STATUS CONSISTENCY AND SECONDARY STRATIFTCATTON CHARACTERISTICS IN AN URBAN ' METROPOLIS ‘ T ' ’ Thesis for the Degree of Ph. D. MTCHIGAN STATE UNIVERSITY , RONALD BURTON 1972 =:.':I‘ 54.41: LP“ ‘RY Michigan State University lhthll This is to certify that the thesis entitled Status Consistency and Secondary Stratification Characteristics in an Urban MetrOpolis presented by Ronald Burton has been accepted towards fulfillment of the requirements for Ph.D Jeme in Sociology Mdmummuwt Date {fl 7/72; 0-7639 One theoretical "status cry Lenski Rese 5°”ing has to the Fess CUSSEd iatel ass“"‘Ptions tiOns of ti ”Search a; “he might Ca]] The seCQndary (1" I (2; ABSTRACT STATUS CONSISTENCY AND SECONDARY STRATIFICATION CHARACTERISTICS IN AN URBAN METROPOLIS By Ronald Burton One of the purposes of this dissertation has been theoretical, empirical and methodological analysis of "status crystallization," with special emphasis on the Lenski Research Program. Some of the Program's key rea- soning has been investigated to ascertain its relevance to the resolution of certain critical problems to be dis- cussed later. Because the Program's original "core“ assumptions are not supported by the "auxiliary" assump- tions of the available survey-research (Lakatos. 1968) there are certain built-in weaknesses in this whole research approach. When a person's ranks have low visibility we might call this "secondary stratification". The following set of assumptions relates to these secondary stratification ranks: (1) Secondary ranks are emphasized by P only when a highly visible rank is low. (2) A secondary rank is recognized by only a limited-membership-group. The is based, v probability Detroit, an The status- income and tions towar social part' and "other 1 tallization’ and ethnicit Ronald Burton (3) When a person steps outside that limited- membership group. the validation of secondary- rank claims is problematic. ‘ The empirical data upon which the current analysis is based, was gathered from five cross-sectional probability-sample surveys of the adult p0pulation in Detroit, and is essentially the same data that Lenski used. The status-attributes used were high and low states of income and education; the dependent variables were inclina- tions toward social change (voting Democratic) and lack of social participation (frequency of contacting "relatives" and "other friends.") We controlled for "structural crys- tallization" (Smith, l969) (above and below age forty-five) and ethnicity (Protestant and Catholic.) Lenski had predicted that persons with inconsisten- cies involving a high state of achieved status and a low state of ascribed status would vote Democratic more often than Republican. However, when we introduce controls for age and provide alternative formulations based on the sali- ence of ascribed versus achieved status, either the inter- action effects disappear or the alternative formulations explain the data almost as well as the main formulations. As was suggested in Chapter III on voting behavior, of the two formulations provided, the "status" (ascribed) alterna- tive operates, for an unexplained reason, to the exclusion of the "class" (achieved) formulation. Concerning social participation, Lenski predicted that persons whose status is poorly crystallized will i ebeit a tend wume. He ar bdavior decli hconsistency reflm of inter hstinction be Wosulation“ i wntributions isfics which ; fmends" and resPonse to l He fi PTOductivity “Search in t i“ a survey-a SUf‘TEStrict fings. In or hcfive and h ConditiOns sh (HA Ronald Burton exhibit a tendency to withdraw from, or avoid, social inter- course. He argued that negatively sanctioned or unrewarded behavior declines in frequency, and that status- inconsistency is associated with denial of rewards in the realm of interpersonal relations. One finds here neither distinction between the processes of "isolation" and "insulation" (Zelditch and Anderson, 1966) nor use of Blau's contributions (l964) on secondary-stratification character- istics which predict validation problems among "other friends" and not among visiting "relatives," based on the response to lack of education when paired with high income. We find that the Lenski formulation generally lacks productivity and persistence because representative research in this tradition has been carried on primarily, in a survey-analysis setting, and its replications are self-restricted, not extending to other standardized set- tings. In order to suggest that a research program be pro- ductive and have momentum, at least the following three conditions should be met: (l) Applicability to different data-gathering set- tings. (2) Openness to further research in elimination of alternative explanations of specific findings. (3) Response