mathematical computations required for Hotelling's or Thurstone's methods, and because it is particularly suited to a large number of variables. In the following description of his method, Tryon refers to the dimensions underlying clusters as "operational unities. ll These are defined as those components which result in two or more variables showing the same pattern of correlation coefficients with all the other variables in an investigation. Two variables, A and B, are said to be wholly or partially determined by an operational unity if both correlate high with variable M, low with N, intermediate with 0, and so on throughout the other variables. In such a-case, clearly what is general in A and B is behaVing in an identical and unitary fashion. Correlation profile analysis is a simple method for discovering and grouping together variables which have identical patterns or profiles of correlations. 2 This method therefore involves the plotting of the intercorrelations between related variables as pictorial "profiles. 1I On the basis of these profiles, the lIoperational unities" underlying certain clusters (comprised of variables with congruent profiles) can be inferred. lRobert Choate Tryon, Cluster Analysis: Correlation Profile and Orthometric (Factor) Analysis !££ the Isolation of Unities in ~ and Personality, Ann Arbor, Michigan: Edwards Brothers, Inc., 1939. 2Ibid., p. 2. -:----:~.::::-=:-::::-~ 70 The Cluster Analysis: Hypotheses and Findings In Chapter III, it was proposed that the Extreme Right be defined in terms of four clusters of attitudes related to underlying operational qnities of government, international relations, modern social principles, and modern social structure and operation. The purpose of the present chapter is to assess whether these clusters of attitudes are empirically valid. This empirical analysis will take the form of a secondary analysis of data gathered in the Eugene- Springfield metropolitan area in the summer of 1959.1 This data was gathered primarily to study local political attitudes and behavior and does not contain any attitudinal items regarding "international rela­ tions." Therefore, this hypothesized element of Extreme Right attitudes cannot be tested in the present analysis. The following two hypotheses will be tested in this section: 1. All indicators of Right-Wing Extremist attitudes under study will be positively intercorrelated. That is, it is assumed that Right- Wing Extremists will be consistent in their responses to these items. 2. The correlation profiles of those attitudes hypothesized to be determined by the factors of IIgovernment," lImodern social principles ll and IImodern social structure and operation ll respectively will be more consistent with one another than with those of the remaining attitudes. l"Eugene-Springfield Metropolitan Area Study,t1 Institute for Community Studies, University of Oregon, 1959. The sample on which this study was based is discussed in Appendix A, pp. 153-156. 71 In the following presentation, the clustering of thirty attitudinal items from the Eugene-Springfield study vlill be analyzed. Such documentation arid proof as are required to illuminate this presen­ tation v,'ill be given in the text. For more extensive documentation and proof, the reader is referred to Appendix B, TlCluster Analysis Data,Tl pages 157-172. The identifying numbers for these attitudinal items refer to IB.\1 card columns, and indicate the order in v1hich the questions ""ere asked on the interview schedule. All of the follm·ring items "lere coded on a Likert-type scale of responses. Items 18, 20, 21, 22, 23, 24,25,26 and 27 "l'lere coded as follm-1s: "Strongly Approve,1f lfApprove,1f "Undecided," "Disapprove" ahd "Strongly Disapprove." In addition, these items contained three residual response categories of "Don't Care," "Don't Know" and "No Answer." The remaining i terns "rere coded as follmJs: "Agree Strongly," "Agree Somewhat," "Agree Slightly, If "Disagree Slightly," "Disagree Somewhat" and "Disagree Strongly," as well as residual categories of "Don't Kno"," and "No Answer." In order to compute the inter-item correlation matrix for the cluster analysis, responses to these items v1ere dichotomized by placing the "Strongly Agree (Approve)" or the "Strongly Disagree (Disapprove)" response (Vlhichever ,'ras applicable to the sense of the question) in the "Extremist" category, and the remaining re sponse s C\vi th the exception of "No Answer") in the "Non-Extremist" category. Thus, only those resp::mdents Vlith a "strong" attitude tOvrard the item in question "lere regarded as having given an extremist response. It should be noted that responding "stronS;lJ'" to anyone item does not classify the 72 respondent as being an Extreme Rightist. This is merely considered as being the response an extremist would make to the item in question. The parenthetical coded form of each item (which form will be used in subsequent reference to these items) indicates the direction of extremist response. On inspection, the thirty attitudinal items under analysis appear to be manifestly related to the three hypothesized clusters of Extreme Right attitudes as indicated below. Within each of the hypoth­ esized clusters, attitudinal items are organized according to sub­ headings. These sub-headings refer to the basic Extreme Right attitudes discussed in Chapter 111. 1 I • ATTITUDE S REGARDING GOVERNMENT Opposition to increased government spending, higher taxes: 20. What do you feel about increasing taxes to provide improved city services? (Disapprove taxes city services.) 25. What do you feel about spending more money on special education? (Disapprove money special education.) 27. What do you feel about increasing taxes to provide public kindergartens? (Disapprove taxes public kindergartens.) Opposition to metropolitan government: 21. What do you feel about annexation to the city of suburban areas? (Disapprove annexation suburban areas.) 51. It would be a good thing for the residents of both cities if Eugene and Springfield merged and became one city. (Disagree Eugene and Springfield merge.) lCf. supra, pp. 54-55. 73 Opposition to urban renewal: 18. What do you feel about urban renewal? (Disapprove urban renewal.) 36. The urban renel.;al program is one of the worst "tax and spend" enterprises yet devised by government planners. (Agree urban renel'lal worst "tax and spend.") 37. The urban renewal program represents interference and regimentation by government. (Agree urban renewal government interference.) 38. In urban renewal programs, the federal government helps the local community. (Disagree urban renewal helps community.) 39. Urban renewal is a much needed program for community better­ ment and development. (Disagree urban renewal needed program.) 40. The urban renewal program will make the community a better place in which to live. (Disagree urban renewal improve cormnunity.) II. ATTITUDES REGAHDING MODERN SOCIAL PRINCIPLES Opposition to modern education: 29. The public schools are not teaching the fundamentals as well today as they used to. (Agree schools not teaching funda­ mentals.) 30. Nowadays children get pampered too much in the public schools. (Agree children pampered in school.) 31. There is too much emphasis on cooperation in our public schools and not enough emphasis on competition. (Agree too much emphasis cooperation in school.) 32. ~~blic schools change too many children away from their parents' ideas. (Agree schools change children's ideas.) Opposition to racial integration: 45. The government in Washington should stay out of the question of whether white and colored children go to the same school. (Agree federal government out of school desegregation.) 74 47. If Negroes are not getting fair treatment in jobs and hous­ ing, the government in Washington should see to it that they do. (Disa.gree federal government role Negro fair treatment.) Political cynicism: l 48. Politicians spend most of their time getting re-elected or re-appointed. (Agree politicians spend time re-election.) 49. People are very frequently manipulated by politicians. (Agree politicians manipulate people.) 50. Most politicians in the community are probably more interested in getting known than in serving the needs of their constitu­ ents. (Agree politicians' interest get known.) III. ATTITUDES REGARDING MODERN SOCIAL STRUCTURE AND OPERATION Opposition ~ socialized medicine and/or health insurance: 2 44. The government ought to help people get doctors and hospital care at low cost. (Disagree government doctors and hospitaL) Opposition to collective bargaining: 22. What do you feel about labor unions in Eugene? (Disapprove labor unions.)3 Support of "~ enterprise": 23. What do you feel about public housing? (Disapprove public housing. ) 26. What do you feel about city-owned parking lots? (Disapprove r : city-owned parking lots.) l!.~., cynicism about politicians' self-interest. 2Since the indicator available here would not tend to discrim­ inate between related attitudes surrounding "socialized medicine" and federal "health insurance," this attitudinal. component was expanded from the.original one of "opposition to socialized medicine." 3Since a primary aspect of organized labor is the principle of collective bargaining, it was felt that an item regarding labor unions is a valid indicator of the attitude towards collective bargain­ ing. 75 33. A municipal power system is a form of socialism. (Agree municipal power socialism.) 34. The federal povler program in the Pacific Northwest is one of the best possible solutions to the region's economic prob­ lems. (Disagree federal power economic solution.) 46. The government should leave things like electric power and housing for private businessmen to handle. (Agree power and housing private business.) Opposition to "full employment": 43. The government in Washington ought to see to it that every­ body who wants to work can find a job. (Disagree federal government find jobs.) Opposition to federal ~ to health ~ ~ca~~~n: 41. If cities and towns around the country need help to build more schools, the government in Washington ought to give them the money they need. (Disagree federal government finance schools.) Opposition to modern "progressive" innovations: 24. ~~at do you feel about fluoridation of the community's water supply? (Disapprove fluoridation.) The limitations of secondary analysis become apparent in the above assignment of attitudinal indicators to clusters. What has been attempted in this procedure is to assign each item to the attitudinal component to which it most closely corresponds. Thus, urban renewal items were assigned to "government," since this program is most closely associated with the federal government. Items 20 (Disapprove taxes city services), 25 (Disapprove money special education) and 27 (Disapprove taxes public kindergartens), although they are concerned with different aspects of local politics, all involve the element of increasing local government expenditures and consequently of increasing taxes. Therefore, 76 they have been assigned to the cluster on "government," under the specific component of "opposition to increased government spending, higher taxes. 11 Item 47 (Disagree federal government role Negro fair treatment), while not directly related to racial integration, does deal with racial discrimination, and is thus included under the former heading. The first step in the cluster analysis was to layout a correlation matrix of the intercorrelations between the thirty vari­ ables under study.l On the basis of this matrix, the first of the hypotheses related to this cluster analysis (l.~., that all indicators of Right-Wing Extremist attitudes will be positively intercorrelated) seems to be borne out. Out of the 435 intercorrelations comprising the matrix of thirty variables, only 29 were found to be negative. Further­ more, reference to the sums of intercorrelations for each variable with all other variables (Lr in Table l8, page 159 of Appendix B) indicated that most of the variables are highly and positively intercorrelated. The main exceptions to this were items 21 (Disapprove annexation sub­ urban areas), 44 (Disagree government doctors and hospital), 47 (Dis­ agree federal government role Negro fair treatment) and 51 (Disagree Eugene and Springfield merge). Because of its extremely low correla­ tion with all other variables, item 51 was eliminated from the analysis lCf. Tryon, op. cit., p. 41. Further reference to this source willbe given iZlthetext as "Tryon" and cited in parentheses with relevant page numbers. The correlation matrix data for the present study is given in Appendix B, "Cluster Analysis Data," Table 18, page 159. 77 at this stage. Items 21, 44 and 47 were retained since, in spite of fairly low overall intercorrelation sums, they were all quite highly correlated with certain groups of variables. A possible exception was item 47, which may have to be eliminated later in the analysis. Preparatory to testing the second hypothesis (regarding the consistency of the profiles of variables posited to be related to the factors of "government," modern social principles" and "modern social structure and operation"), the variables under study were assigned to tentative clusters. The "B" coefficient ("coefficient of belonging") method developed by Holzinger was used for this purpose. l This pro­ cedure provides an objective way of clustering correlation coefficients and hence the configurations of variables which they represent. It is based on the assumption that the variables related to a given factor should have higher intercorrelations between themselves than with other variables under analysis, and involves the systematic introduction of variables assumed related to a given cluster. A more detailed description of this technique is given in Appendix B, "Cluster Analysis Data," pp. 158-166. lKarl J. Holzinger, Student Manual of Factor Analysis, Department of Education, University of Chicag~ 1937, Chapter III, section L See also Karl J. Holzinger and Harry H. Harman, Factor Analysis: A Synthesis of Factorial Methods, Chicago: The University of Chicago Press, 1941,-Pp. 23-24; Benjamin Fruchter, Introduction to Factor Analysis, Princeton, New Jersey: D. Van Nostrand Company, Inc., 1954, Chapter 2; Robert Choate Tryon, Cluster Analysis, op. cit., pp. 43-45, 47-48. -- --­ 78 On the basis of the "B" coefficient analysis, five tentative clusters of variables were discerned. The relationship of certain variables (indicated by"?" in the following presentation) to these clusters was questionable. These variables will be included in the tentative clusters but subjected to further correlation profile analysis. TENTATIVE CLUSTER A: 36. (Agree urban renewal worst "tax and spend.") 37. (Agree urban renewal government interference.) 38. (Disagree urban renewal helps community.) 39. (Disagree urban renewal needed program.) 40. (Disagree urban renewal improves community.) 18. (Disapprove urban renewal.) 23. (Disapprove public housing.) TENTATIVE CLUSTER B: 48. (Agree politicians spend time re-election.) 49. (Agree politicians manipulate people.) 50. (Agree politicians' interest get known.) 46~ (Agree power and housing private business.) 45? (Agree federal government out of school desegregation.) TENTATIVE CLUSTER C: 25. (Disapprove money special education.) 27. (Disapprove taxes public kindergartens.) 79 20. (Disapprove taxes city services.) 26. (Disapprove city-owned parking lots.) 21. (Disapprove annexation suburban areas.) 24. (Disapprove fluoridation.) TENTATIVE CLUSTER D: 43. (Disagree federal government find jobs.) 44. (Disagree government doctors and hospital.) 41. (Disagree federal government finance schools.) 34? (Disagree federal power economic solution.) 33? (Agree municipal power socialism.) 22? (Disapprove labor unions.) TENTATIVE CLUSTER E: 30. (Agree children pampered in school.) 31. (Agree too much emphasis cooperation in school.) 32. (Agree schools change children's ideas.) 29. (Agree schools not teaching fundamentals.) 45? (Agree federal government out of school desegregation.) 49? (Agree politicians manipulate people.) One variable--item 47 (Disagree federal government role Negro fair treatment)--did not appear to be related to any of the clusters deter­ mined by the "B" coefficient method, and was excluded from further analysis. The next step in the analysis was to calculate the mean correlation of each variable with each of the tentative clusters. 80 These mean correlations are presented in Appendix B, Table 21, page 168. These mean correlations were then plotted as mean correlation profiles (cf. Appendix B, Figure 3, page 169), by means of which additional relationships between the variables in the five tentative clusters were noted. The profiles for items 23 (Disapprove public housing) and 46 (Agree power and housing private business) showed considerable discrepancy from those of the other variables in their tentative clusters, and were therefore eliminated from the analysis at this stage. These two variables, along with the previously-excluded item 47 (Disagree federal government role Negro fair treatment), will constitute "residual variables" in the cluster analysis. On the basis of the mean correlation profiles, questionable items 22 (Disapprove labor unions), 33 (Agree municipal power socialism) and 34 (Disagree federal power economic solution) were retained in Tentative Cluster D. Finally, the profiles indicated that questionable items 45 (Agree federal government out of school desegregation) and 49 (Agree politi­ cians manipulate people) appeared to be related to both Tentative Clusters Band E. The mean correlation profiles for these clusters, when combined (cf.Appendix B. Figure 4, page 1701 verified this. For the balance of this analysis, therefore, Tentative Clusters Band E will be considered as one cluster. The last step in the cluster analysis will be to plot the final correlation profiles of individual variables according to the clusters established in the previous operation. To facilitate this, the correlation coefficients for items arranged by clusters were laid ,--, - .­ 81 out as in Table 22 of Appendix B (page 171). The final correlation profiles, in which the clusters are named according to their congruence with the initially hypothesized clusters (plus the observed cluster on "urban renewal") are presented in Figure 1, page 82 below. (cf. Tryon, page 5). For the most part, the second hypothesis being tested in this cluster analysis (i.!., that the correlation profiles of those attitudes hypothesized to be determined by the factors of "government," "modern social principles" and "modern social structure and operation" respectively will be more consistent with one another than with those of the remaining attitudes) seems to be verified. The major exception to this is the appearance of a fourth cluster of attitudes related to "urban renewal." Otherwise, most of the variables under analysis cluster together as initially posited. Although some variation in the profiles of the variables constituting each cluster are evident, the overall patterns of consistency between the clustered variables seem tangible enough to warrant the conclusion that the four clusters discerned by this analysis constitute four distinct attitudinal com­ ponents of the Extreme Right. To conclude this section, each of the four clusters discerned by this analysis will be discussed. CLUSTER I. ATTITUDES FEGARDING URBAN RENEWAL 18. (Disapprove urban renewal.) 36. (Agree urban renewal worst "tax and spend.") 37. (Agree urban renewal government interference.) -"""""'."",,"­ .­ _--r __ ----.- .,,---,,'=.- .:")SJ-.4£ '-¥_* -' -,-"':r;~'_~~ ~< ~ FIGURE 1. FINAL CORRELATION PROFILES OF VARIABLES BY CLUSTERS Cluster I Cluster II Cluster III Cluster IV Urban Renewal Modern Social Structure and Operation Government Modern Social Principles Residua Vari­ ables 3D -­ 3/ - - ­ -_. 32.···· ... 2'3 _. -'-'­ 4-8-­ 50 - --­ 3' -­ 4° ---­ 38 36 37 /8 21 ----­ 26········ 20-­ 2/ - --- Cl. I Cl. II 43-­ 4-4---­ +1 . 34­ _' ,~, __ ",;:-........... '-. _'_ n. ="..,... I , I \,.. '.., I .~ ·8 ·7 ·6 S ·4 ·3 .2 ./ , ~ ~ I I I ......':' I .;. I I I I I I . ""- ': . I r 0 I I ! i I ~ . '" j I v I I : ...v.o:1 I -.1 -.2­ -.3 -:4 -5 r r r CD f\) r 83 ~" 38. (Disagree urban renewal helps community.) 