to specific challenges either method- ological or theoretical in nature- Nithin such a perspective, the value of Lenski's formulation does not seem apparent. It has not been applied toa variety 0 rntricted to nthout posing chfllenges of Lemki theory, theory (1957, building on ct its "core" the Ronald Burton to a variety of standard settings. Its use has been restricted to similar types of analyses of survey-research without posing alternative formulations or answering the challenges of alternative formulations. In this way, the Lenski theory, unlike the Festinger "cognitive dissonance" theory (1957, l964), has not been fecund in answering and .building on challenges while maintaining the vitality of its "core" theoretical areas. STA inp STATUS CONSISTENCY AND SECONDARY STRATIFICATION CHARACTERISTICS IN AN URBAN METROPOLIS By Ronald Burton A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Sociology T972 C) C0pyright by RONALD BURTON 1972 I w hers of my structive c 399919. Cha McKee and F Dr. thesis adv- VNUable t bEen comp] To immeaSUrat and gratn tions have tion 0f tl F “thou h Dossible. ACKNOWLEDGMENTS I wish to express sincere appreciation to the mem- bers of my guidance committee for their interest and con- structive criticisms. These members were Drs. J. Allen Beegle, Chairman, 80 Anderson, Research Director, James B. McKee and Frederick B. Waisanen. Dr. Beegle has been a most understanding and helpful thesis advisor and without the most generous offering of his valuable time and suggestions, this thesis could neVer have been completed. To Bo Anderson, for his intellectual stimulation and immeasurable help, I wish to express my utmost admiration and gratitude. His theoretical and methodological contribu- tions have provided the motivation necessary for the comple- tion of this thesis. A Finally, to my wife Mona, the acknowledgment that without her tender loving care, none of this would have been possible. ii CRAPTER I. THEOI His "CC Ass Met Asc Met Con 11. DATA The Met Smi: E Addi An t Conc IIL VOTING INC Intr SOci Stru Metr Alte Ahal CHAPTER I. II. III. TABLE OF CONTENTS PAGE THEORY . . . . . . . . . . . . . . . . . . . . . 1 Historical Background . . . . . . . . 1 "Core“ Assumptions in Theories of. Status Consistency . . . . . . 3 Assumptions About Behavior Consequences of Status Inconsistency . . . . . . . . . . ll Methodologies . . . . . . l9 Ascribed- Achieved Status Distinctions . . . . 3l Methodological Challenges to the "Lenski Tradition" Research . . . . . . . . 37 Conclusion . . . . . . . . . . . . . . . . . . 47 DATA . . . . . . . . . . . . . . . . . . . . . . 52 The Data to be Used . . . . . . . . . . . . 52 Methodological Challenges . . . . . . 55 Smith's "Structural Crystallization" as a Methodological Contribution . . . . . . . 64 Additional Blalock Methodological Challenges . . . . . . . . . . . . . . . . 67 An "Ideal Test“ of the "Lenski" Formulation . . . . . . . . . . . . . . . 69 Conclusion and Review . . . . . . . . . . . . 7l VOTING DEMOCRATIC AND STATUS INCONSISTENCY . . . . . . . . . . . . . . 74 Introduction . . . . . . . . 74 Social and Economic Formulations of Voting Behavior . . . . . . 76 Structural Crystallization and Voting Behavior . . . . . . . . . . . Bl Methodological Challenges . . . . . . . 86 Alternative "Achievement“ Assumption (Model 2) . . . . . . . . . . . . . . . . . 9l Analysis of Data . . . . . . . . . . . . . . 93 Review and Conclusion . . . . . . . . . . . .100 iii CHAPTER IV. SOCI IN TABLE OF CONTENTS - continued CHAPTER PAGE IV. SOCIAL PARTICIPATION AND STATUS INCONSISTENCY . . . . . . . . . . . . . . . . 103 Introduction . . . . . . . 103 Jackson' 5 "Isolation" Assumptions . . . . . . 105 Alternative "Insulating" Assumptions . . . . 110 Secondary Stratification Assumptions 117 Methodological Challenges . . . . . . . . . . 123 Analysis of Data . . . . . . . . . . . . . 124 Review of Data . . . . . . . . 135 Conclusion . . . . . . . . l V. CONCLUSION AND EVALUATION . . . . . . . . . . . 138 Introduction . . . . . 138 When Are Research Programs Productive?. . . . 141 Reasons Why the Lenski Formulation Has Not Been Fruitful . . . . . . . . . 149 Applications of Lenski' 5 Theory . . . . . . . 150 Alternative Extensions of the Lenski Formulation . . . . . . . . . . . . 154 Future Research Directions . . . . . . . . . 158 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . T70 iv TABLE EF EF EF E l TABLE LIST OF TABLES EFFECTS ON VOTING DEMOCRATIC OF DIFFERENT COMBINATIONS OF INCOME- EDUCATION INCONSISTENCY FOR PROTESTANTS AND CATHOLICS . . . . . . . . . . EFFECTS ON CONTACTING CATHOLIC RELATIVES OF DIFFERENT INCONSISTENCY COMBINATIONS . . .