39. (Disagree urban renewal needed program.) 40. (Disagree urban renewal improves community.) This cluster is comprised of the items from Tentative Cluster A (excepting item 23--"Disapprove public housing"--Hhich became a "residual variable"), v:hich \jere initially hypothesized to belong to the cluster of items related to the factor of "government." Although the profiles for these urban renewal items are similar, in many re­ spects, to :'hose in the cluster regarding government, the extremely high intercorrelation beh!een these items '!!arrant the ir inclusion as a separate cluster. At least two reasons may be adduced to explain this high intercorrelation. First, these items are undoubtedly Guttman scale items, included in the metropolitan area study because they yield a high coefficient of reproducibility. Second, urban renei'lal is a specific issue vlhich, unlike many of the more general "ideological" issues raised in the other clusters, can evoke a definitive and con­ sistent response. This fact is probably intensified for the Fugene­ Springfield area, i'There a strong public opinion vas polarized on a local urban renewal project at the time the metropolitan survey was made. CLUSTER II. ATTITUDE S PEGARDING MODERN SOCIAL STRUCTUP-E AND OPERATION 22. (Disapprove labor unions.) 33. (f~ree municipal power socialism.) 34. (Disagree federal power economic solution.) 84 41. (Disagree federal government finance schools.) 43. (Disagree federal government find jobs.) 44. (Di sagree government doctors and hospitaL) The above items (from Tentative Cluster D) were initially hypothesized to be in this cluster. Four hypothesized items, however, failed to chlster in this group. Variables 24 (Disapprove fluoridation) and 26 (Disapprove city-owned parking lots) fell into the cluster on "government," and ,.,ill be discussed under that heading. Variable 23 (Disapprove pIJ.blic housing) and 1~6 (Agree pmier and housing private business) failed to cluster, and were classified as "residual vari­ abIes." Regarding item 23, it may be that the meaning of "public housing" was not understood by, or vIas not made clear to, the respond­ ents interviewed. The failure of item 46 to cluster raises two points. First, if llpublic housing" is not opposed by Right-1iJing Extremists in the Eugene-Springfield area (a possible interpretation of the finding regarding item 23), then the extremist who opposed government operation of electric power--as suggested by the clustering of items 33 (Agree municipal pOi.;rer socialism) and 34 (Disagree federal power economic solution)--1'7oIl1d have been "cross-pressured" by item 46, which is a "double-barreled" question concerning both electric pO'/7er and housing. The second point has reference to the unique situation regarding public pOvTer in the Eugene area. Although the Right-Hing Extremist may be ideologically opposed to municipal power in general, the Eugene extremist faces the fact that the electric rates charged by the Eugene 85 I'later and Electric Board (a municipal pOl-rer enterprise) are among the lowest in the nation. Thus, the failure of this item to cluster may reflect a conflict between the Right-Wing Extremist's ideology and his pocketbook. CLUSTER III. ATTITUDE S REGAP..DING GOVERNMENT 20. (Disapprove taxes city services.) 2L (Disapprove annexation suburban areas.) 24,. (Disapprove fluoridation.) 25. (Disapprove money special education.) 26. (Disapprove city-avmed parking lots.) 27. (Disapprove taxes public kindergartens.) This group is comprised of the variables from Tentative Cluster C. Items 20, 21, 25 and 27 were initially hypothesized to be in this cluster. Items 24 and 26 (from "modern social structure and operation") undoubtedly fall into this cluster on government because they concern local political issues. In Chapter II, it 1V'as suggested that Extreme Right opposition to local political issues was primarily tactical--the greatest opportunity for extremist elements to gain access to political power being at the local level. Therefore, it may be expected that the Right-l,'ling Extremist will attack the kinds of issues represented by items 24 and 26 on what is basically a political leveL CLUSTER IV. ATTI'IUDES REGAPJ)ING HODERN SOCIAL PRINCIPLES 29. (Agree schools not teaching fundamentals.) 86 30. (Agree children pampered in school.) 31. (Agree too much emphasis cooperation in school.) 32. (,~ree schools change children's ideas.) 45. (Agree federal government out of school desegregation.) 48. (Agree politicians spend time re-election.) 49. (Agree politicians manipulate people.) 50. (Agree politicians' interest to get knmln.) All of the above items were originally hypothesized to be in this cluster, which consists of Tentative Clusters Band E (with the exception of item 46 from Cluster B--"Agree pOl'1er and housing private business"--";lhich vlas classified as a "residual variable"). Item 47 (Disagree federal government role Negro fair treatment) washypothe­ sized as belonging to this cluster, but was eliminated during the analysis as a "residual variable." The failure of this item to cluster is probably due to the fact that racial tensions are less evident in the Pacific Northwest than in other areas of the nation. Thus, racism is not a central concern of Right-Hing Extremists in this area. In summary, the limitations of secondary analysis notwith­ standing, the empirical investigation of attitudes undertaken through this cluster analysis would appear to provide some validation for de­ fining the Extreme Right in terms of an ideology opposing "collec­ tivistic fl tendencies in distinct areas of the social system. The existence of separate extremist attitudinal "sets" related to "govern­ 87 ment," "modern social principles" and "modern social structure and operation l1 has, to a large extent, been verified by this analysis. The Right-vJing Extremism Score In developing a scale from the twenty-six attitudinal items discussed in the previous section, a simple scoring procedure was used. Likertl a~d Edwards2 have observed that the simple assignment of inte­ gral weights to response categories correlates highly with more com­ plicated scoring procedures. For this study, therefore, "Extremist" responses were assigned a score value of 1, and "Non-Extremist" responses a score value of O. A total score for each respondent in the sample was obtained by computing the sum of his scores for the twenty-six individual items. Before presenting the Right-Wing Extremism scores obtained by this method, the problem of non-response to the relevant attitudinal items should be discussed. In testing the relationship between Right- Wing Extremism and status crystallization, an attempt will be made to , ,·!jl control on all three relevant status variables (income, occupation, education) simultaneously. This will require eight comparisons of the proportion of Right-Wing Extremists for both crystallized and uncrys­ tallized status categories. In order to fill the sixteen cells of the lRensis Likert, "A Technique for the Measurement of Attitudes,11 Archives ~ Psychology, 140 (June, 1932). 2Allen L. Edwards, Techniques of Attitude Scale Construction, New York: Appleton-Century-Crofts, Inc .-,-1957, Chapter 6, "The Method of Summated Ratings," pp. 149-171. 88 table required for this test, it will be necessary to retain as many cases as possible in the sample under study. To this end, "Don't Knmv" and "No jmswer" categories have been retained as non-extremist responses, and only the "Strongly Agree" (or "Strongly Disagree") categories have been used as extremist responses. One plausible interpretation of non-response to these attitudinal questions is that it in fact represents a "true ll extremist attitude. That is, refusal to answer may reflect a reluctance on the part of the respondent to express what may be considered by others a "socially unacceptable" (i.e., extremist) response. l In the present research, however, the decision has been made to "lose" any potential extremists represented by non-response, for several reasons. In the first place, most of the non-response by individuals consists of one or two randomly distrib­ uted "No Answer" responses. Therefore, non-response could affect only a few of the scores immediately bordering the "Non-Extremist" side of the cutting point bet~een extremist and non-extremist scores (i.e., if a respondent has already obtained an extremist score without the aid of "No AnsVTer" responses, his position v7ill not be affected by such non-response). Second, if the null hypothesis to be tested (that there is no difference between crystallized and uncrystallized status groups with respect to Right-yling Extremism) is true, then "non-response" (assumed to indicate an extremist response) may be considered to be lFor a discussion of this interpretation of non-response see Hans Zeisel, Say It With Figures, revised fourth edition, New York: Harper and Brothers, 1957, pp. 59-60. 89 ---,.--------­ randomly distributed in both crystallization groups, as is the overt "Extremist" response. Therefore, non-response will not affect the percentages to be compared. Finally, if the null hypothesis is false, and the group characterized by uncrystallized status does contain a higher proportion of Right-Wing Extremists than the crystallized group (this hypothesis will be fully developed in Chapter V), classifying non-response as non-extremist will not bias the results in favor of this hypothesis, since more "potential ll extremists will be lost in the uncrystallized than in the crystallized group. In fact, such a pro­ cedure will actually have the effect of "loading" the outcome in favor of the null hypothesis. In sum, while the amount of bias introduced by retaining non-response in the sample may be considered as negligible, this method of retaining it, while serving to preserve sample size, will not operate to bias the results of the investigation in favor of the hypothesis to be tested. The Right-Wing Extremism Scale, constructed by summing individual scores on twenty-six attitudinal items, is presented in Table 8 below. For the sample of 673 cases for which complete data on occupation, income and education was available (~.~., data for the status crystallization variables discussed in Chapter II), a range of scores from 0 to 22 was obtained. The higher the scale score, the more consistently IIExtremist ll were the individual's responses. The final problem to be addressed in this chapter is that of arriving at a cutting point on this scale which will distinguish extremists from non-extremists. Although no definitive estimation 90 TABLE 8. FREQUENCY DISTRIBUTION OF SCORES ON TEE RIGHT-WING. EXTREMISM SCALE Score Number 0 157 1 125 2 95 3 90 4 63 .. 5 36 6 32 7 20 8 20 9 7 10 4 11 9 12 1 13 5 14 3 15 1 16 2 17 18 19 1 20 21 1 22 - 1 Total: 673 of the proportion of Extreme Right supporters (i.~., persons who hold attitudes consistent with the Extreme Right ideology, ~ only members of Extreme Right groups) is available, a certain amount of evidence exists that suggests a cutting point that will divide extremists from non-extremists in the proportion of about one to four. l M examination IFor example, see Martin Trow, "Small Businessmen, Political Tolerance, and Support for McCarthy,1l American Journal of Sociology, 64 (November, 1958), pp. 270-281; Charles H. Stember,-"Anti-Democratic Attitudes in America: A Review of Public Opinion Research,1l Bureau of Applied Social Researc~, New York: Columbia University, 1954 (mimeD: .--..------.." __.. _~~. _.,_.. __..~ _.2._L.2 1 ---...., 91 of Table 8 indicates that something approximating a "natural break ll exists between scale scores 4 and 5, the frequency for score 5 (36 cases) being a drop of nearly forty-five per cent over the frequency for score 4 (63 cases). For the purposes of this study, therefore, IlRight-TfJing Extremist ll will be that group receiving scores of 5 or higher on the "Right-Wing Extremismll scale. This group represents twenty-one per cent of the sample under study, a proportion which is consistent with the findings of other investigations. Thus far in the present stUdy, the independent and dependent variables, respectively, of status crystallization and Right-Wing Extremism have been defined and operationalized. It nOVl remains to develop and test a hypothesized causal relationship bet",reen these variables. The development of such a hypothesis will be the subject of the next chapter. graphed), p. 52 (cited in Trow, ibid., p. 271); Hanan C. Selvin and Harren 0. Hagstrom, "Determinants of Support for Civil Liberties,1l The British Journal of Sociology, 11 (March, 1960), pp. 51-73; Fred J. Cook, "The Ultra"'S": Aims, Affiliations and Finances of the Radical Right," The Nation, June 30, 1962, p. 571. 92 CHAPTER V STATUS CRYSTALLIZATION AJID RIGHT-vITNG EXTREMISM: DEVELOPMENT OF THE HYPOTHESIS Problems Engendered by Lm.". Status Crystallization In a number of studies using the status crystallization model, it has been pointed out that individuals with inconsistent status sets face certain social and psychological difficulties. Jackson discerns the following: unsatisfactory social relationships, unstable self- images, rewards out-of-line with investments, and social ambiguity.l This section will be concerned with a review of the findings related to these phenomena. Lenski, in his 1956 study, tested the following hypothesis: • persons vIi th a low degree of status crystallization are more likely to be subjected to disturbing experiences in the interaction process and have greater difficulty in establish­ ing rewarding patterns of social interaction than others. 2 This hypothesis, vlhich was borne out in Lensld f s study, was tested through observation of behavior and attitudes in the areas of associ­ lSee Elton F. Jackson, "Status Consistency, Vertical Mobility and Symptoms of Stress," unpublished Ph.D. dissertation, University of Michigan, 1960, and "Status Consistency and Symptoms of Stress," American Sociological Review, 27 (August, 1962), pp. 469-480. 2Gerhard Lenski, "Social Participation and Status Crystalliza­ tion, II American Sociologic,al Review, 21 (August, 1956), -p. ~·59. ~ :1, , " :~. 'r·' ,,'J "II!I :'II~ " ,'~~ 'l 93 ational or interactional activities, particularly those related to participation in voluntary associations. He found respondents with low crystallization to be generally non-participants in voluntary relationships, to have given up long standing voluntary ties, and to have established and maintained ties for non-sociable (i.e., ulterior) rather than for sociable reasons. Zaleznik, Christensen and Roethlis­ berger, in developing their "Theory of Social Certitude," also empha­ sixed this aspect of social relationships. In discussing the reasons why people seek status congruence (i.e" status crystallization) they state: To us it is the need on the part of man for clear and unambiguous relationships. This is the meaning and striving for status congruence. When this condition is realized, a condition of social certitude is felt. Under this condition members of a group know where they stand in relation to each other. They know how they should behave and are expected to behave toward others. Little anxiety exists in such relation­ ships except that engendered by the need to live up to what is expected. l Peter Blau has found that persons whose statuses are made inconsistent through social mobility face special dilemmas for interpersonal rela­ tions. 2 Finally, Everett Hughes has cogently summarized the relation of the status inconsistent (i.e., one whose statuses are uncrystallized) to his fellows in the following passage: lAo Zaleznik, C. R. Christensen and F. J. Roethlisberger, with the collaboration of George C. Homans, The Motivation, Productiv­ ity and Satisfaction of Workers: A Prediction-Study, Boston: Harvard unIversity Press, 195~ p. 63. - ----­ 2peter M. Blau, "Social Mobility and Inter-Personal Relations," American Sociological Review, 21 (June, 1956), PI'. 290-295. 91~ In order that men may communicate freely and confidently, they must be able to take a good deal of each other's sentiments for granted. They must feel easy about their silences as well as about their utterances. These factors conspire to make colleagues, with a large body of unspoken understandings, uncomfortable in the presence of what they consider odd kinds of fellows. The person who is the first of his kind to attain a certain status is often not drawn into the informal brother­ hood in which experiences are exchanged, competence built up, and the formal code elaborated and enforced. He thus remains forever a marginal man. l With respect to unstable self-images, Goffman, in discussing the sources of stress experienced by status inconsistents, points out that uncrystallized status may lead the individual to incorporate diverse and conflicting views of himself, thus making a unified self- image impossible. 2 Lipset and Bendix point out the same problem: • . . status discrepancies may cause difficulties in personal adjustment because high self-evaluations in one sphere of life conflict with low ones in another.3 Fenchel, Monderer and Hartley, in discussing the phenomenon of re- equilibration of out-of-line statuses, observe: :/ /j ll' ~~ In terms of personality stability, the phenomenon is suggestive l~ of a reference point for the individual which permits him to maintain a constant self-image. 4 lEverett C. Hughes, IIDilemmas and Contradictions of Status," American Journal of Sociology, 50 (March, 1945), p. 356. 2Irwing W. Goffman, "Status Consistency and Preference for Change in Pm"er Distribution," American Sociological Review, 22 (June, 1957), pp. 275-281. 3Seymour Martin Lipset and Reinhard Bendix, Social Mobility in Industrial Society, Berkeley and Los Angeles: University of California Press, 1960, p. 65. 4Gerd H. Fenchel, Jack H. Monderer and Eugene L. Hartley, "Subjective Status and the Equilibration Hypotheses," Journal of Abnormal and Social Psychology, 46 (October, 1951), p. 478. ~_- 95 Turning to the problem of rewards and investments, we find that the problem of imbalance between occupational investments and the rewards therefrom forms the basis of the "Theory of Distributive Justice ll developed by Zaleznik and his colleagues. l In their formula­ tion of this theory, the "investment" statuses that a worker brings to his job are age, seniority, sex, ethnicity and educat ion. The" revTard" status that he gets in return for these investments is that of "pay" or income. When a worker's "revTards" are out-of-line with his "invest­ ments" (either higher or lower), then a condition of "status incongruence" is said to exist. The authors hypothesize that such a condition vTill lead to "trouble'! or dissatisfaction. The concept of the balance between rewards and costs (investments) as a major factor in social behavior is developed to some length by Homans in other writings. 2 Finally, a number of authors have made reference to the prob­ lem of social ambiguity as a consequence of uncrystallized status. Hughes suggests that a complex of "auxiliary characteristics" adhere to any given status, and that behavior consistent with these character­ istics comes to be expected of the incumbent of that status.3 However, others find it difficult to relate to people who hold inconsistent lZaleznik, et a1., op. cit., pp. 50-56. 2See , for example, George C. Homans, "Social Behavior as Exchange," American Journal of Sociology, 63 (May, 1958), pp. 597-607, and Social Behavior: Its Elementary Forms, New York: Harcourt, Brace and World, Inc., 1961.--- ----­ ~ughes, op. cit. ~--_.---- ._-~----- ­ 96 statuses and hence contradictory status expectations. Persons holding inconsistent statuses based on sex, race, ethnicity and occupation (as, for example, the Negro doctor and the female engineer) are par­ ticularly prone to these ambiguities of status expectations. Irwin Goffman, in discussing the social behavior of persons with inconsistent statuses, observes that: . • • Individuals who are inconsistent in status are subjected to inconsistent pressures by those with whom they interact, i.e., others cannot appropriately "define" the inconsistent individual and anticipate his responses. This may lead to stress for the individual by subjecting him to inconsistent demands. l Zaleznik, et ~., point out that a condition of status incongruence is a condition of ambiguity which produces anxiety, not only for the in­ consistent individual but also for the group with which he interacts. 2 It would appear, then, that a major problem created by occupying out-of-line statuses is that of ambiguous, unclear and in­ consistent "normative expectations" for both the incumbent of these statuses and those with whom he must interact. These expectations, and the extent to which they are or are not fulfilled, determine not only the nature of the individual's social relationships and the "rewardll increment he gets from engaging in social behavior, but also his own self-image. Regardless of his relative position in the stratification system, when an individual's separate statuses are crystallized the problems of ambiguity and its concomitants do not arise. However, lGoffman, £E. ~., p. 276. 2Zaleznik, et al., £E. cit., pp. 63-64. 97 when an individual's statuses are sufficiently uncrystallized, he will suffer these problems and their behavioral consequences. The Psychological ~ Social Consequences of Uncrystallized status Much of the research in the area of status crystallization points to a general psychological reaction to uncrystallized status of stress, frustration, uncertainty and insecurity.l Jackson discusses the frustration that the inconsistent person experiences because of the contradictory expectations of others and his uncertainty with respect to what he can expect of others and what they can expect of him. He concludes: Inconsistency, then, produces conflicting expectations which result in frustration and uncertainty for the individual and thus increase his psychological stress. 2 As discussed earlier, Goffman points out that the inconsistent demands made of individuals whose statuses are out-of-line may lead to stress.3 LrhiS is not to say, however,.that these effects of uncrys­ tallized status are necessarily adverse for the individual, nor is any jUdgement being made as to whether any possible overt consequences of these psychological reactions are lTbetter" or "worselT for society than any other conditions. Another consideration which may be raised here is that the majority of the literature on status crystallization deals with motivated behavior as a consequence of subjective psychological states created by status inconsistencies. This position has consider­ able utility for the development of theoretical predictions. However, the possibility that behavior is a consequence of objective reality situations ~ ~ should not be overlooked. 2Jackson, lTStatus Consistency and Symptoms of Stress," op. cit., p. 470. ­ 3Goffman, op. cit. 98 Peter Blau has found that the lesser social integration of those vU10 are socially mobile manifests itself in a stronger feeling of insecur­ ity.l He observes that this insecurity can evidence itself in various ways, as, for example, in hostility toward minority groups. Finally, Zaleznik, et al., observe that the mixed expectations for behavior related to inconsistent statuses create ambiguities for the individual which result· in anxiety in social relationships.2 What are the possible social consequences which may be caused by the psychological stress arising from uncrystallized status? Among those suggested in the literature reviewed for the present investiga­ tion are social isolation, desire for change, motivation to action, upward mobility, and political extremism, both "left" and "right. tI These manifestations of status inconsistency '.Jill now be examined more fully. Lenski has found that persons with low status crystallization are more likely to be non-participants in voluntary relationships and are more likely to have given up long standing voluntary ties than per­ sons vlith high status crystallization.3 HO~'1ever, an additional observation should be carefully noted. That is, Lenski suggests that when persons with uncrystallized status are motivated to establish and maintain voluntary ties, they do so for non-sociable rather than IBlau, op. cit ., p. 295. 2Zaleznik, et a1., op. cit., p. 359. 3Lenski, op. cit. 99 sociable reasons. This suggests that such persons are not necessarily non-participants in social life, but rather that they may participate for some ulterior purpose. In Lenski's words: Thus, persons with poorly crystallized status might be expected to join voluntary associations not because of the opportunities which such groups provide for meeting new people and making new friends, but rather because the group and the social rela­ tionships which go with it are the necessary means to the end of personal advancement, the reform of society, or some other nonsociable end. l Zaleznik, et a1., also report social isolation to be related to low status crystallization: The data on group membership, apart from the sUb-system comparisons, showed that those workers who were high in status congruence tended to be members of groups while those low in status congruence tended to be isolates. 2 Turning to the possibility of desire for social change as a reaction to poorly crystallized status, Lenski, in his 1954 study, suggests that the use of the status crystallization model might lead to the discovery of basic relationships between social structure and li,social change. He hypothesizes: "~' '11'1 • . • one might predict that the more frequently acute status inconsistencies occur within a population the greater would 1111,1,1 be the proportion of that population willing to support programs of social change.3 Following up this hypothesis, Goffman found that preference for change lIbid., p. 461. 2Zaleznik, et a1., op. cit., p. 366. 3Gerhard Lenski, "Status Crystallization: A Non-Vertical Dimension of Social Status," American Sociological Revievl, 19 (August, 1954), p. 411. 100 in the distribution of power in society was a solution to status inconsistency at high strata levels. l stuart Adams' observations re­ garding the possible effects of status incongruency (low crystalliza­ tion) are also related to the phenomenon of change. He suggests that changes in congruency are accompanied by changes in individual motiva­ tion, in that consciousness of such incongruency may drive the individual to action of some kind. 2 One of the most widely-suggested consequences of uncrystal­ lized status is the phenomenon of social mobility. Goffman, in the study cited above, found that upward vertical mobility was a more common solution to status inconsistency at lower strata levels. This finding led him to make the following reformulated hypothesis: The degree of status consistency is inversely related to the extensiveness of preferences for change in the distribution of power when experienced opportunities for upward mobility are 101'1.3 This observation is supported by that of Jackson, who suggests that males vlith high educational and low occupation status can avoid a psychosomatic response to status inconsistency through upward social mobility.4 Fenchel, et al., found greater status strivings by indi­ viduals in those reference groups which represented their least lGoffman, Ope ~. 2Stuart Adams, "Status Congruency as a Variable in Small Group Performance, II Social Forces, 32 (October, 1953), pp. 16-22. 3Goffman, Ope cit. p. 279. 4Jackson, Ope cit. 101 crystallized status than in those which represented their higher crystallized statuses. l Finally, Homans, in a study of a group of "lOrking girls, found that one girl, the members of whose family had higher statuses than those of the other girls' families, eventually applied for a transfer to a division where the chances for advancement were greater. 2 In other 1vords, this girl's response to having an occupational status out-of-line with her other statuses was upward occupational mobility. Of central concern to this study is the suggestion made by a number of researchers that political extremism may be a common response to uncrystallized status. Goffman's findings regarding a desire for change in power distribution may be regarded as a form of political expression, although he does not specify whether this expression is that of the Tlleft Tl or of the "right."3 Lenski, in his 1954 study of the Detroit area, found that persons with low status crystallization were more likely to vote for the Democratic party and to express liber­ al social attitudes than were persons with high status crystallization. 4 This finding is supported by Jackson's more recent study, which found that persons with high achieved status or an ascribed status lower than IFenchel, et ~., op. cit. 2George C. Romans, liThe Cash Posters: A Study of a Group of Working Girls,1f American Sociological Review, 19 (December, 1954), pp. 724-733. 3Goffman, .£E. cit. 4Lenski, op. ~. 102 achieved status were politically liberal. In commenting on his and Lenski's findings, Jackson provides the following caution: These data serve as a reminder that a single social stimulus can produce more than one behavioral response. A wide range of alternative responses are available to status inconsistents. There is no reason to think that the source or the degree coerces the individual into a single pattern of reaction. l An observation in a similar vein is made by Lenski in a footnote to his 1956 study. In discussing the tendency toward a liberal political bias found in persons with uncrystallized status, he notes: In connection with this finding, Gordon has commented that this may be but part of a still more general tendency for persons with poorly crystallized status to adopt extreme political positions, whether on the extreme left or the extreme right. In the present data, there was no clear evidence of such a tendency, but no fully satisfactory test was possible because of the very small number of persons adopting an extremely conservative position on the three controversial issues. 2 The possibility of an extreme right political response to uncrystal­ lized status has not yet been investigated. It is to this problem that we shall next turn. status Crystallization and Right-Wing Extremism As we have seen, low status crystallization is related to psychosomatic symptoms of stress, social isolation, desire for change, upward social mobility, motivation to action, and political liberalism. IJackson, "Status Consistency, Vertical Mobility and Symptoms of Stress," op. cit., p. 120. 2Lenski, "Social Participation and Status Crystallization," op. ~., footnote 3, p. 459. 103 In connection with the latter phenomenon, it has been suggested that low status crystallization may also be related to political extremism on the right. In addition to the foregoing discussion of research findings directly related to the crystallization hypothesis, a number of other observations have been made linking Right-Wing Extremism to status discrepancies. Lipset and Bendix have observed that class discrepancies seem to predispose groups or individuals to accept extremist views. l Elsewhere, Lipset has noted that five national right-wing movements (McCarthyism, Poujadism, Italian Fascism, and German and Austrian Nazism) appealed mainly to the self-employed urban and rural middle classes (small businessmen and farmers).2 Lipset observes that individuals in these classes, whose status and influence within the larger community is declining, tend to feel cut off from the main trends of modern society. This observation is entirely consistent with those regarding the social consequences of uncrystallized status noted earlier in this chapter. Furthermore, the il ,,,use of the concept of "status politicslT in relation to Right-Wing Extremism was noted in Chapter III. lTStatus politics" refers essen­ tially to the projection of people's status anxieties and frustrations into .the political sphere. Considering that status anxieties and frustrations seem to be consequences of poorly crystallized status, lLipset and BendiX, "The Consequences of Social Mobility," in Social Mobility in Industrial Society, op. cit., pp. 64 ff. 2S. M. Lipset, "Social Stratification and 'Right-1Ving Extrem­ ism' ," British Journal of Sociology, 10 (December, 1959), pp. 1-38. I 104 the concept of f'status politicsl! would appear to be related to that of II status crystallization." Additional support for hypothesizing a relationship between uncrystallized status and Right-Wing Extremism is thus established, since a number of analysts have advanced the concept of II status politics" as a possible cause of extreme right political tendencies. l The foregoing evidence strongly suggests that an empirical relationship may exist between status crystallization and Right-Wing Extremism. Before proceeding to test this relationship, certain theoretical assumptions we have made regarding the nature of these two phenomena must be brought together. As Jackson has observed, a number of possible modes of response to low status crystallization exist. 2 The objective of this present study is to test the proposition that Right-Wing Extremism is one such alternative response. Earlier in this chapter, it was established that status inconsistencies result in frustration and uncertainty for the individual. Numerous studies of collective behavior and the processes in the development of social norms have established the fact that a basic reaction to such uncer­ lSee, for example, Daniel Bell, IIInterpretations of American Politics,1I in Bell (ed.), The New American Right, New York: Criterion Books, 1955, 1'1'. 3-32; S. W:-Lipset, "The Sources of the 'Radical Right',11 in Bell, 01'. cit., 1'1'.166-233; Richard Hofstadter, liThe Pseudo-ConservativeReVOlt," in Bell, 01'. cit., PI'. 33-35. 2Jackson, IIStatus Consistency and Symptoms of Stress,1I 01' • cit., p. 479 • 105 tainty is to attempt to restructure the situation, to find meaning. l We have seen that upward mobility, expression of a desire for change, activism, and political liberalism are some of the social expressions resulting from such attempts to restructure the situation. We have also seen that outward symptoms of psychological stress are evidenced by those who are blocked from such expressions. In Chapter III, the nature of the ideology of the Extreme Right as a positive, militant and millenarian solution to political frustration was discussed. There, it was pointed out that the nature of the American political system precluded an "institutionalized" expression of divergent, extremist politics, and that the Extreme Right provided an activistic political philosophy of "individualism" in a society characterized by "collec­ tivism." Finally, it has been suggested that those individuals who suffer status inconsistencies, although socially "detached;' may be "activistic," particularly where political issues are concerned. In light of this summary, it is the basic assumption of this study that the Extreme Right exists as a form of political expression offering simplistic, highly structured solutions to the frustrations arising from lack of status crystallization. 2 Persons suffering from lSee, for example, the studies reported in Ralph H. Turner and Lewis M. Killian, Collective Behavior, Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1957; Muzafer Sherif, The Psychology of Social Norms, New York: Harper and Brothers, 1936:-- --­ 2The concept of "simplistic," as it is used in this context, refers to the Extreme Right tendency to reduce complex factors and issues to a single element (e.g., the Right-Wing Extremist places the failure of the United States to "win" the "Cold War" to an internal 106 status inconsistencies will be more likely to maintain this political ideology than those whose statuses are crystallized. Therefore, in terms of the variables developed earlier in this stUdy, the basic hypothesis to be tested in the next chapter is as follow·s: Individuals characterized by low status crystallization are more likely to be Right-Wing Extremists in their political attitudes than individuals characterized by high status crystallization. In Chapter VI, conditions related to controls for the effects of status will be specified, and the hypothesis will be rephrased as a "null hypothesis ll for testing. "conspiracyll). This tendency has also been suggested by a number of the contributors to Daniel Bell (ed.), The Radical Right (The New American Right expanded and updated), Garden City, New Yor~ Doubleday and Company, Inc., 1963. For example, Talcott Parsons refers to Right­ Wing llindividualismll as the llidealization of pristine simplicity as against organizational and other compleXity" ("Social Strains in America: A Postscript,lI p. 195) and Alan F. Westin discusses the llfundamental­ istic ll characteristics of the Extreme Right, such as the belief that there are ali-rays solutions for social problems, and the advocacy of lldirect actionll (llThe John Birch Society: 'Radical Right' and 'Extreme Left I in the Political Context of Post World War II," p. 203). ___•••__ ,c.",.,....- ~ -, 107 CRAPIER VI STATUS CRYSTALLIZATION AND RIGHT-WING EXTREMISM: ANALYSIS OF DATA AND FINDINGS The Sample As indicated in Chapter IV, the original sample on which this secondary analysis is based was one taken in the Eugene-Springfield metropolitan area in the Summer of 1959 by the Institute for Community Studies, University of Oregon. The original study was designed to investigate local political attitUdes, opinions and behavior. Sampling was done in five geographically distinguishable residential areas: Eugene, Springfield, the River Road and Ferry Street Bridge areas, and East Springfield. Each of these areas was sampled by means of probabil­ ity sampling techniques. l However, a proportional sample was obtained only in the Eugene area. Residential strata in this area were deter­ mined by the use of census enumeration districts as the sampling frame. A systematic sample of households was made in the remaining areas, using utility company records to define the sampling unit. Different sampling intervals were employed in each of these areas. The initial sample obtained in the Eugene-Springfield metro­ politan area contained 1231 respondents, distributed among the lFor a full discussion of these techniques, see Appendix A, pp. 153-156. 108 residential areas sampled as indicated in Table 9. The lack of propor­ tional distribution of respondents by residential area may be observed by comparing the sample sizes for Eugene and Springfield to the total populations in these cities, which, in 1960, ,.Tere 95,686 and 19,616 respectively. 1 TABLE 9. SAMPLE SIZE BY AREA, EUGENE- SPRINGFIELD METROPOLITAN AREA STUDY Area Sample Size Eugene city 526 Springfield city 448 Glenwood (a Springfield suburb) 12 River Road 73 East Springfield 80 Willakenzie (Ferry Street Bridge) ~ Total: 1231 Considerable attrition of this initially large sample occurred due to non-response to the mail-back questionnaire administered as part of the survey. Since this mail-back contained a number of questions relative to Right-Wing Extremist attitudes, respondents who failed to return it had to be eliminated from the sample on which this present study is based. This non-response part of the sample contained 490 cases, or forty percent of the initial sample. Unfortunately, nothing can be done about the possible effect this attrition may have on the Lunited States Bureau of the Census, United States Census of Population: 1960. Detailed Characteristics. Oregon. Final Report-­ PC(l) - 39B. -united States Government Printing Office, Washington, D.C., 1962. 109 "representativeness" of the sample. However, questions relevant to the variables of occupation, income and education were on the basic interview schedule. Therefore, the distribution of non-response with .respect to these independent stratification variables is known, and can be accounted for in the test of the hypothesis advanced in this study. As might be expected, non-respondents were overrepresented by low occupational, income and educational status, as indicated in Tables 10, 11 and 12 below. Before presenting these tables, a note should be made of the cutting points used to discriminate between "high" and "low" status on each of these variables. These cutting points correspond roughly to the medians for the distributions of occupation, income and education given in Tables 4, 5 and 6 of Chapter II (cf. pages 31-33 supra). Thus, occupation is dichotomized into "\'lhite Collar" (Clerical and Sales and above) and "Blue Collar" (Craftsmen and below); income is dichotomized into "High" ($6000 and over) and "Low" (less than $6000); education is dichotomized into "High" (beyond High School) and "LO"T" (High School or less). In addition to usage in the following tables, these cutting points are to be taken for all further reference to these three variables. The possibility that the disproportionate representation of of status variables indicated by these tables might invalidate the findings of the present study should be taken into account. If Right­ Wing Extremism is associated with either low income, low education or 10,"T occupational status, then the lack of an adequate representation /il . ~}1 ii' ~j - - 110 TABLE 10. OCCUPATIONAL STATUS BY MAIL-BACK RESPONSE AND NON-FE SPONSE , AND TOTAL SAMPLE Response Non-Response Total Sample 'VJhi te Collar 41% 31% 37% Blue Collar 53 62 56 Not Classifiable 6 - 7 7 100% (741) 100% (490) 10CfJ!o (1231) TABLE 11. INCOME STATUS BY MAIL-BACK RESPONSE AND NON-RESPONSE, AND TOTAL SAMPLE Response Non-Response Total Sample High Income 37% 26% 33% Low Income 60 66 62 No Answer 8 _5 -l - 100% (741) 10CfJ!o (490) 10CfJ!o (1231) TABLE 12. EDUCATIONAL STATUS BY MAIL-BACK RESPONSE AND NON-RESPONSE, A1iJD TOTAL SAMPLE Resp~ Non-Response Total Sample High Education 39% 22% 32% Low Education 60 77 67 No Ans'tver 1 1 1 - 100% (741) 100% (490) 10CfJ!o (1231) 111 of respondents having such status attributes could operate against any hypothesis testing the incidence of Right-Wing Extremism in the sample. On the other hand, if extremism were related to high position on these variables, then an overrepresentation of respondents with such status attributes could introduce bias in favor of this hypothesis. Tables 13, 14 and 15 present the relationship between Right-Wing Extremism and each of these status variables for the final sample used in this study. TABLE 13. PERCENTAGE OF RIGHT-HING EXTREMISTS BY OCCUPATION White Collar Blue Collar %Right-Wing Extremist 23 (297) 20 (376) TABLE 14. PERCENTAGE OF RIGHT-WING EXTREMISTS BY INCOME High Income Low Income %Right-Hing Extremist 20 (266) 22 (407) TABLE 15. PERCENTAGE OF RIGHT-WING EXTREMISTS BY EDUCATION High Education ~ Education %Right-Wing Extremist 17 (266) 24 (407) An examination of Tables 13 and 14 ShovlS that Right-Wing Extremism is not significantly related to either occupation or income. On the other 112 hand, the data in Table 15 show that low educational status does seem to be related to Right-Wing Extremism. However, since the sample being used in this study is underrepresented by low education respondents (cf. Table 12), this relations~ip will tend to introduce bias against the hypothesis being tested. The same holds true for the variable of income, although, as indicated, it does not seem to be significantly related to Right-Wing Extremism (cf. Table 14). This effect does not operate for occupation, however. Since Right-Wing Extremism seems to be slightly more related to high occupation than to low occupation, the overrepresentation of white collar respondents in the sample will tend to introduce bias in favor of the hypothesis being tested. However, it may be considered that this bias will not be severe, since the relation­ ship between white collar occupation and Right-Wing Extremism is negligible (cf. Table 13). A further safeguard against the possible biasing effects of status will be introduced by "holding constant" the variables of occupation, income and education in the elaboration of the relationship between status crystallization and Right-Wing Extremism. This subject will be discussed in the follo1>Ting t,oJ'O sections. Status and Crystallization Since status crystallization, as it has been defined and operationalized for this study, is a function of given status variables (in this case, occupation, income and education) some method of holding status constant must be used when studying the relationship between 113 -_ .... _--_._-- -------_._---­ crystallization and political attitudes. This is necessary since it would not otherwise be possible to determine if any relationship was due to.crystallization dif~erences or to status differences. Realizing this problem, Lenski attempted to control for status by eliminating a number of respondents with the highest income scores from his high crystallization group, and a number of respondents with the lowest income scores from his law crystallization group. This procedure was based on the observation that high crystallization respondents had higher mean income and occupation scores than did law crystallization respondents. Therefore, Lenski contends that by elimi­ nating income extremes at both ends of the crystallization score distribution, mean status scores are equalized, thus controlling for the effects of status in the relationship between crystallization and political attitudes by 1I1oadingll the comparison against his hypothesis. Lenski feels that this procedure controls not only for income, but also for all status variables simultaneously.l Thus, as Selvin has pointed out, Lenski used "frequency matching" rather than "precision matching" (or'cross tabulation).2 However, this use of ~~ (mean status scores) to make inferences about the distribution of individuals (respondents with the highest and lowest income scores) is an example laerhard Lenski, "Status Crystallization: A Non-Vertical Dimension of Social Status," American Sociological Review, 19 (August, 1954), p. 409~ 2Hanan C. Selvin, "Problems in the Use of Individual and Group Data," (mimeographed), University of California, Berkeley, (January, 1960), PP. 5-7. 114 of an "ecological fallacy."l That is, under no circumstance can the knowledge of group distributions (ecological correlations) be used to make inferences about the distribution of characteristics with respect to individuals. Basically, status crystallization is a measure of the degree of variation between rank positions on given vertical status hier­ archies. More specifically, it is a measure of the degree of deviation of vertical status hierarchy scores from a line with a slope of 45 degrees passing through equivalent scores for ~ status variables. The region in an ~-dimensional space traversed by this line (i.e., the region where the line passes through equivalent scores on all statuses) is the region of status crystallization. This is the basis of the formula for scoring status crystallization, presented in Chapter II, which expresses crystallization as a function of deviations from the mean of given status scores--the lesser the deviation, the higher the crystallization; the greater the deviation, the lower the crystalliza­ tion. Figure 2, although necessarily restricted to two dimensions, will illustrate this meaning of status crystallization. The shaded portion of this figure, the area where deviations are least, represents the area of highest status crystallization. Beyond this area, where deviations are greatest, is the region of low status crystallization. lFor discussions of this concept, see Selvin, ibid., and W. S. Robinson, "Ecological Correlations and the Behavior-of' Individuals," American Sociological Review, 15 (June, 1950), PP. 351­ 357. ---------.. _------------­ 115 FIGURE 2. SCHEMATIC REPRESENTATION OF STATUS CRYSTALLIZATION AS A FUNCTION OF DEVIATIONS FROM AN EQ,UIVALENT STATUS SCORE LINE • High /0 8 6Status I . Scores­ - 4 2 Low ". Z I 4- i B 10 I i Low " High Status II Scores The implication of this conceptualization of status crystal­ lization for the relationship bet'ween status and crystallization is simple. Given that an individual's statuses are crystallized, and given his score on one status dimension, one can predict (Within reasonable limits) that individual's status scores on the remaining status dimensions. That is, for Figure 2, if a "high crystallization" respondent has a low score on Status I, his score on Status II must also be low. For an individual whose statuses are uncrystallized, however, no such prediction is possible, particularly in a multi­ dimensional space. ~ 116 In light of these observations, it may be seen that Lenski's method of controlling for status could hold only for the high crystal­ lization group. That is, eliminating those respondents with the high­ est income scores in the high crystallization group will effectively bend the slope of the equivalent status score line at the top for in­ come as well as for other correlated variables. However, eliminating the lowest income respondents in the low crystallization category will have virtually no effect on the slope of the line since, by definition, to be uncrystallized is to have unequal scores on the status variables being measured. Thus, a respondent with uncrystallized status and a low score on income may have moderate or high scores on the other status variables. In no way can the effect of removing him from the sample under study be predicted. This analysis of the relationship between status and crystal­ lization is important, since it points out that no determination of the distribution of individual status scores can be made from knowledge of group scores, particularly for the low crystallization category. Thus, status controls made by the manipulation of status scores within crystallization groups, as Lenski has done, are not effective. However, it is imperative that controls for the effects of status be made in testing the relationship between crystallization and political atti­ tudes. Therefore, some method of controlling other than Lenski's must be used. ..""1 117 The ~ of Elaboration in Survey. Research In the classical controlled experimental design, cause and effect relationships are established by comparing the responses of an experimental group, which has been exposed to the stimulus being tested, with the responses of an equivalent control group, which has not re­ ceived the stimulus. The experimenter employs matching and/or random­ ization techniques to make the two groups initially identical on those ~I !' factors which may create a "spurious ll causal relationship betvleen the independent and dependent variables. In the social sciences, however, opportunities are seldom afforded to control in advance the composition of groups being studied. Thus, there is always a danger that observed relationships may be caused by initial differences in the groups being compared, and are therefore "spurious." The factors making for initial differences between groups being studied may be called "invalidating factors,lI in that they may invalidate the relationships found between independent and dependent variables. In the present study, the analytical procedure known as explanation will be used for controlling on possible invalidating factors. l In this procedure, an initial cross-tabulation between two lFor detailed discussions of this procedure, see Paul L. Lazarsfeld, IIInterpretation of Statistical Relations as a Research Operation,ll in Paul F. Lazarsfeld and Morris Rosenberg, The Language of Social Research, Glencoe, Illinois: The Free Press, 195~pp. 115-125; Patricia L. Kendall and Paul F. Lazarsfeld, "Problems of Survey Analysis,lI in Robert K. Merton and Paul F. Lazarsfeld (eds.), Continu­ ities in Social Research, (Studies in the Scope and Method of liThe AmerIcan Soldier"), Glencoe, Illinois: The Free Press, 1950, pp. 133­ 118 variables is stratified by different values of one or more "test" • variables (or possible invalidating factors). Comparisons between the "cause II and "effect" variables being measured are then made in the sub­ groups obtained. If the initial relationship between these variables is maintained in each of the subgroups, then it may be concluded that this relationship is a "true" causal one (i.e., to the extent that it is not caused by differences in the "test" variables). If, on the other hand, the original relationship is not maintained, then it is said to be a "spurious" one, caused by the effect of the "test" variable on both the "cause tl and "effect" variables. This method of making comparisons between subgroups (obtained by stratifying an initial cross-tabulation by different values of a test variable) approximates the controlled experiment. What is effec­ tively obtained in the subgroups is an "experimentalll group having one value of a test variable and a "control" group not having this value. The researcher should, however, be aware of several limitations to this method of controlling. First, since randomization techniques (i.e., methods of selecting the control and experimental groups so that initial differences between them may be considered to be randomly distributed) are not employed, the effect on the cross-tabulations of those variables which have not been controlled cannot be assessed. Second, in order to control for interaction effects (i.e., the joint 196; Herbert Hyman, Survey Design and Analysis, Glencoe, Illinois: The Free Press, 1955, Chapter VI, IiThefutroduction of Additional Variables and the Problem of Spuriousness," pp. 242-274. 119 effect of two or more test variables operating together) in the cross­ tabulation, it is necessary to control on all test variables simul­ taneously. For example, certain relation~hips between A and B may be observed for test values Xl and X". However, in order to ascertain whether the different values for A and B for XI and A and B for X" are "true" differences or merely the result of a differential effect of a third variable, Y, on XI and X", the subgroups for XI and X" must be further stratified by y l and Y". Since the addition of each test vari­ able geometrically increases the number of subgroup comparisons to be made (i.e., controlling on one variable requires two subgroup cells, controlling on two variables requires four cells, controlling on three requires eight cells, etc.), simultaneous controlling on a number of test variables requires an extremely large sample size. Third, "explanation" requires that the test variable be antecedent to both the "cause" and "effect" variables. If the test variable is an inter­ vening variable, then its effects are an essential link in "interpreting" the relationship between the "cause" and "effect" variables. In the present study, we are concerned with the relationship between status crystallization (the "cause" variable) and Right-Wing Extremism (the "effect" variable) at a static point in time, rather than as a sequential development. Since status crystallization has been defined for the purposes of the present study, as a function of status vari­ ables (cf. Chapter II), it may be assumed that the status variables which will be controlled for in the relationship between crystalliza­ 120 tion and Right-lUng Extremism are antecedent to the variable of status crystallization. In surrunary, the elaboration of a relationship by stratifying on antecedent test variables provides a procedure, based on the logic • of the controlled experiment, which permits the social scientist to "control" for the invalidating effects of any given factors. Effec­ tively, this allows the researcher to make causal inferences, at least within the limits of the factors for which he has controlled. Lazarsfeld has expressed this point as follo..,s: We can suggest a clearcut definition of the causal relationship between two attributes. If we have a relationship between fiX" and "y"; and if for any antecedent test factor the partial relationships between x and y do not disappear, then the original relationship should be called a causal one. l In the present study, controlling for the effects of status in the relationship between status crystallization and Right-Wing Extremism ,,,ill be done by attempting to "explain" any relationship found between these variables in terms of the status factors of occupa­ tion, income and education. If the hypothesized relationship between crystallization and extremism still remains in the partials after stratifying simultaneously on these status variables, then it will be assumed that this relationship is a "true" causal one (i.e., that it is not attributable to either occupational, income or educational difference s) • lLazarsfeld, ~. cit., pp. 124-125 (italics in the original). 121 Testing ~ Hypothesis At the end of Chapter V, it was indicated that the basic hypothesis developed regarding the relationship between status crystal­ lization and Right-Wing Extremism would need to be specified in terms of the variables which are to be controlled, and rephrased as a null hypothesis for testing. The latter procedure is related to the test of statistical significance, and will be discussed later in this chapter. The original hypothesis developed in Chapter V asserted that individuals with 1mV' status crystallization would be more likely to be Right-\'ling Extremist in their political attitudes than individuals with high status crystallization. When Right-Wing Extremists (cf. data from Chapter IV, pages 89-91 supra) are stratified by low and high crystal­ lization (cf. data from Chapter II, pages 34-37 supra), proportions of extremists for e.ach crystallization category obtain as shown in Table 16 below. The original hypothesis that individuals with low status crystallization are more likely to be Right-Wing Extremist than indi­ viduals with high status crystallization is borne out in the data of this table, although the percentage difference between the two crystal­ lization groups (4%) is relatively small. TABLE 16. PERCENTAGE OF RIGHT-WING EXTREMISTS BY STATUS CRYSTALLIZATION Per Cent Right-Wing Extremist Low crystallization 24 (176) High crystallization 20 (497) 122 Earlier in this chapter, the necessity of controlling for the possible invalidating effects of the status variables used in measuring crystallization 1,as discussed. From this observation, the basic hypothesis (Hl) to be tested in this study may nm'7 be expressed as folloTtlS: Individuals characterized by 101'1 status crystallization are more likely to be Right-Hing Extremist in their political attitudes than are individuals characterized by high status crystallization, when the invalidating effects of status differences in occupation, income and education are controlled. The data for the test of this hypothesis are presented in Table 17. In this table, Right-I,ring Extremists for low and high crystallization categories are stratified by occupation, income and education. Thus, column (a) contains data for respondents with white collar occupations, high income and high education, column (b) contains data for respondents Tidth vlhite collar occupations, high income and 101-[ education, and so forth. l The widely disparate cell frequencies between low and high crystallization groups at the extremes of Table 17, i.e., columns (a) and (h), may seem incongruous when compared to the more or· less equal distributions in the remaining six columns. However, when we consider the meaning of status crystallization, it becomes ap~arent that the particular combination of status variables characterizing columns (a) and (h) are the only hw of the eight combinations that Hill yield llpure ll crystallized status. Unfortunately, this fact also determines that fev7 cases will be recorded in the low crystallization group for 1 ~-For a discussion of the dichotoTi1izations of these three variables, see p. 109 supra. TABLE 17. PERCENTAGE OF RIGHT-WING EXTREMISTS BY LOIrf AND HIGH CRYSTALLIZATION GROUPS, CONTROLIED FOR STATUS DIFFEPENCES IN OCCUPATION, INCOME AND EDUCATION Occupation: Hhite Collar Blue Collar Income: High Low High Low Education: High LOI... High Low High Low High Low Per Cent Right-Wing Extremist by: (a) (b) (c) (d) (e) (f) (g) (h) Low Crystallization -­ (---) 40 (15) 15 (46) 38 (32) 29 (14) 21 (38) 12 (26) 40 ( 5) High Crystallization 20 (111) 24 (33) 21 (34) 19 (26) o (10) 9 (1~5) 8 (25) 25 (213) Percentage Difference -­ 16 -6 19 29 12 4 15 E = .06* * The statistical test of significance used here is the "Sign Test," as discussed by Sidney Siegel, Nonparametric Statistics for the Behavioral Sciences, New York: McGraw-Hill Book Company, Inc., 195~p. 68-75_ I-' I\) w 124 columns (a) and (h). The complete loss of cases for the low crystal­ lization category of column (a) caused by this feature reduces to seven the number of comparisons that can be made in this table. With respect to the effect of individual status variables on the distribution of Right-Wing Extremists in the status crystalliza­ tion categories, little in the way of consistent trends can be discerned. The data of Table 15 (page III supra) suggested that low educational status might be related to Right-Wing Extremism. Comparison of the percentage differences between columns (c)/(d) and (g)/(h) indicates that this relationship may remain w'hen other status variables are controlled. The evidence is not complete, however, since this relation­ ship is reversed between columns (e)/(f) and no comparison is possible for columns (a)/(b). Nevertheless, one patterned relationship for combinations of status variables may be discerned in Table 17. The two lowest percentage differences (-6% and 4% for columns (c) and (g) respectively) are observed for the "low income, high education" combi­ nation. A possible explanation for this finding is that people in this circumstance, regardless of their occupational level, feel the need for some redistribution of economic opportunities, and have a sufficient level of educational sophistication to believe that contemporary American liberalism can bring this about. Therefore, low crystalliza­ tion as a consequence of low income and high education does not seem to result in a highly disproportionate incidence of Right-Wing Extremism. This finding is significant when compared to the findings of two earlier 125 status crystallization studies conducted by Lenski and Jackson. l Both these researchers found that high occupation or education combined with low racial-ethnic status predisposed respondents with status inconsist­ encies to a 111eftl1 or liberal political response. It may be that, under given conditions, a combination of high income and low education may predispose Right-Wing attitudes, whereas a combination of low income and high education may predispose Left-Wing attitudes. The implications of this for further studies of political extremism are discussed in the concluding chapter of this study. The final problem to be considered herein is whether the number of predicted differences observed in Table 17 is statistically significant. The null hypothesis (HO)' based on Hl' to be tested is as follows: No difference between low and high crystallization groups will be observed with respect to the dependent variable of Right-Wing Extremism when the invalidating effects of status differences in occupation, income and education are controlled. This form of the hypothesis implies that the crystallized and uncrystal­ lized groups are, in fact, similar with respect to extremist attitudes, and that any differences found between them are due to chance variations. If, at a given level of significance, it can be shown that observed differences between these two groups would be unlikely if the groups were, in fact, the same then the null hypothesis can be rejected. luerhard Lenski, op. cit., pp. 405-413; Elton Jackson, "Status Consistency and Symptomsof Stress,l1 American Sociological Review, 27 (August, 1962), pp. 469-480. 126 Although rejecting the null hypothesis does not prove the orignal hypothesis to be true, it does mean that it has not been disproved. In establishing the acceptable level of significance for the statis­ tical test used in this analysis, the convention of R = .05 will be follow'ed. In testing the significance of the present findings, the con­ cern will be whether the number of positive values obtained (i.e., the number of percentage differences in the direction predicted under HI) could have occurred by chance. Thus, a one-tailed test will be used, since this concern involves prediction in only one direction. An appropriate test for use when the data involves merely the number of plus or minus values observed is the Sign Test, which is based on the binomial distribution. l In the present data, plus (predicted) values are associated with a preponderance of Right-Wing Extremists in the low crystallization group; minus values are associated with a preponderance of Right-Wing Extremists in the high crystallization group. On the null hypothesis (RO), we would expect positive and negative occurrences with equal probability. Since HI posits that positive differences will occur more frequently, a one-tailed test will be used. For the data in Table 17, seven pairs of comparisons are being made (N = 7), one of which is negative (x • 1). Reference to Siegel shows that the chance probability of the occurrence under HO of x = 1 IFor a discussion of this test, see Sidney Siegel, Nonparametric Statistics for the Behavioral Sciences, New York: McGraw-Hill Book Company, Inc::-1956, pp. 68-75. l 127 or fewer when N = 7 is E = .062. 1 This level of significance does not quite meet the specified level of E = .05. However, the fact that known sample biases on the variables of education and income operate in favor of HO (see ~, pages 109-112), the rejection of HO at the .06 level of significance would appear ''iarranted. The conclusion which may be dravm from the foregoing analysis of the relationship between status crystallization and Right-Wing Extremism is, therefore, that individuals whose statuses are uncrystal­ lized are more likely to be Right-Wing Extremist in their political attitudes than individuals ''ihose statuses are crystallized. The implications and significance of this finding will constitute, in part, the next and final chapter of this study. lIbid., pp. 68-69 and Table D, p. 250. 128 CHAPl'ER VII Sm~RY, DISCUSSION A}ID CONCLUSIONS Summary The phenomenon of the "Extreme Right ll has become increasingly manifest on the American political scene during the past few years. The present study has been an attempt to predict, using sociological variables, the incidence of Extreme Right political attitudes among individuals. This investigation constitutes a secondary analysis of survey data collected by the Institute for Community Studies, Univer­ sity of Oregon, in their "Eugene-Springfield Metropolitan Area Study" of 1959--a study of local political attitudes and behavior. Relevant literature was reviewed, in an attempt to arrive at a suitable working definition of the Extreme Right. On the basis of an analysis of the ideological characteristics of the Extreme Right, it was defined as follows: The Extreme Right is a militant and millenarian political ideology, espoused by numerous Right-Wing groups and individuals, which maintains as an ideal the principle of "limited individualism"; this principle being articulated as opposition to lIcollectivism" in government, international relations, modern social principles, and modern social structure and operation. A cluster analysis of the data at hand was made, and twenty- six items related to Extreme Right attitudes were found to cluster into ---- .~-----_.._------­ 129 four groups, as follows: I. Attitudes regarding urban renewal. II. Attitudes regarding modern social structure and operation. III. Attitudes regarding government. IV. Attitudes regarding modern social principles. On the basis of a sum of Likert-type responses to these items, the sample under study was divided into two categories: "Right-Wing Extremist" and "Non-Extremist." In this research, Right-Wing Extremism constitutes the dependent variable. The independent variable of this study is the sociological factor of "status crystallization," defined as the extent to which an individual's rank positions on given status hierarchies are at a comparable level. This conceptualization provides a horizontal model of social stratification, in contrast to the traditional uni-dimensional vertical model. The status characteristics used in this study were occupation, income and education. Intervals for each of these three statuses were ranged in rank order and cumulative percentile ranges were computed for each interval. Scores for each interval were assigned on the basis of the midpoints of these percentile ranges. A status crystallization score was then computed for each respondent in the sample, using the following formula: Status Crystallization Score = 100 - )L c:(~, where d represents deviations from the mean of the three vertical status scores, and 100 a score inversion factor. Using a frequency 130 distribution of these scores, the sample was divided into two catego­ ries of nlown and t1highn crystallization, the former group constituting the least crystallized quartile of the total sample. The literature on status crystallization indicates that status inconsistents suffer from unsatisfactory social relationships, unstable self-images, rewards out-of-line with investments, and social ambiguity. These conditions conspire to create ambiguous, unclear and inconsistent t1normative expectationsn for both those who hold inconsist­ ent statuses and those with whom they interact. It has been further suggested that the psychological reaction to this situation is one of stress, frustration, uncertainty and insecurity; which in turn is overtly manifested in social isolation, desire for change, motivation to action, upward mobility, and political extremism (both "left" and "right") • The basic assumption of the present research was that the Extreme Right exists as a form of political expression offering sim­ plistic, highly structured solutions to the frustrations arising from lack of status crystallization, and it was hypothesized that individuals characterized by low status crystallization are mOre likely to be Right­ Wing Extremists in their political attitudes than individuals character­ ized by high status crystallization. In testing this hypothesis, the possible invalidating effects of status per ~ were controlled by stratifying the relationship be­ tween status crystallization and Right-Wing Extremism on the variables of occupation, income and education. This is the analytic procedure 131 of elaboration known as ltexplanation" or ttcontrol for spuriousness. 1I The failure of cases to be recorded in one colulllI). of the table in which this hypothesis was tested made comparison between high and low crystal­ lization groups on the variable of Right-Wing Extremism impossible for one set of status characteristics. In six of the seven remaining pairs of comparisons, differences in the predicted direction as large or larger than the original relationship between the two crystallization groups were observed. The null hypothesis positing no difference in Right-Wing Extremism between these two crystallization groups was rejected, with a statistical significance of E = .06. On the basis of these findings, it was concluded that Right-Wing Extremist attitudes are significantly related to uncrystallized status. Discussion In general, it may be said that secondary analysis is weak for predictive purposes, but strong from an exploratory point of view. In the secondary analysis undertaken in this study, we have been restricted to using available data, which was not originally designed for the present research purpose. The effects of this have been felt with respect to both the independent and the dependent variables. Re­ garding the former, the information available on occupation, income and education was not as detailed as could have been desired. For example, it would have been preferable if actual years of schooling had been given for education beyond high school, rather than type of schooling. Also, since we dealt with the family as a status unit, 1 I II 132 full status data on the ~ of household would have been desirable, thus avoiding the need for certain extrapolations. Moreover, data on occupational history would have been helpful, especially with regard to retired persons. Too often in survey research, this type of factual or "census" information is skimpy, since the length of time required to extract full information reduces the time available to gather other information vlhich may be considered "more important." However, in the type of study where status information is vital, such sacrifices cannot be made. In future research on status crystallization, considerable attention must be paid to collecting sufficiently detailed status information. It is always possible to collapse overly-detailed informa­ tion, but one cannot do the reverse. With respect to the dependent variable under study, the attitudinal information available may be considered to be, at best, only an approximation of vlhat would be desirable. This information taps only a fraction of the full complex of twenty-five attitudinal components discerned as characterizing the Extreme Right. Several of the attitudinal components which were represented in the data were "over-measured" (e.g., urban renewal), in that a number of relevant questions were asked where one could suffice. In future research on the Extreme Right, more precise and specific indicators, tapping all attitudinal components, will be required. HOvlever many shortcomings secondary analysi s may have with respect to the present study, it does provide invaluable guidelines for future research. Some of the insights gained in the course of the 133 present research are as follows: 1. The purpose of future research has been clarified. The pres­ ent study lends considerable support to the conviction, expressed in earlier literature, that differential status crystallization has a definite bearing on behavior and attitudes, particularly those of a political nature. The way is now paved for further studies that will investigate this relationship more extensively. 2. The nature of the concepts with which we have been working has been clarified. An important phase of the present study has been that dealing with a working definition of the Extreme Right--a problem which may otherwise have been neglected under the pressures of design­ ing a research study and collecting data. 3. A number of problems with respect to the kinds of data needed for a study of this type have been brought to light, and these diffi­ culties can be avoided in future research. 4. Related to the above point is the fact that an understanding of the exact information needed in future research has been reached. Thus, a more specific, and consequently shorter, format may be used for gathering data in the future, rather than relying on a "shotgun" approach. 5. A greater awareness of what variables may be important is also an exploratory gain achieved by secondary analysis. Thus, the thorough investigation of the concept of status crystallization undertaken for the present study points out the fact that status variables other than those used in the present study may be important. This point will be 134 referred to later in this chapter. 6. Finally, this type of investigation points to profitable areas for future research. In line with this, the possibility of studying both "1eft II and "right" extremism simultaneously \-Till be discussed later in this chapter. One important point which must be considered in any future use of the status crystallization model is the question of which sta­ tuses are most salient in producing the overt consequences of inconsist­ ency. Most of the studies published to date have used "objective ll statuses (either "achieved and/or lIascribed"), primarily because they are easy to measure. Lenski and Jackson both used occupation, income, education and ethnicity;l Kenkel used occupation, education, rental value of dwelling, and d,-relling area prestige; 2 occupation, income and education were used in the present study. An apparent assumption relied upon in many studies using the status crystallization model is that status inconsistency has adverse effects on those whose statuses are out-of-line. However, as pointed out in Chapter V, this may not be so. For example, one possible reaction to status inconsistency which has not yet been researched is "creativity." Whether the creative lElton F. Jackson, "Status Consistency and Symptoms of Stress," American Sociological Review, 27 (August, 1962), pp. 469-480; Gerhard E. Lenski, IfStatus Crystallization: A Non-Vertical Dimension of Social Status," American Sociological Review, 19 (August, 1954), pp. 405-413. 2vJilliam F. Kenkel, "The Relationship Between Status Consist­ ency and Politico-Economic Attitudes," American Sociological Revie"T, 21 (June, 1956), pp. 365-368. 135 impulse is adverse for individuals or society could be debated. On the other hand, lack of status crystallization, as measured objectively, may be "normal" for many persons. For instance, teachers may "expect" relatively 10\01 pay in return for their high education, and other "non­ economic" rewards may compensate for their lack of high income. As an alternative to "objective" status it is possible that the use of "subjective" or "self-rated" status vlould increase the power of the crystallization model, since this vTould assess whetqer a respondent perceived problems of status inconsistency. Little has been done in regard to subjective status as yet, although Fenchel, Monderer and Hartley had their sample of college students rate their present and their desired statuses in five different reference groups (the "House" group to which they belonged, other school or outside groups, the general student body at the college, friends and acquaintances, and family).l Thus, although "objective" status is important, the possible contribution to be made by using "subjective" status should not be overlooked in future research. Conclusions One possible contribution which the present research may have achieved is that of controlling for the invalidating effects of the status variables used in the crystallization model. The importance of laerd H. Fenchel, Jack H. Monderer and Eugene L. Hartley, "Subjective Status and the Equilibration Hypothesis," Journal of Abnormal and Social Psychology, 46 (October, 1951), pp. 476-4791. 136 this factor has been pointed out by Elton Jackson: The biggest problem in this kind of research, of course, is how to observe the effects of status inconsistency while controlling for the effects of status ~~; I don't think the final answer to this question has been uncovered yet. l As well as can be determined, the present research is the first to employ controls for all status variables simultaneously. However, the method by 1"hich this has been achieved requires exceptionally large sample sizes, and it is not likely that controlling can be done for more than three or four status variables at one time. In fact, the author is willing to concede that the low frequencies recorded in some of the cells of Table 17, Chapter VI, produced by this method of con­ trolling, may cast doubts on the significance of the findings of this study. Nevertheless, the present findings seem to point to a definite relationship between inconsistent status and Right-Wing political extremism. One additional point which may be raised in this section is that of the relationship between the present findings and those of Lenski, who, in his 1954 study, found low status crystallization to be related to political liberalism. 2 On the surface, the apparent relation­ ship of uncrystallized status to both Right-Wing Extremism and political liberalism may seem inconsistent. However, it is possible that this inconsistency can be explained in terms of certain evidence that strong lpersonal communication, JUly 12, 1962. ,. ·t 2LenSKl, op. Cl • 137 similarities exist between Left and Right Extremism. Westin has dis­ cussed similarities in ideologies, programs, strategies and tactics between the Right of today and the Left of 1945-1948;1 Ringer and Sills have discerned a number of similarities in social characteristics be­ tween Right and Left Extremists in Iran. 2 !\lthough the Extreme Left is virtually nonexistent in contemporary American society, it is possible that Lenski's indicators of liberalism did not discriminate between the moderate "liberaltl Left and vThat vestiges of the Extreme Left may have been in the population. If this is the case, and since there is reason to believe that Lenski's Detroit Area data may have contained a large proportion of working-class respondents, much of what Lenski measured as "liberalism" may actually be Lipset's "working-class authoritarianism. "3 Implications for Future Research At several points in the present study, observations have been made regarding the relationship of status crystallization to both "left" and "right" political extremism. It would appear that a significant lAlan F. Westin, "The John Birch Society: 'Radical Right' and 'Extreme Left' in the Political Context of Post World War II, If in Daniel Bell (ed.), The Radical Right (The New American Right expanded and updated), Garden City, New York: Doubleday and Company, Inc., 1963, pp. 201-226. 2Benjamin B. Ringer and David L. Sills, "Political Extremists in Iran: A Secondary Analysis of Communications Data," Public Opinion Quarterly, 16 (Winter, 1952-1953), pp. 689-701. 3s M. Lipset, "Democracy and ~'lorking-Class Authoritarianism," American Sociological Review, 21~ (August, 1959), pp. 482-501. 0 138 ---_ .•. --------­ contribution to our knmvledge of these forms of political behavior could be advanced by studying both of them simultaneously, using the status crystallization model. A major question to which such a study could be directed would be to discover what elements, working in con­ junction with status inconsistency, predispose a "left" or a "right" political response. It is possible that the crucial factor may be certain configurations of inconsistent statuses as opposed to others. Both Lenski and Jackson found that high occupation or education com­ bined with low racial-ethnic status predisposed status inconsistents to a "left" or liberal political response. l The present study indicates that ~ educational status may be a salient factor in predisposing a "r ight fl political response. As indicated in Chapter VI, it is possible that the combination of high education and 1m., income is responsible for the low percentage difference between high and low crystallized groups in column (g) and the reversal of predicted percentage differ­ ence in column (c) of Table 17 (page 123). These observations regarding education also point to its implications for social change, especially with reference to vertical mobility. Education may be regarded as a nonreversible achieved status. That is, it is possible to gain an education, but it is not possible to lose it. With educational opportunities becoming available to an in­ creasingly larger segment of the population, it is possible that upward educational mobility may be a more significant indicator of IJackson, op. cit. and Lenski, op. cit. 139 social change than occupational mobility, since the former often pre­ cedes upvlard occupational changes. It would appear, therefore, that the role of educational status is an important one in determining the possible reaction to status inconsistency. In conclusion, even in light of the relatively small amount of research that has been done using the status crystallization or status consistency model, it appears to be an important and useful predictive tool for the social sciences. This conviction is shared by Lenski,l who pioneered and stimulated much of the research in this area. Considering the variables to which this predictive model has been applied (psychosomatic symptoms of stress, social isolation, desire for change, upward social mobility, motivation to action, politi­ cal liberalism, and Right-Wing Extremism), the implications for the crystallization model for the study of a ,·Tide range of "deviant ll (both social and psychological) behavior are considerable. lpersonal communication, April 11, 1962. In part, Lenski wrote, • I am more convinced than I "l'la·s in the mid-50s that theII. • concept of status crystallization has real utility.1I 140 BIBLIOGRAPHY Books Adorno, T. W., et al. The Authoritarian Personality, New York: Harper and Brothers, 1950. Agar, Herbert. The Price of Power: America Since 1945, Chicago: The University of Chicago Press, 1957. ---­ ,II il Alpert, Harry. Emile Durkheim and His Sociology, New York: Russell and Russell, Inc., 19bI:--- Aptheker, Herbert. The Era of McCarthyism, New York: Marzani and Munsell, Inc:, 1962:"" Arendt, Hannah. The Origins of Totalitarianism, New York: Meridian Books, Inc., 1958. ­ Arnold, Thurman. The Folklore of Capitalism, lITe", Haven: Yale University Press, 1959: Barbu, Zevedei. Democracy and Dictatorship: Their Psychology and Patterns of Life, New York: Grove Press, 1956. --­ Becker, Howard and Boskoff, Alvin (eds.). Modern Sociological Theory in Continuity and Change, New York: Holt, Rinehart and 11inston, Inc., 1957. Bell, Daniel (ed.). The NevI American Right, lITe'''' York: Criterion Books, 1955. -- ­ ~ End of Ideology, New York: Collier Books, 1961. (ed.). The Radical Right (The New American Right expanded and updated), Garden City, NeTll York: Doubleday-aDd Company, Inc., 1963. Bendix, Reinhard. Max Weber: An Intellectual Portrait, Garden City, NevT York:--noubleday and Company, Inc., 1960. Bendix, Reinhard, and Lipset, S. M. (eds.). Class, Status and Power A Reader in Social Stratification, Glencoe: The Free Press, 1953. -­ 141 Berelson, B. R., Lazarsfeld, P. F., and McPhee, W. N. Voting, Chicago: University of Chicago Press, 1954. Burdick, Eugene, and Brodbeck, Arthur J. (eds.). American Voting Behavior, Glencoe: The Free Press, 1959. Burtt, E. A. The Metaphysical Foundations of Modern Physical Science, Garden City, New York: Doubleday-and Company, Inc., 1955. Carr, Edward Hallett. ~ _N_e_w _S_o_c_i_e_t~y, Boston: Beacon Press, 1957. ., TCash, 1,.\. v. ~ Mind of the South, New York: Vintage Books, 1960. Christie, Richard, and Jahoda, Marie (eds.). Studies in the Scope and Method of liThe Authoritarian Personality, fi (COnTInuities in Social Research, No.2), Glencoe: The Free Press, 1954. Cochran, William G. Sampling Techniques, New York: John Wiley and Sons, Inc., 1953. Cohen, Morris R., and Nagel, Ernest. An Introduction to Logic and Scientific l'lethod, New York:--Harcourt, Braceand Company, 193~. Cook, Fred J. ~ Warfare State, New York: The Macmillan Company, 1962. Deming, William Edwards. Some Theory of Sampling, New York: John Wiley and Sons, Inc., 1950. -- Dollard, John, et al. Frustration and Aggression, New Haven: Yale University Press, 1961. --­ Donner, Frank J. The Un-Americans, New York: Ballantine Books, 1961. Durkheim, Emile. The Rules of Sociological Method, eighth edition, Translated by SarahlA. Solovay and John H. Mueller, Glencoe: The Free Press, 1938. The Elementary Forms of the Religious Life, Translated by Joseph Ward Swain, Glencoe: The Free-Press, 1954. Edwards, Allen L. Techniques of Attitude Scale Construction, New York: Appleton-Century-Crofts, Inc., 1957. Ellsworth, Ralph E., and Harris, Sarah M. The American Right Wing: A Report to the Fund for the RepublIC; 1vashington, D.C-.-:-­ Public Affairs Pre s~1962. 142 Feigl, Herbert, and Brodbeck, May (eds.). Readings in the Philosophy of Science, New York: Appleton-Century-Crofts, Inc., 1953. Forster, Arnold and Epstein, Benjamin R. Danger on the Right: The Attitudes, Personnel and Influence of the Radical Righ~nd Extreme Conservatives~ew York: Random House, 1964. --­ Francis, Roy G. The Rhetoric of Science: A Methodological Discussion of the Two By Two Table, Minneapolis: The University of :t-Unnesota :PresS;-196L""" Fromm, Erich. Escape From Freedom, Nel,r York: Holt, Rinehart and Winston, Inc~941. Fruchter, Benjamin. Introduction to Factor .~alysis, Princeton, New Jersey: D. Van Nostrand-Company, Inc., 1954. Gerth, H. H. and Mills, C. Wright (Translators and Editors). From Max Weber: Essays in Sociology, New York: Oxford University--­ Press, 1958. -- --- --- Girvetz, Harry K. From Wealth to Welfare: The Evolution of Liberalism, Stanford,-caIifornia:--Stanford University Press:-1950. Goldman, Eric F. The Crucial Decade: America, 1945-1955, New York: Alfred A. Knopf, 1956. Green, Robert W. (ed.). Protestantism and Capitalism: The Weber Thesis and its Critics, (Problems in European Civilization, Number 3), Boston: D. C. Heath and Company, 1959. Grove, Gene. Inside the John Birch Society, Greenwich, Connecticut: Fawcett Publications, Inc., 1961. Hansen, Morris H., HurWitz, William N., and Madow, William G. Sample Survey Methods and Theory, Volumes I and II, New York: John Wiley and Sons, Inc., 1953. Harman, Harry H. Modern Factor Analysis, Chicago: The University of Chicago Press, 1960. Herberg, Will. Protestant, Catholic, Jew, New York: Doubleday and Company, Inc., 1960. --- Hoffer, Eric. The True Believer, New York: Mentor Books, 1958. Hollingshead, August B. Elmtown's Youth: The Impact of Social Classes ~ Adolescents, New York:~n Wiley and Sons, 1949. 143 Hollingshead, August B., and Redlich, Fredrick C. Social Class and Mental Illness: A Community Study, New York: John Iviley and Sons, Inc., 1958. Holzinger, Karl J. Student Manual of Factor Analysis, Department of Education, University of Chicago, 1937. Holzinger, Karl J., and Harman, Harry H. Factor Analysis: A Synthesis of Factorial Methods, Chicago: The University of Chicago Press, 1941. Homans, George C. Social Behavior: Its Elementary Forms, New York: Harcourt, Brace and World, Inc., 1961. Horowitz, Irving Louis (ed.). Power, Politics and People: The Collected Essays of Q. vIright Mills, Ne'·' York: Ballantine Books, 1963. Hyman, Herbert. Survey Design and Analysis: Principles, Cases and Procedures, Glencoe: The Free Press, 1955. ----- --- Janson, Donald and Eismann, Bernard. The Far Right, New York: McGraw­ Hill Book Company, Inc., 196~ --- ----- Kahl, Joseph A. The American Class Structure, New York: Holt, Rinehart and Winston, 1953. Katz, Elihu, and Lazarsfeld, P. F. Personal Influence, Glencoe: The Free Press, 1955. King, C. Wendell. Social Movements in the United States, New York: Random House, 1956. -- --- Kornhauser, Hilliam. The PoUtics of ~J1ass Society, Glencoe: The Free Press, 1959.--- -- ---- Krech, DaVid, Crutchfield, Richard S. and Ballachey, Egerton L. Indi­ vidual in Society: A Textbook of Social Psychology (A major reVision-of Theory andProblems-of Social Psychology by David Krech and Richard S:-Crutchfield~New York: McGraw-Hill Book Company, Inc., 1962. Lazarsfeld, P. F., Berelson, B. R. and Gaudet, Hazel, The People's Choice, 2nd edition, New YOrk: Columbia University Press, 1948. Lazarsfeld, Paul F., and Rosenberg, Morris (eds.). The Language of Social Research: A Reader in the Methodology of Social-­ Research, Glencoe:- The Free-Press, 1955. -- Lenski, Gerhard. The Religious Factor, New York: Doubleday and Com­ pany, Inc., 1961. 11 11+4 I Lerner, Max (ed.). The Portable Veblen, New York: The Viking Press, 1948. ­ Levin, Murray B. The Alienated Voter: Politics in Boston, New York: Holt, Rinehart and \rlinston, Inc., 1960.- Lippmann, Walter. The Public Philosophy, New York: The New American Library, 1956. Lipset, S. M. Political Man, Garden City, Ne't'l York: Doubleday and Company, Inc., 1960. Lipset, S. M. and Bendix, R. Social Mobility in Industrial Society, Berkeley and Los Angeles: UniversitY-of California Press, 1960. Lynd, Robert S. and Helen Merrel. Middletown, New York: Harcourt, Brace and World, Inc., 1929. Middletown in Transition, New York: Harcourt, Brace and Company, 1937":" Merton, Robert K. Social Theory and Social Structure, Glencoe: The Free Press, 1957. --­ Merton, Robert K., Broom, Leonard, and Cottrell, Leonard S., (eds.). Sociology Today: Problems and Prospects, New York: Basic Books, Inc., 1959. --­ Mills, C. Wright. The Power Elite, New York: Oxford University Press, 1956. --- ----­ Mueller, John H. and Schuessler, Karl F. Statistical Reasoning in Sociology, Boston: Houghton Mifflin Company, 1961. - Myrdal, Gup~ar. Value in Social Theory: A Selection of Essays on Methodo~ New York: Harper and Brothers,-r958. Nagel, Ernest. The Structure of Science: Problems in the Logic of ScientIfIc Explanation, New York: Harcourt: Brace and -­ World, Inc., 1961. Pope, Liston. Millhands and Preachers, New Haven: Yale University Press, 1942. --- Rapoport, Anatol. Operational Philosophy: Integrating Knowledge and Action, New York: Harper and Brothers, 1953. 145 Reichenbach, Hans. The Rise of Scientific Philosophy, Berkeley and Los AngeleS:- University of California Press, 1951. Riley, Matilda White, et al. Sociological Studies in Scale Analysis: Applications-,-Theory, Procedures, New BrUnSwick, New Jersey: Rutgers University Press, 1954. Rokeach, Milton. The Open and Closed Mind: Investigations into the Nature of Belief SYStems and-personality Systems, New York: Basic BOOks, 1960. --­ Rosenberg, Morris. Occupations and Valu.es, Glencoe: The Free Press, 1957. --- Rossiter, Clinton. Conservatism in America, New York: Alfred A. Knopf, 1955. -- Rovere, Richard H. Senator Joe McCarthy, Cleveland, Ohio: The World Publishing Company-rMeridian Books), 1960. Saloutos, Theodore. Farmer Movements in the South: 1865-1933, (Uni­ versity of California Publications in History, Volume 64), Berkeley and Los Angeles: University of California Press, 1960. Sherif, Muzafer. The Psychology of Social Norms, New York: Harper and Brothers, 1936. -- ----- Sorokin, Pitirim A. Society, Culture, and Personality: Their Structure and Dynamics, New York: Harper and Brothers, 1947. Stein, Maurice R., Vidich, Arthur J., and White, David Manning, (eds.). Identity and Anxiety: Survival of the Person in Mass Society, Glencoe: ~e Free Press, 1960. -- --- -- ---- Stonequist, Everett V. The MaTginal Man: A Study in Personality and Culture ConfliCt: New York:~ussell and Russell, Inc., 1961. Stouffer, Samuel A., et a1. Measurement and Prediction, Volume IV of Studies in sacral Psychology i~rld War II, Princeton, New Jersey: Princeton University Pres~1950. Tavmey, R. H. Religion and the Rise of Capitalism, Hest Drayton, Middlesex: Pelican Books, 1938. Taylor, Telford. Grand Inquest: The Story of Congressional Investiga­ ~, New York: Ballantine Books-,-1961. 146 Thurstone, L. L. Multiple Factor Analysis: A Development and Expan­ sion of The Vectors of Mind, Chicago: The University of ChI'Cago Press, 1947.--­ The Measurement of Values, Chicago: The Unive rsity of Chicago Press, 1959. Thurstone, L. L. and Chave, E. J. The Measurement of Attitude, Chicago: The University of Chicago Press-,-1929. Tryon, Robert Choate. Cluster Analysis: Correlation Profile and Orthometric (Factor) Analysis for the Isolation of Unities in Mind and Personality, Ann Arbor:;Michigan: Edwards Brotters::Inc., 1939. Turner, Ralph M. and Killian, Lewis M. Collective Behavior, Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1957. U. S. Bureau of the Census. Alphabetical Index of Occupations and Industries, prepared by Alba M. EdWardS: U. S. Department of Commerce, Washington, D. C.: U. S. Government Printing Office, 1940. U. S. Census of Population: 1960, Detailed Characteristics, Oregon, Final Report PC tl) - 39D, Washington, D. C.: U. S. Government Printing Office, 1962. Vahan, Richard. The Truth About the John Birch Society, New York: MacFadden Books, 1962. -- -- -­ Warner, W. Lloyd, et al. Social Class in America: The Evaluation of Status, New-York: Harper and~rothers, 19bO: Washburne, Norman F. Interpreting Social Change in America, New York: Random House, 1954. ­ Weber, Max. The Protestant Ethic and the Spirit of Capitalism, trans­ lated by Talcott Parsons-;-New York: Charles Scribner's Sons, 1958. 'V'Telch, Robert. The Blue Book of the John Birch Society, Sixth Printing, 1961. - -­ -­ - -­ -­ --­ West, James. Plainville, U. Press, 1945. - S. A., New York: - - Columbia University vThite, Winston. Beyond Conformity, New York: Glencoe~ Inc., 1961. The Free Press of 147 Wolff, Kurt (ed.). The Sociology of Georg Simmel, Glencoe: The Free Press, 1950. -­ Yule, G. Udny and Kenda1J, M. G. An Introduction to the Theory of Statistics, fourteenth edition, revised and--erilarged, New York: Hafner Publishing Company, 1950. Zaleznik, A., Christensen, C. R. and Roethlisberger, F. J., with the collaboraticn of George C. Homans. The Motivation, Product­ ivity and Satisfaction of Workers: A Prediction Stu~-- 13'O'St'On-:-Harvard UniverSity Press, 1958. --- Zeisel, Hans. Say It With Figures, revised fourth edition, New York: HarperandBrot'hers, 1957. Zetterberg, Hans L. On Theory and Verification in Sociology, New York: The Tressler-Press, 1954. -­ Journal and Periodical Articles Adams, Stuart. "status Congruency as a Variable in Small Group Perform­ ance," Social Forces, 32 (October, 1953), pp. 16-22. Alpert, Harry. "Operational Definitions in Sociology,lI American Sociological ReView, 3 (December, 1938), pp. s55=Ebr:­ 'America's Future, Inc. New Rochelle, New York, selected pamphlets. Benoit-Smullyan, Emile. "Status, Status Types, and Status Interrel­ ations," American Sociological Review, 9 (April, 1944), pp. 151-161. Bible Recordings, Inc. Baltimo:;:-e, Maryland, selected pamphlets. Blau, Peter N. "Social Nobility and Inter-Personal Relations," American Sociological Review, 21 (June, 1956), pp. 290-295. Christie, Richard and Cook, Peggy. "A Guide to Published Literature Relating to the Authoritarian Personality Through 1956," Journal of Psychology, 45 (April, 1958), pp. 171-199. Cook, Fred J. "The Ultras: Aims, Affiliations and Finances of the Extreme Right," The Nation, 194 (June 30, 1962). DaviS, Kingsley. "A Conceptual Analysis of Stratification," American Sociological Review, 7 (June, 1942), pp. 309-321. 148 Dynes, Russell R. "Church-Sect Typology and Socia-Economic Status," American Sociological Review, 20 (October, 1955), pp. 555-560. Fenchel, G. H., Nonderer, J. H. and Hartley, E. L. "Subjective Status and the Equilibration Hypothesis," Journal of Abnormal and Social Psychology, 46 (October, 1951), pp. 476-479. --­ Gibbs, Jack P. and Martin, Walter T. "A Theory of Status Integration and Its Relationship to Suicide," American Sociological Review, 23 (April, 1958), pp. 140-147. Goffman, Irwin W. "Status Consistency and Preference for Change in Power Distribution," American Sociological Review, 22 (June, 1957), pp. 275-281. Gouldner, Alvin W. "Cosmopolitans and Locals: Toward an Analysis of Latent Social Roles," Parts I and II, Administrative Science Quarterly, 2 (December, 1957 and March, 1958), pp. 281-306 and 444-480. Rield, Wayne. "The Study of Change in Social Science," British Journal of Sociology, 5 (March, 1954), pp. 1-11. Romans, George C. "The Cash Posters: A Study of a Group of Working Girls," American Sociological Review, 19 (December, 1954), pp. 724-733. "Social Behavior as Exchange," American Journal of Sociology, 63 (May, 1958), pp. 597-607. -- Hotelling, Harold. "Analysis of a Complex of Statistical Variables into Principle Components," Journal of Educational Psychology, 24 (September and October, 1933), pp. 417-441 and 498-520. House Committee on Un-American Activities, selected transcripts of testimony. Hughes, Everett C. "Dilemmas and Contradictions of Status," American Journal of Sociology, 50 (March, 1945), pp. 353-359. "Social Change and Status Protest: An Essay on the Marginal Man," Phylon, 10 (First Quarter, 1949), pp. 58-65. Independence Foundation, Inc. Portland, Indiana, selected pamphlets. Jackson, Elton F. "Status Consistency and Symptoms of Stress," American Sociological Review, 27 (August, 1962), pp.469-480. 149 Johnson, Benton. "Ascetic Protestantism and Political Preference," Public Opinion Quarterly, 26 (Spring, 1962), pp. 35-46. "Ascetic Protestantism and Political Preference in the Deep South," "~erican Journal of Sociology, 69 (January, 1964), pp. 359-366. -- Kenkel, William F. "The Relationship Between Status Consistency and Politico-Economic Attitude s," American Sociological Reviei\T, 21 (June, 1956), PI'. 365-368. Landecker, Werner S. "Class Crystallization and Its Urban Pattern," Social Research, 27 (Autumn, 1960), PI'. 308-320. Lenski, Gerhard E. "American Social Classes: Statistical Strata or Social Groups?" American Journal of Sociology, 58 (September, 1952), PI'. 139-144. -­ "Status Crystallization: A Non-Vertical Dimension of Social Status," American Sociological Review, 19 (August, 1954), PP. 405-413. . "Comment" [On \\Tilliam F. Kenkel, "The RelationShip Betvleen -------- Status Consistency and Politico-Economic Attitudes"?, mnerican Sociological Review, 21 (June, 1956), pp. 368-369.- -------­ ttSocial Participation and Status Crystallization," American Sociological Review, 21 (August, 1956), pp. 458-464. Likert, Rensis. tt A Technique for the Measurement of Attitudes," Archives of Psychology, No. 140 (June, 1932). Lipset, S. M. "Socialism--Left and Right--East and Hest," Confluence, 7 (Summer, 1958), PI'. 173-192. "Democracy and Working-Class Authoritarianism," American Sociological Review, 24 (August, 1959), pp. 482-501. "Social Stratification and 'Right-Wing Extremism'," British Journal ~ Sociology, 10 (December, 1959), pp. 1-38. "Reply" fio Schegloff and Kruytbosch, "Some Comments on . 'Working-Class Authoritarianism"7, Berkeley Journal of Sociology, 6 (Spring, 1961), pp.-l07-l08. ttThe Value Patterns of Democracy: A Case Study in Compar­ ative Analysis," American Sociological Review, 28 (August, 1963), PI'. 515:531. -----­ 150 "'vTorking-Class Authoritarianism r : A Reply to Miller and Riessman," British Journal of Sociology, 12 (September, 1961), pp. 277-281. - Lipset, S. M. and Bendix, Reinhard. "Social Status and Social Struc­ ture: A Re-Examination of Data and Interpretations: I and II," British Journal of Sociology, 2 (June and September, 1951), pp. 150-168 an~230-254. Maslow, A. H. "The Authoritarian Character Structure," Journal of Social Psychology, 18 (November, 1943), pp. 401-411-.--­ Mayer, Albert J. and Sharp, HaTry. "Religious Preference and Worldly Success," American Sociological Revie'w, 27 (April, 1962), pp. 218-227. McClosky, Herbert. "Conservatism and Personality," American Political Sc~~ Review, 52 (March, 1958), pp. 27-4~------ McDill, Edward L. "Anomie, Authoritarianism, Prejudice, and Socio­ Economic Status: An Attempt at Clarification," Social Forces, 39 (March, 1961), pp. 239-245. McDill, Edward L. and Ridley, Jeanne Clare. "Status, Anomia, Political Alienation, and Political Participation," American Journal of Sociology, 68 (September, 1962), pp. 205-213. Miller, S. M. and Riessman, Frank. "'Working-Class Authoritarianism': A Critique of Lipset," British Journal of Sociology, 12 (September, 1961), pp. 263-273. Mischler, Elliot G. "Personality Characteristics and the Resolution of Role Conflicts,1I Public Opinion Quarterly, 17 (Spring, 1953), pp. 115-135. Nagi, Saad Z. "Status Profile and Reactions to Status Threats,1I Ameri­ can Sociological Review, 28 (June, 1963), pp. 440-443. -----­ National Opinion Research Center. "Jobs and Occupations: A Popular Evaluation," Opinion~, 9 (September 1, 1947), PP. 3-13. Parsons, Talcott. "Some Considerations on the Theory of Social Change," ,Rural Sociology, 26 (September, 1961), pp. 219-239. Putney, Snell and Middleton, RusselL "Dimensions and Correlates of Religious Ideologies,1I Social Forces, 39 (May, 1961), pp. 285-290. 151 Rhodes, Lewis. "Anomia, Aspiration, and Status,rt Social Forces, 42 (May, 1964), Pl'. 434-440. Ringer, Benjamin B. and Sills, David L. rtpolitical Extremists in Iran: A Secondary Analysis of Communications Data," Public Opinion Quarterly, 16 (Winter, 1952-1953), PI'. 689-701. Robinson, H. S. "Ecological Correlations and the Behavior of Individ­ uals," American Sociological Eeview, 15 (June, 1950), Pl'. 351-357. Schegloff, Emanuel A. and Kruytbosch, Carlos. "Some Comments on 'Horking-Class Authoritarianism' ," Berkeley Journal of Sociology, 6 (Spring, 1961), PI'. 99-105. Selvin, Hanan C. "Durkheim's 'Suicide' and Problems of Empirical Re­ search," American Journal of Sociology, 63 (May, 1958), Pl'. 607-619. -- Selvin, Hanan C. and Hagstrom, Warren O. "Determinants of Support for Civil Liberties," British Journal of Sociology, 11 (March, 1960), Pl'. 51-73. -­ Siegel, Saul M. "The Relationship of Hostility to Authoritarianism," Journal of Abnormal and Social Psychology, 52 (May, 1956), PI'. 368-372. --­ Spearman, Charles. "General Intelligence Objectively Determined and Measured," American Journal of Psychology, 15 (1904), PI'. 201-293. -­ Stark, Rodney. "Class, Radicalism, and Religious Involvement in Great Britain," American Sociological Review, 29 (October, 1964), Pl'. 698-706. Stone, G. P. and Form, 1'1. H. "Instabilities in Status," American Sociological Review,.18 (April, 1953), PI'. 149-162. Stouffer, Samuel A. "An Analysis of Conflicting Social Norms," American Sociological Review, 14 (December, 1949), Pl'. 707-717. "Some Observations on Study Design," American Journal of Sociology, 55 (January, 1950), PI'. 355-361.- -­ Taylor, Janet A. "A Personality Scale of Manifest AnXiety," Journal of Abnormal and Social Psychology, 48 (April, 1953), PI'. 285- -­ 290. ­ 152 The Twentieth Century Reformation Hour, CollingsvlOod, Ne"l1 Jersey, selected radio broadcasts and pamphlets. Toby, Jackson. "Universalistic and Particularistic Factors in Role Assignment," A'Tlerican Sociological Revie"T, 18 (April, 1953), PP. 134-141. TrOiv, Martin. "Small Businessmen, Political Tolerance, and Support for McCarthy," American Journal of Sociology, 64 (November, 1958), pp. 270-281. -- Hengert, Robert. "Reply" [to Schegloff and Kruytbosch, "Some Comments on 'Working-Class Authoritarianism' ,"7, Berkeley Journal of Sociology, 6 (Spring, 1961), pp. l09~112. -­ T:Jestin, Alan F. "The Deadly Parallels: Radical Right and Radical Left," Harper's Magazine, 22'-+ (April, 1962), pp. 25-32. Unpublished Manuscripts Bisno, Herbert and Gibbs, Jack p. "A Macrocosmic Analysis of Socio­ pathic Deviation, " (unpublished manuscript). Frost, Leslie. "Social Conflict, Social Systems Theory, and Contemp­ orary Sociology," (unpublished Honors thesis, Harvard University, 1960). Jackson, Elton F. "Status Consistency, Vertical Mobility and Symptoms of Stress," (unpublished Ph.D. Dissertation, University of Michigan, 1960). "Sampling Methods Used in Eugene-Springfield Hetropolitan Area Study of 1958," Institute for Community StUdies, Eugene, Oregon: University of Oregon~mimeographed). Selvin, Hanan C. "Problems in the Use of Individual and Group Data," Berkeley, California: University of California, 1960, (mimeographed) • Stember, Charles H. "Anti-Democratic Attitudes in America: A Revie", of Public Opinion Research," Bureau of Applied Social Re­ search, NeVI York: Columbia Unive:csity Press-,-1951+, (mimeo­ graphed) • Syme, Sherman Leonard. "The Concept of Status Crystallization: An Exploratory Analysis," (unpublished Master's Thesis, Univ­ ersity of California, Los Angeles, June, 1955). 153 APPENDIX A THE SAMPLE The data used in the present study vias obtained from a sample of the Eugene-Springfield metropolitan area, drawn by the Institute for Community Studies at the University of Oregon in the Summer of 1959. Sampling was done in five residential areas of the Eugene-Springfield metropolitan complex: Eugene, Springfield, East Springfield, the River Road area, and the Ferry Street Bridge area. In the following dis­ cussion, the different combinations of probability sampling techniques used in each area will be discussed separately.l Eugene. The sampling procedures used in Eugene combined propor­ tional stratified sampling, multi-stage cluster sampling, systematic sampling and simple random sampling. Strata were formed according to the socio-economic status of the individuals within the Eugene popula­ tion. This was effected by the use of proposed census enumeration districts, as drawn up by the Lane County Planning Commission, for the sampling frame. These enumeration districts corresponded roughly to neighborhoods employing similar land use. The rationale for this lThis discussion is based mainly on "Sampling Methods Used in Eugene-Springfield Metropolitan Study of 1958," Institute for Community Studies, Eugene, Oregon: University of Oregon (mimeographed). 154 . .._--­ procedure was that it would delimit areas within which individuals have homogeneous socia-economic status. This use of contiguous land area ta classify socio-economic status was suggested by Cochran's ob servation: In surveys which cover a geographic region, adjacent units are often more alike than units that are far apart. • .. People in the same uart of town tend to be of similar economic level and have certain things in co~~on in their attitudes and tastes. 1 From each of these stratified enumeration districts, a proportionate random sample of blocks was taken. From the blocks thus selected, each interviewer took a systematic sample of households, alternating male and female respondents from each selected household. Springfield. In this area, a systematic sampling of every seventh household was obtained, in the following manner. On a utility company map showing every house within the Springfield city limits every block within the city was numbered in a "serpentine" order. Then, a system­ atic sample of every seventh h~~se was taken, moving in numerical order block by block. To compensate for multiple dwelling units, a map from the Lane Count~r Planning Commission showing all multiple dwelling units in Springfield was used. The number of households within each of these dwelling units 'VlaS indicated on the map, and every seventh one of these additional households was added to the original sample. Interviewers were given the specific address of each household lWilliam G. Cochran, Sampling Techniques, New York: John Wiley and Sons, 1953, p. 96. 155 in the sample, and were to systematically alternate male and female re spondent s • River Road and Ferry Street Bridge Areas. Eugene Water and Elec­ tric Board meter books (which indicated all users of water and elec­ tricity "rithin these areas) were used as the basis for a systematic sample of households in these areas. Since most multiple dwellings in these areas had separate meters for each household, it was felt that the use of meter books gave a total population of all households. In taking the systematic sample, pages which indicated the customer to be a commercial establishment rather than a residential household were eliminated. The sampl!_ng proportions of households to be interviewed within the River Road and Ferry street Bridge areas were thirty-three and twenty-two respectively. These numbers were arrived at by dividing the number of interviews desired in each area plus tw'enty per cent into the number of households in each area (obtained from a then recent survey by the Bureau of Municipal Research, University of Oregon. East Springfield. In spite of the slight complications introduced by the fact that this area was serviced by three utility companies, the same method of sampling employed in the River Road and Ferry Street Bridge areas was used. That is, a systematic sample of hOllseholds was drawn from utility company meter records. The sampling proportion used in this area was every tenth household. This pattern was deviated from on the McKenzie Highway, where every twentieth name was used. A smaller sample was taken from this specific area because it contained many commercial establishments, which were not differentiated from 156 residential utility users in the files of the Pacific Power and Light Company. The report by the Institute of Community Studies on the samp­ ling methods used in their study is conc1nded as follows: Sampling devices used in Springfield, River Road, Ferry Street Bridge and East Springfield were employed primarily for their simplicity and their ease of access, and inclus­ iveness of the total population of elements we were interested in. Using EDs {enumeration district!7 seemed extremely inappropriate in the fringe areas due to large variation of certain types of individuals living within each ED. The type of systematic sampling employed in Springfield gives us estimates of the population parameters closely approximating those that would be secured from a proportionate stratified sample. It should be emphasized that although a variety of sampling techniques have been employed, all are of the probability type, suppl1ing us with measurements of the precision of estimates. 1llSampling Methods Used in Eugene-Springfield Metropolitan Study of 1958, II op. cit., p. 10. 157 APPENDIX B CLUSTER ANALYSIS DATAl The purpose of the following analysis is to discover the operational unities among the thirty attitudinal indicators of the Extreme Right being studied. The data consists of 435 correlation coefficients between these variables, which are laid out in Table 18, page 159. 2 As discussed in Chapter IV, page 71, extremist and non- extremist responses to these items were dichotomized by placing f1Strong­ ly Agree" or f1Strongly Disagree fl responses (according to the contenl of the item) in the extremist category and the remaining responses in the non-extremist category. Since this method of dichotomizing responses may be expected to yield disparate marginal frequencies in the comput­ ation of inter-item correlations (i.e., a low proportion of extremist lThe method of cluster analysis used herein is that outlined in Robert Choate Tryon, Cluster Analysis: Correlation Profile and Orthometric (Factor) Analysis for the Isolation of Unities in Mind and Personality, Ann Arbor, Michigan: Edvlards Brothers, Inc., 1939:--­ Further reference to this source will be given in the text as "Tryon" and cited in parentheses with relevant page numbers. 2The comnlete correlation matrix in Table 18 consists of coefficient entrie~ in 870 cells, as determined by the formula (n2 - n), where n equals the number of variables defining the matrix. The-30 ­ principal diagonal cells (extending from upper-left to lower-right) are, of course, subtracted in this computation, since they represent the correlation of each variable vlith itself. Since Table 18 is a sym­ metric matrix (having equal values in corresponding positions relative to the principal diagonal), there are actually (n2/2 - n/2), or 435, discrete variables in the matrix. -­ 158 to non-extremist responses), it vlOuld appear that Yule's "Q" coeffic­ ient provides the best measure of association for the correlation matrix, since "Q" remains stable under changes in marginal ratios.l Thus, shifts in marginal ratios, 'dhich may be expected from even slight changes in the number of extremist to non-extremist responses, will not be reflected in the correlations, "Q" being based on cell ratios. Most of the variables under study appear to be highly and positively intercorrelated. Reference to the sums of the intercorrel­ ations for each variable with all other variables (Lr in Table 18) indicates that there are four exceptions. Because of its extremely low correlation with all other variables, item 51 (~r = 4.62) ,,,ill be elim­ inated from the analysis at this time. Items 21 (~r =7.06), 44 (~r :' 8.29) and 47 (!r = 8.35) I·rill be retained, hovlever. In spite of fairly 10v1 overall intercorrelation sums, these items are all quite highly correlated 'i1ith certain groups of variables. A possible exception to this is item 47, which may have to be eliminated later in the analysis. Before correlation profiles for clusters can be dravm, the variables will be sorted on the basis of the intercorrelations and assigned to groups by means of "B" coefficients. The first step in this operation is to list the intercorrelations between pairs of vari­ ables by size (cf. Tryon, pages 43-41~, 47). This distribution of IFor discussions of the properties of Yule's "Q", see G. Udny Yule and M. G. Kendall, An Introduction to the Theory of Statistics, fourteenth edition, New York: Hafner Pu~lishing Company, 1950, p. 30; John H. r-rueller and Karl F. Schuessler, Statistical Reasoning in Sociology, Boston: Houghton Mifflin Company, 1961, pp. 242-249, 258. - Variables 18 20 21 22 23 24 25 26 27 2'9 I - 18 ( .93) .48 .60 .39 .71 .61 .51 .53 .33 .4o 20 .48 ( .80) .70 .17 .47 .52 .80 .69 .78 .J.j 3 21 .60 .70 (.75) .24 .59 .66 .63 .75 .52 .34 22 .39 .17 .24 (.71) .18 .35 .59 .71 .18 .J.j 2 23 .71 .47 .59 .18 ( .83) .62 .60 .79 .76 -.13 24 .61 .52 .66 .35 .62 ( .87) .87 .71 .56 .47 25 .51 .80 .63 .59 .60 .87 (.87) .76 .87 .58 26 .53 .69 .75 .71 .79 .71 .76 ( .82) .82 .43 27 .33 .78 .52 .18 .76 .56 .87 .82 ( .87) ':< 8.~ .40 .43 .34 .42 -.13 .47 .58 .43 .38 (.77) 30 29 .15 .38 .09 .43 -.07 .34 .71 .38 .43 .77 31 .26 .29 .28 .39 .37 .32 .77 .65 .41 .73 32 .27 .25 .01 .39 -1.00 .40 .83 .44 .38 .E 3 33 .62 .32 .31 .63 .53 .16 .40 .62 .45 .51 34 .~.43 -.01 .31 .66 .73 .16 -.29 .16 .04 ':< 7 .92 .50 .58 .14 .81 .60 .79 .65 .57 .45 37 36 .93 .49 .49 .24 .83 .43 .80 .56 .45 .5o 38 .91 .67 .30 .46 .79 .57 .76 .37 .60 .46 ,.. .0::.93 .44 .06 .36 .76 .63 .76 .62 .56 839 40 .91 .52 -.14 .41 .65 .67 .87 .31 .60 .42 ,..41 .0::.33 -.06 -.15 .49 .50 .18 .31 .29 .26 2 43 .24 .07 -.19 .52 .45 .13 .22 .43 .21 .1.4 44 .25 .07 -.25 .52 .48 .12 .32 .51 .29 .1 3 45 .34 .51 .12 .28 .37 .44 .68 .46 .31 .4 5 . 46 .43 .33 .11 .48 .31 .55 .39 .41 .42 .49 ,.. .0:: -.17 .38 -1.00 .62 .38 -.01 .56 .44 .40 547 48 .46 .47 .45 .08 .15 .58 .62 .08 .32 .41 49 .43 .25 .42 .55 -.27 .23 .63 .05 .02 .57 50 .51 .59 .17 .23 .12 .54 .65 -.02 .32 .57 ,.. .0::.25 .25 .16 .13 -.09 .43 .38 -.16 .11 251 13.96 11.75 7.06 11.24 11.39 12.84 17.37 13.44 12.35 11.89 ,Er- (b) 14.89 12.55 7.81 11.95 12.22 13.71 18.24 14.26 13.22 12.66 Lr (a) - ---------- ------------ J"] 159 I i ~o .15 1. 38 .09 .43 .07 .34 .71 .38 .43 .77 .82) .82 .77 .44 .32 .38 .46 .15 .22 .20 .11 .17 .04 .40 .35 .36 .49 .56 .39 .08 .32 .14 TABLE 18. ORIGINAL CORRELATION MATRIX: 3l ~2 ~3 34 36 37 .26 .27 .62 .43 .92 .93 .29 .25 .32 -.01 .50 .49 .28 .01 .21 .31 .58 .49 .39 .39 .63 .66 .14 .24 .37 -1.00 .53 .73 .81 .83 .32 .40 .16 .16 .60 .43 .77 .83 .40 -.29 .79 .80 .65 .44 .62 .16 .65 .56 .41 .38 .45 .04 .57 .45 .73 .63 .51 .37 .45 .50 .82 .77 .44 .32 .38 .46 ( .82) .78 .39 .35 .54 .62 .78 ( .83) .50 .50 .41 .50 .39 .50 (.71) .76 .54 .66 .35 .50 .76 (.78) .40 .48 .54 .41 .54 .40 (.98) .98 .62 .50 .66 .48 .98 (.98) .68 .43 .68 .78 .93 .93 .51 .36 .65 .62 .93 .93 .57 .62 .71 .61 .94 .95 .07 .34 .54 .52 .13 .29 .25 . .22 .48 .69 .16 .34 .06 .10 .45 .60 .13 .31 .49 .54 .32 -.14 .50 .43 .48 .28 .71 .74 .62 .56 .24 .48 .48 .42 .19 .07 .49 .66 .34 -.23 .65 .53 .54 .58 .33 .25 .41 .42 .40 .70 .43 .26 .64 .55 -.11 -.08 -.05 .22 .16 .02 12.64 11.29 13.81 10.71 15.65 15.75 13.46 12.12 14.52 11.49 16.63 16.73 THIRTY VARIABLES 38 39 40 .91 .93 .91 .67 .44 .52 .30 .06 ' -.14 .46 .36 .41 .79 .76 .65 .57 .63 .67 .76 .76 .87 .37 .62 .31 .60 .56 .60 .46 .28 .42 .15 .22 .20 .68 .51 .57 .43 .36 .62 .68 .65 .71 .78 .62 .61 .93 .93 .94 .93 .93 .95 (.98) .98 .97 .98 (.99) .99 .97 .99 (.99) .60 .50 .59 .58 .54 .63 .57 .54 .63 .55 .38 .56 .69 .44 .62 .06 .20 -.01 .67 .55 .66 .47 .33 .57 .67 .56 .74 .23 .26 .36 17.51 15.89 17.13 18.49 16.88 18.12 41 .33 -.06 -.15 .49 .50 .18 .31 .29 .26 .22 .11 .07 .34 .54 .52 .13 .29 .60 .50 .59 (.78) .78 .72 .15 .48 .49 .13 .37 .23 .25 9.66 10.44 43 44 .24 .25 .07 .07 -.19 -.25 .52 .52 .45 .48 .13 .12 .22 .32 .43 .51 .21 , .29 .14 I I .13I .17 .04I I .25 .06 .22 .10 .48 .45 .69 .60 .16 .13 .34 .31 .58 .57 .54 .54 ' :~3 .63 .78 .72 ( .89) .89 .89 : ( .89) -.05 ; -.16 .•45 I ; .51 .55 I .50 -.15 I -.24I .36 .24I -.02 -.17I .09 .13 9.18 ' 8.29 10.07 I 9.18 45 .34 .51 .12 .28 .37 .44 .68 .46 .31 .45 .40 .49 .54 .32 -.14 .50 .43 .55 .38 .56 .15 -.05 -.16 (.77) .40 .77 .59 .58 .65 .19 11.11 11.88 46 .43 .33 .11 .48 .31 .55 .39 .41 .42 .49 .35 .48 .28 .71 .74 .62 .56 .69 .44 .62 .48 .45 .51 .40 (.77) .50 .69 .67 .77 .20 14.08 14.85 47 -.17 .38 -1.00 .62 .38 -.01 .56 .44 .40 .25 .36 .24 .48 .48 .42 .19 .07 .06 .20 -.01 .49 .55 .50 .77 .50 (.77) .17 .53 .41 .09 8.35 9.12 48 .46 .47 .45 .08 .15 .58 .62 .08 .32 .41 .49 .49 .66 .34 -.23 .65 .53 .67 .55 .66 .13 -.15 -.24 .59 .69 .17 (.93) .87 .93 .38 11.80 12.73 49 .43 .25 .42 .55 -.27 .23 .63 .05 .02 .57 .56 .54 .58 .33 .25 .41 .42 .47 .33 .57 .37 .36 .24 .58 .67 .53 .87 ( .91) ·91 .29 12.16 13.07 50 .51 .59 .17 .23 .12 .54 .65 -.02 .32 .57 .39 .40 .70 .43 .26 .64 .55 .67 .56 .74 .23 -.02 -.17 .65 .77 .41 .93 .91 (.93) .23 12.96 13.89 51 .25 .25 .16 .13 -.09 .43 .38 -.16 .11 .22 .08 -.11 -.08 -.05 .22 .16 .02 .23 .26 .36 .25 .09 .13 .19 .20 .09 .38 .29 .23 ( .43) 4.62 5·05 Lr- (a) 13.96 11.75 7.06 l1.24 11.39 12.84 17.37 13.44 12.35 11.89 10.32 12.64 11.29 13.81 10.71 15.65 15.75 17.51 15.89 17.13 9.66 9.18 8.29 11.11 14.08 8.35 11.80 12.16 12.96 4.62 (a) exc1. diag. (b) inc1. diag. ~~ l I f """'-_""""_...,jl_' ........", 160 variables is plotted in Table 19, page 161. The entries on the head of this table are correlation intervals. Thus, the correlation between variables 18 and 36 falls bebieen .90 and .9110, the correlation between variables 20 and 27 falls between .75 and .79, and so forth. This table facilitates the selection of variables to be assigned to tenta­ tive clusters by the "B" coefficient method. The "B" coefficient operation is based on the assumption that the variables related to a given factor should have higher intercorre­ lations between themselves than with other variables under analysis. The "B" coefficient ("coefficient of belonging") is defined by Fruchter as follo'\.;s: The B-Coefficient gives the ratio of the average inter­ correlations of the variables in a cluster to their average correlation with the variables not included in the cluster. A B-coefficient of 1.00 would indicate that the variables in the cluster correlate no more highly among themselves than they do with the vari­ ables outside of the cluster. l Basically, the "B" coefficient procedure involves the systematic intro­ duction, one at a time, of variables assumed related to a given group until the coefficient obtained shows a significant drop over that ob­ tained for the previous set. The calculation of "B" coefficients for the present study is given in Table 20, pages 164-166 (cf. Tryon, pages 44-45, 48). The grouping is ordinarily begun by selecting the two variables '\.;hich have the highest correlation, to '\.;hich are systeraatic­ ally added variables for '\oihich the sum of correlations with the preced­ lBenjamin Fruchter, Introduction to Factor Analysis, Prince­ ton, New Jersey: D. Van Nostrand Company, Inc., 1954, p. 14. _. - ----------- I' ! i I i . al J.l «S -.01 - -LOO .00 - .04 .05 - .09 .10 - .14 .15 - .19:> . 18 47 30 4J 20 34 41 43 44 22 32 40 41 43 ·i· 21 44 47 32 30 39 45 46 50 22: "22 48 36 20 23 27 2Li , 23 29 30 32 49 50 22 48 i 24 47 43 44 33 34 41 49 25 34 43 26 50 48 49 34 27 34 49 "\; 22 43' 29 23 43 44 41 I 30 23 44 21 18 38 41 40 43 39' 31 41 44 47 , 32 23 21 44 201 33 24 21 34 20 25 45 48 27 24 26 49} 36 22 41 44 43 47 .~: I 37 47 22 38 47 30 39 21 30 I 40 21 47 30 , 41 20 21 31 30 36 48 24 45 29 43 21 45 48 50 20 24 29 30 36 18 44 21 45 48 50 30 20 31 24 29 32 36 18 45 34 43 44 21 41 46 21 47 18 21 24 40 37 38 36 48 "29: 48 34 43 44 22 26 41 23 47 49 23 27 26 20 26 43 44 , 50 23 21 22 1 161 TABLE 19. DISTRIBUTION OF CORRELATI~NS OF EACH VARIABLE: TWENTY-NINE VARIABLES '0 - .24 .25 - .29 ·30 - .34 .35 - .39 Correlation Scale .40 _ .44 ,.45 - .49 .50 - .54 .55 - .59 .60 - .64 .65 - .69 .70 - .74 .'15 - .79 .80 _ .84 .85 - .89 .90 - .94 .95 - .99 . ~ > .4 49 31 32 31 27 41 45 33 46 2229 30 47 34 46 49 29 39 20 48 18, 23 37 48 25 26 50 24 36 40 45 21 50 24 33 21 26 38 23 , 25b 36 37 38 3940 18 20 33 37 50 '47 31 38 45 41 41 44 39 29 34 24 46 22 30 31 41 44 40 18 45 48 56 21 18 31 32 39 31 45 47 46 30 38 29 32 18 27 34 4849 29 30 '40 32 37 45 33 ~g ~~ 43 30 31 '46 47 20 22 26 36 40 48 37 I 38 41 46 j 204344 I I29 I I45 I 33 37 i 24 3B 45 46; 27 4344 33 41 20 46 50 18 1844 21 33 37 18 23 36 2549 21 27 36 38 48 22 29 47 37 24 36 38 3940 25 49 50 25 3347 24 25 40 18 23 39 ~~ ~6 48 31 33 36 39 32 20 24 34 21 40 45 20 26 26 18 34 26 30 21 22 24 31 26 ~7 38 39 I 20 ~6 31 36 37 ~8 39~3 5 20 3 30 36 37 32 27 2"6 25 24 27 40 25 21 22 23 24 25 26 27 29 24 34 46 ~ 26367 50 22 27 33 45 3748 l 49 25 29 ~2 • 31 30 18 20 21 43 24 34 22 23 33 27 50 45 46 48 36 39 49 40 26 37 38 29 25 2 30 31 43 18 46 41 20 45 48 49 22 27 39 31 24 26 36 25 30 44 50 34 37 47 27 43 47 33 45 23 29 32 36 41 49 2940 18 22 26 48 37 38 39 50 4046 30 [l1 34 : 25 32 33 50 21 30 31 29 36 1847 32 37 41 ' 39 40 44 22 43 23 46 3338 34 30 34 29 32 49 45 20 31 33 21 24 21 26 46 48 50 25 : 23 18 38 39 40 37 36 47 41 21 29 43 44 49 26 22 32 45 24 45 49 32 2046 ~g~~~ , 22 29 49 41 2948 31 43 44 48 26 46 50 24 27 41 43 44 45 27 50 31 24 26 34 33 ,~ ~§ §6 33 25 : 23 ~5 34 23 r5 23 18 38 39 18 36 37 18 36 37 36 40 3940 3840 37 36 39 26 22 29 20 ~n§ 41 ~nu~ 2448 33 50 25 16 36 37 38 39 40 50 26 27 37 18 25 32 49 22 39 46 47 23 33 34 3840 44 43 : 41 25 27 32 49 31 39 31 27 22 32 37 25 37 18 27 33 20 23 30 49 23 39 25 20 23 30 26 46 33 24 30 37 46 18 26 27 39 43 45 26 27 34 50 23 33 23 I 26 29 31 36; 22 29 31 41!, I 32 33 41 I' 22 39 22 26 39 46 47 20 32 24 44 47 44 46 49 3847 34 38 38 40 48 49 37 20 25 43 40 40 .36 40 34 25 50 38 48 49 41 33 34 41 I i 47 ! 50 i 45 44 43 43 44 45 46 47 27 33 2129 18 20 30 311; 37 39 2445 25 36 32 38 40 46 49 50 48 2434'44 33 39 4143 18 21 36 37 38 I' 31 47 . ~n6 ~g 25 46 48 50 49 41 34 27 30 31 33 47 I I 18 24 20 29 37 39 25 36 3845 32 40 46 4849 50 162 ing group are highest. In the present analysis,some groups consisting of items v.lith extremely high intercorrelations are be gun I·d th more than hlO variables. The follmving outline of the "B" coefficient procedure discusses the steps represented by columns (1) through (11) of Table 20. COL1J1I1N PROCEDURE (1) List of variables in the cluster, with ~ being the last added. (2) Sum of the correlations in column i (from Table 18, page 159). The first row in the analysis ~lill-contain the sums of the columns for all variables in the cluster (ordinarily two). (3) Sum of the correlations between variable i and those already in the cluster. ­ (4) Sum of all the correlation coefficients in the cluster. (5) Sum of the correlations of the variables in the cluster with the variables not in the cluster. (6) The number of variables in the cluster. (7) The number of intercorrelations among variables in the cluster. (8) The number of remaining intercorrelations (i.e., the number of intercorrelations between the variables in the cluster and those not in the cluster). (9) The average of the correlation coefficients in the cluster. (10) The average of the correlations between the variables in the cluster and the remaining variables. (11) The ratio of the average of the correlations in the cluster to the average of the remaining correlations of the variables in the cluster (i.e., the ratio of column 9 to column 10). This is the "B" coefficient. As indicated earlier, the procedure of adding variables is continued until the coefficient obtained shows a significant drop over that obtained for the previous set of variables. This procedure is illus­ 163 trated in the first set of computations in Table 20, page 164, where a drop in the "B" coefficient from L 958 to L 850 warrants the ex- elusion of variable 25 from the cluster of variables 36, 37, 38, 39, 40, 18 and 23. Although Holzinger and Harmon set the minimum signifi­ cant value of a "B" coefficient at L 30,1 some .iudgement as to what constitutes a "significant drop" must be exercised by the analyst. The small drop obtained when variable 26 is added to the cluster of vari­ ables 25, 27 and 20 is not as significant, for instance, as the large decrease recorded when variable 34 is added to the cluster of variables 43, Q,4 and 41 (cf. Table 20, page 165). Moreover, the tentative ex- elusion of an item by the "B" coefficient method is not conclusive evidence that it will not cluster with the group in question. The tentative clusters of variables discerned by the compu­ tation of "E" coefficients are summarized at the end of Table 20, page 166. The items indicated by a question mark (variables 22, 33, 34, 45, 46 and 49) are those variables which appear to be related to a given cluster, although the "B" coefficient obtained ,·!hen they "Tere included decreased over the previous coefficient for the set. These variables will receive further consideration when the mean correlation profiles of variables by clusters are calculated and plotted. Item 47 does not appear to be related to any of the tentative clusters. It will there­ fore be eliminated from further investigation and considered a "resid-' ualrr variable for this analysis. ~arl J. Holzinger and Harry H. Harmon, Factor Analysis: A Synthesis of Factorial Methods. Chicago: The University of Chicago Press, 1941; p. 27. ' TABLE 20. TENTATIVE ALLOCATION OF VARIABLES TO CIDSTERS BY THE "BII COEFFICIENT METHOD (1) Prec.e ding 9rouf 36 37 38 39 40 + i. Sum of Correlations: (2) (3) (4) (5) (6 ) (7) (8) (9) (10) (11) i. with i. with between be+w~en k = k (1<-/) k (f1-k)= (4h (7) (5)+(8) (~) ~ (10) n group varia-hies Vo.t'iA 6/f'$ nII"" b,u' 2. /c(2'-Ic) VQ.riQ.bles in 9rOIJP nof ,1'1 of =1'10. of ",eel nVAriable =no. I"flf'f)f'J 8 coefF. = (3) pluS grovp = of I'tmQ;f1. infer- J­(1)-2.(3) in re",4'n. preced - inJ~r -n infer- r~ r '''9 (4) + prec. (5 9rovP 81.93 9.53 9.53 62.87 5 10 120 .953 .524 1.819 36 37 38 39 40 18 13.96 4.60 14.13 67.63 6 15 138 .942 .490 1.922 36 37 38 39 40 18 23 11.39 4.55 18.68 69.92 7 21 154 .889 .454 1.958 36 37 38 39 40 18 23 25 17.37 5.09 23.77 77.11 8 28 168 .849 .459 1.850 Check: ~ (2) =124.65 = 2 (23.77) + 77.11 = 124.65 48 49 50 36.92 2.71 2.71 31.50 78 .903 I .403 I 2.2403 3 48 49 50 46? 14.08 2.13 4.84 41.32 4 6 100 . 806 1 .413 11.951 Check: [, (2) = 51.00:: 2 (4.84) + 41.32 =. 51.00 48 49 50 45? Check: :r (2) :a 11.11 48.03 1.82 4.53 38.97 J 4 = 2 (4.53) + 38.97 = 48.03 I 6 I 100 I .755 I .390 I 1.935 (CONTINUED NEXT PAGE) f-J CJ\ +;­ TABLE 20. (CONTINUED) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) 48 49 50 47 8.35 loll 3.82 37.63 4 6 100 .636 .382 1.664 Check: l (2) = 45.27 == 2 (3.82) + 37.63 = 45.27 25 27 25 27 20 25 27 20 26 25 27 20 26 21 25 27 20 26 21 24 29.72 11.75 13.44 7.06 12.84 .87 1.58 2.27 2.60 3.32 .87 2.45 4.72 7.32 10.64 27.98 36.57 45.47 47.33 53.43 2 3 4 5 6 1 3 6 10 15 56 78 100 120 138 .870 .816 .786 .732 .709 .499 .468 .455 .391 .385 1.743 1.743 1.727 1.872 1.841 Check: ~ (2) = 74.81 = 2 (10.61t) + 53.43 = 74.81 1+3 44 41 27.13 2.39 2.39 22.35 3 3 78 .796 .286 2.783 43 44 41 34? 43 44 41 34 33? 43 44 41 34 33 22? 10.71 13.81 11.24 1.81 2.23 2.82 4.20 6.43 9.25 29.44 38.79 44.39 4 5 6 6 10 15 100 120 138 .700 .643 .617 .294 .323 .322 2.380 1.990 1.916 Check: ~ (2) == 62.89 == 2 (9.25) + 44.39 = 62.89 (CONCLUDED 1~XT PAGE) I-' 0'\ \Jl l---­ Check: ~ (2) :: 60.22 :: 2 (6.10) + 48.02 :: 60.22 TABLE 20. (CONCWDED ) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) 30 31 22.96 .82 .82 21.32 2 1 56 .820 .380 2.157 30 31 32 11.29 1.55 2.37 29.51 3 3 78 .790 .378 2.089 30 31 32 29 30 31 32 29 46 11.89 14.08 2.13 1.60 4.50 6.10 37.14 48.02 4 5 6 10 100 120 .750 .610 .371 .400 2.021 1.525 30 31 32 29 45? 11.11 1.88 6.38 44.49 5 10 120 .638 .371 1.719 Check: ~ (2) = 57.25 = 2 (6.38) + 44.49 :: 57.25 30 31 32 29 49? 12.16 2.25 6.75 44.80 5 10 120 .675 .373 1.809 Check: ~ (2) :: 58.30 = 2 (6.75) + 44.80 = 58.30 TENTATIVE CLUSTERS: A: 36 37 38 39 40 18 23 D: 43 44 41 -­ 34? 33? 22? B: 48 49 50 -­ 46? 45? E: 30 31 32 29 -­ 45? 491 C: 25 27 20 26 21 24 RESIDUAL: 47 ~ 0' ~ 167 The next step in the cluster analysis will be to examine the congruence of each variable with'the tentative clusters determined by the "B" coefficient procedure. To this end, the mean correlation of each variable with each of the tentative clusters is calculated in Table 21, page 168 (cf. Tryon, pages 45-46, 49). The data for columns (1) to (5) of this table consists of the sum of the correlations of each variable with all other variables in its tentative cluster, including diagonal entries (from Table 18, page 159 above). For ex­ ample, the sum of correlations for variable 18 with group A (variables 18,23,36,37,38,39 and 40 is (.93) + .71 of .92 + .93 + .91 + .93 + .91 = 6.24. For columns (1) through (5), "s" represents the number of variables in each group. Each entry in coillmns (1) through (5) is then multiplied by "l/s" (e.g., l/s::. .143 for group A), the product of which is entered in a corresponding cell in columns (6) through (10). These mean correlations from columns (6) through (10) are next plotted as trial correlation profile s in Figure 3, page 169 (cf. Tryon, page s 46, 50). At this stage, additional relationships between the variables in the five tentative clusters may be noted. First, although item 23 appeared to be related to Tentative Cluster A on the basis of its "B" coefficient, its mean profile is markedly disparate from those of the other variables in this cluster. This item will therefore be excluded from further analysis. Second, item 46, a questionable variable in Tentative Cluster B, also shows considerable discrepancy from the other variables in its group, and will therefore be eliminated. Together "\'lith item 47 (eliminated from analysis by the "Btl coefficient procedure) 168 TABLE 21. MEAN CORRELATION BETlAJEEN EACH VARIABLE "lITH EACH CLUSTER Group s 18 20 21 22 23 24 25 26 27 29 30 31 r/J 32 (l) r1 '='3 .c ..J c::i 4'~i 3 ?--: ('j:> 36 37 38 39 40 41 43 44 45 46 48 49 50 (1) A " '1~ 36 31 :18 39 40 7 6.24 3.57 2.18 2.42 5.38 4.13 5.09 3.83 3.87 2.38 1.49 3.55 1.59 4.39 4.05 6.49 6.53 6.49 6.51 6.40 2.94 2.94 2.91 3.13 3.67 3.67 2.36 3.79 (2) B 4-8 4-9 5D 3 1.40 1.31 1.04 .86 .00 1.35 1.90 .11 .66 1.55 1.44 1.43 1.94 1.10 .28 1.70 1.50 1.81 1.44 1.97 .73 .19 -.17 1.82 2.13 2.73 2.69 2.77 Lr­ (3) C 20 ZI 24 z.!i Zb 2.7 6 3.06 4.29 4.01 2.24 3.83 4.19 4.80 4.55 4.42 2.63 2.33 2.72 2.31 2.16 .37 3.69 3.22 3.27 3.07 2.83 .83 .87 1.06 2.52 2.21 2.52 1.60 2.25 (4 ) D 4/43 -r4 3 .82 .08 -.59 1.53 1.43 .43 .85 1.23 .76 .49 .32 .38 .66 1.47 1.81 .42 .94 1.75 1.58 1.85 2.28 2.56 2.50 -.06 1.44 -.26 .97 .04 (5) E 2~ ~o 3/ 32 4 1.08 1.35 .72 1.63 -.83 1.53 2.89 1.90 1.60 2.90 3.18 3.15 3.01 1.84 1.54 1.78 2.08 1.72 1.37 1.81 .74 .78 .33 1.88 1.60 2.05 2.25 2.06 (6) A 18 Z3 36 37 38 .39 40 IS .143 .89 .51 .31 .35 .77 .59 .73 .55 .55 .34 .21 ·51 .23 .63 .58 .93 .93 .93 ·93 .92 .42 .42 .42 .45 .52 .52 .34 .54 (7) B 48 43 so .333 .47 .44 .35 .29 .00 .45 .63 .04 .22 .52 .48 .48 .65 .37 .09 .57 .50 .60 .48 .66 .24 .06 -.06 .61 .71 .91 .90 .92 Mr (8) (9) (10) C D E 20 21 4/ 43 193D 242.5" 44- 3/3Z 2-b 2.7 .167 .333 .250 .51 .27 .27 .72 .03 .34 .67 -.20 .18 .37 .51 .41 .64 .48 -.21 .70 .14 .38 .80 .28 .72 .76 .41 .48 .74 .25 .40 .44 .16 .73 .39 .11 .80 .45 .13 .79 .39 .22 .75 .36 .49 .46 .06 .60 .39 .62 .14 .45 .54 .31 .52 .55 .58 .43 .51 .53 .34 .47 .62 .45 .14 .76 .19 .15 .85 .20 .18 .83 .08 .42 -.02 .47 .37 .48 .40 .42 -.09 .51 .27 .32 .56 .38 -.01 .52 .J 169 FIGURE 3. TRIAL CORRELATION PROFILE S OF VARIABLE S WITH GROUPS Tentative Clusters ABC D E I I I I I I-b 3~ -­ .8 ...... - -_ ·7 ' •••• -- 40 --­'~ ·b " .•••• : ...:-:-.:--- J'-.-. ---­ .S " • <;. __ ;''; z,;; __d:>_~ - - - 38 r·4 .,,'. ,/' -.~- -3 • '" . .,,' • ·-·7 .~- - ....-:- Cl. A .36 -2, ..".. •• • • .. ....... .. .. •••, ./ ".'/. ., 37 o ' ;' " ~ , /8 .... -.2 23 -.2 /.0 .., 20-­ ·8 ·7 2/ - - - ­ .~ ......... '" .... Z4 ....... r··If '"~: •..~ Cl. C :3 --- " Zs­ . ..... ~ 2.6 ·l .. -.- ••'.- " -'~ '/ < ...f...o -.I , 27 -.2 v .- -­ /.0 .' 43­ ...,2, I.D ., ·8 ... , , .." 30 -­ ..... ·7 3/.(;, - - ..... ~.;..: ,.....~ ..: .--:-:.~.,..:".... ·s . ~~::5 32Cl. E 29r·' .z. ·3 ./ 45o -/ 49··· .... ";2 48-­ 50 - --- Cl. B 49·· .. · . 46­ 4-5 ---­ /·0 .~ ·8 .7 ·6 .5 r :i -~------------ __"'=::::~~~""'~07::rE:.- _ .l, ./ o ~I 44----­ 4/ Cl. D 34­ 33 - --­ 22. ..•... ·8 .7 ·6 .5" Y·4 , ·3 .... _" " " .. _. ·2 - .......... _ / ....'0 > __ -.I 170 these items 'will constitute "residual variables rr for this cluster anal­ ysis. Third, the profiles for items 22, 33 and 34 seem to correspond closely to those of the basic variables in Tentative Cluster D. These items, therefore, will be retained in this cluster. Finally, items 45 and 49 appear to be related to both Tentative Clusters Band E. This suggests that these hTa clusters are related. The trial correlation profiles for these combined clusters, presented in Figure Lf below, verifies this observation. For the balance of this analysis, there­ fore, Tentative Clusters Band E will be considered as one cluster. FIGUFE 4. TEIAL COREELATION PROFILE S OF VARIABLE S IN TENTATIVE CLUSTERS B AI{D E Grou:ps ABC D E I 1 , t t 1.0 ., ·8 ·1 .f:, 30 3/ 32­ - -­ -­ ........ -5r.1 29 ...-.-.­ ·z 48 ./ o 50 ---­ -./ -·1 4-9 ... , .. -~ -.4 4S -'-'-' ", .~ •• ' ~ •.:...:.~~. __ .. ..,. _ ;j-~ -.- -..;.~~ ~,.~.: ..,.. .. The last step in the cluster analysis Hill be to plot the final correlation profiles of individual variables according to the clusters established in the previous operation. To facilitate this, the correlation coefficients for items'3:rr;~nged by clusters are laid out in Table 22, page 171 (cf. Tryon, page 5). At this stage, the clusters are named according to their congruence with the initially hypothesized ....I 171 TABLE 22. FINAL CORRELATION MATRIX: TljENTY-NINE CLUSTERED VARIABLES 1­ Clusters Urban Renewal Modern Social Structure .an..d Operation Government Modern :Social Principles Residual Variables 39 40 38 36 37 18 43 44 41 '34 22 33 25 27 26 20 21 24 30 31 32 i 29 46 50 49 45 23 46 47 39 .99 .98 .93 .93 .93 .54 .54 .50 ';62 .36 .65 .76 .56 .62 .44 .06 .63 .22 .51 .36 \.28 .55 .56 .33 .38 .76 .44 .20 40 .99 .97 .94 .95 .91 .63 .63 .59 .61 .65 .71 .87 .60 .31 .52 -.14 .67 .20 .57 .62 1.42 .66 .74 .57 .56 .65 .62 -.01 38 .98 ,97 .93 .93 ·91 .58 .57 .60 .78 .46 .68 .76 .60 .37 .67 .30 .57 .15 .68 .43 1. 46 .67 .67 .47 .55 ·79 .69 .06 36 .93 .94 .93 .98 .92 .16 .13 .13 40 .14 .54 .79 .57 .65 .50 .58 .60 .38 •• 54 .41 1. 45 .65 .64 .41 .50 .81 .62 .19 37 .93 .95 ,93 .98 .93 .34 " .31 .29 ,48 .24 .66 .80 .45 .56 .49 .49 .43 .46 .62 .50 1,,50 .53 .55 .42 .43 .83 .56 .07 18 .93 .91 .91 .92 .93 .24 .25 .33 ,46 .39 .62 .51 .33 .53 '.48 .60 .61 .15 .26 .27 1;40 .46 .51 .43 .34 .71 .43 -.17 43 .54 .63 .58 .16 .34 .24 .89 .78 69' .52 .48 .22 .21 .43 .07 -.19 .13 .1'7 .25 .22 i.14 -.15 • _.02 .36 -.05 .45 .45 .55 44 .54 .63 .57 .13 .31 .25 .89 .72 60 .52 .45 .32 .29 .51 .07 -.25 .12 .04 .06 .10 ;.13 -.24' -.17 .24 -.16 .48 .51 .50 41 .50 .59 .60 .13 .29 .33 .78 .72 ,52 .49 .54 ,31 .26 .29 -.06 -.15 .18 .11 .07 .34 .22 .13 .23 .37 .15 .50 .48 .49 34 .62 .61 .78 .40 .48 .46 .69 .60 .52 .66 .76 -.29 .04 .16 -.01 .31 .16 .32 .35 .50 i· 37 -.23 .26 .2'5 _.14 .73 .74 .42 22 .36 .65 .46 .14 .24 .39 .52 .52 .49 1'6 . .63 .59 .18 .70 .17 .24 .35 .43 .39 .39 \.42 .08 .23 .55 .28 .18 .48 .62 33 .65 .71 .68 .54 .66 .62 .48 .45 .54 ]76 .63 .40 .45 .62 .32 .31 .16 .44 .39 .50 i·51 .;14 .43 .33 .32 .53 .71 ,48 25 .76 .87 .76 .79 .80 .51 .22 .32 .31 -29 .59 .40 .87 .76 .80 .63 .87 .71 .TT .83 !.58 .62 .65 .63 .68 .60 .39 .56 27 .56 .60 .60 .57 .45 .33 .21 .29 ;26 '04 .18 .45 .87 .82 .78 .52 .56 .43 .41 .38 1-38 .32 ·32 .02 .31 .76 .41 .40 26 .62 .31 .37 .65 .56 .53 .43 .51 .29 ~6 .70 .62 .76 .82 .69 .75 .71 .38 .65 .44 1. 43 .08 -.02 .05 .46 .79 .41 .44 20 .44 .52 .67 .50 .49 .48 .07 .07 -.06 -pl .17 .32 .80 .78 .69 .70 .52 .38 .29 .25 .43 .47 .59 .25 .51 .47 .33 .38 21 .06 -.14 .30 .58 .49 .60 -.19 -.25 -.15 t~ .24 .21 .63 .52 .75 .70 .66 .09 .28 .01 .. 34 .45 .i7 .42 .12 .59 .11 -1.00 24 .63 .67 .57 .60 .43 .61 .13 .12 .18 .35 .16 .87 .56 .. 71 .52 .66 .34 .32' .40 \.47 .58 .54 .23 .44 .60 .55 -.01 30 .22 .20 .15 .38 .46 .15 .17 .04 .11 ~: .43 .44 .71 .43 .38 .38 .09 .34 .82 .TT i.77 .49 .39 .56 .40 -.07 ·35 .36 31 .51 .57 .68 .54 .62 .26 .25 '.06 .07 .39 .39 .77 .41 .65 .29 .28 .32 .82 .76 i·73 .49 .40 .54 .49 .37 .48 .24 32 .36 .62 .43 .41 .50 .27 ,22 .10 .34 l~ .39 .50 .83 .38 .44 .25 .01' .40 .TT .78 ,.63 .66 .70 .58 .54 1.00 .28 .48 29 .28 .42 .46 .45 .50 .40 .14 .13 .22 .42 .51 .58 .38 .43 .43 .34 .47 .77 .73 .63 .41 .57 .57 .45 -.13 .49 .25I 48 .55 ,,66 .67 .65 .53 .46 -.15 _.24 .13 .08 .34 .62 .32 .08 .47 .45 .58 .49 .49 .66 1. 41 .93 .87 .59 .15 .69 .17 50 .56 .74 .67 .64 .55 .51 _.02 -.17 .23 6 .23 .43 .65 .32 -.02 .59 .17 .54 .39 .40 .70 :.57 .93 .91 .65 .12 .77 .41 49 .33 .57 .47 .41 .42 .43 .36 .24 .37 5 '.55 .33 .63 .02 .05 .25 .42 .23 .56 .54 .58 ;.57 .87 ·91 .58 -.27 .67 .53 45 .38 .56 .55 .50 .43 .34 -.05 _.16 .15 4 '.28 .32 .68 .31 .46 .51 .12 .44 .40 .49 .54 1.45 .59 .65 .58 .37 .40 .77 23 .76 .65 .79 .81 .83 .71 .45 .48 .50 3 .18 .53 .60 .76 .79 .47 .59 .60 -.07 .37 -1.00 ,)13 .15 .12 -.27 .37 .31 .38 46 .44 .02 .69 .62 .56 .43 .45 .51 .48 4 .48 ,71 .39 .41 .41 .33 .11 .55 .35 .48 .28 \.49 .69 .77 .67 .40 .31 .50 47 .20 _.01 .06 .19 .07 -.17 .55 .50 .49 2 .62 .48 .56 .40 .44 .38 -1.00 -.01 .36 .24 .48 :.25 .17 .41 .53 .77 .38 .50 , /~--- 172 l clusters. The final correlation profil-es of clustered varia.bles are presented in Chapter IV, Figure 1, page 82 (cf. Tryon, page 7). The discussion of the groupings discerned by this cluster analysis, with particular reference to their relationship to the initially hypothes­ ized clusters, accompanies this figure in the text. The variables identified by the operational unities of "urban renevlal," "modern social structure and operation," "government" and "modern social prin­ ciples" i'rill constitute the Right-Wing Extremism scale developed at the end of Chapter IV. I I I I __1_