—. . . . . . . EFFECTS ON CONTACTING RELATIVES OF DIFFERENT COMBINATIONS OF INCOME- EDUCATION INCONSISTENCY FOR PROTESTANTS . . . . . . EFFECTS ON CONTACTING OTHER FRIENDS OF DIFFERENT COMBINATIONS OF INCOME- EDUCATION INCONSISTENCY FOR PROTESTANTS AND CATHOLICS PAGE 84 125 T30 T33 historical Q The t1 sional strati and Party," w Stratificatio hierarchies. consideration Pendent of th market eCODOn tion of SOCia hllDotheses cc NUitip1e aff“ 0n behavior . high COrreia. amultidimen 1953). C001 that: u posslble col p011“tical h CHAPTER I THEORY Historical Background The theoretical roots of the study of multidimen- sional stratification are found in Weber's "Class, Status and Party," which provides a conceptual model of social stratification consisting of several empirically distinct hierarchies. Two forms of stratification are given primary consideration, class and social status, power being inde- pendent of the two. Class is based on position in the market economy, and social status is based on the distinc- tion of social respect (Weber, 1953). Simmel also poses hypotheses concerning the relationship of conflicting multiple affiliation and the effects of status-inconsistency on behavior (Simmel, 1955:54). Marx's preposition of very high correlation between various rank hierarchies suggests a multidimensional view of social rank (Bendix and Lipset, 1953). Cooley, in his discussions of stratification wrote that: ". . . there are in fact an indefinite number of possible collocations" (Cooley, 1909:249). Sorokin thought of stratification as made up of occupational, economic and political hierarchies. He was one of the first theorists to suggest a t 'innerly antac mutually contvr 1947:289; citr The C' new of strat in an article sees the econ separate and frequently (:1 “may indiv hierarchies a by the fact The of statuses POlitiCa1 ’ blockageS tensiOnS a MS. Stat 2 to suggest a tension related to ill-matched statuses, an "innerly antagonistic . . . multi-bonded Stratum made up of mutually contradictory, uncongenial bond—values" (Sorokin, 1947:289; cited in Jackson, 1960). . The change in orientation from a unidimensional ‘view of stratification reached a theoreticalturning point in an article by Benoit-Smullyan in 1944. Benoit-Smullyan sees the economic, political and prestige dimensions as separate and distinct status-hierarchies, despite their frequently close correlation. There may develop in a society individuals whose status ranks on these several hierarchies are not parallel, a condition made explicable by the fact that as a man's status changes, affecting not only one dimension at a time, imbalances mayoccur. In a dynamic and mobile society, status equilib- rium is always being disturbed since pronounced changes in status occur ordinarily in only one type of status at first, and are only gradually converted into equivalent statuses in other hierarchies (Benoit-Smullyan, 1944:160). There is a pressure, then, toward "equilibration“ -of statuses through these "conversion" processes in the political, economic and social realms.. In addition, if blockages in this "equilibrating" process occur, consequent tensions and frustrations lead to engagement in a course of action designed to bring abbut congruency in the individu- als' statuses: Thus the customary and legal barriers of the 'ancien regime' prevented the prospering bour- geoisie from achieving a position in the politi- cal and prestige hierarchies which would be in line with their economic status. The economic Hhat E emeriencing ' iayattempt t nnhngs in 0 “Me" Assfl 311115 Incw This M N1 subse equilibriuny devElOP Nhe ivy" Proce The ”3301‘ c formuiatio' equilibrat Th formu]atic berry ref. sistency“ baiance-< 0f bh‘an .3 progress of the bourgeoisie would not in itself have produced the ensuing tensions, were it not for the fact that there exist.strong psy- chological and social forces working toward .status equilibrium which were continually blocked or repressed by institutional barriers (Benoit-Smullyan, 1944:160). - What Benoit-Smullyan suggests, then, is that persons experiencing tensions resulting from status "diSequilibrium" may attempt to reduce the tension level by altering their rankings in one or more of the dimensions of status. "Core" Assumptions in Theories of StatusTnconsistency This beginning by Benoit-Smullyan became the matrix of all subsequent studies of status equilibrium and dis- equilibrium: how equilibration originates, how tensions develop when the equilibrium is disrupted, and what "auxil- iary" processes are applied to re-establish equilibrium. The major differences that develop are concerned with the formulation of tension assumptions and the development of equilibrating mechanisms. . There are three major types of theory regarding the formulation of tension assumptions (Anderson, 1971). Kim- berly refers to these as the "psychology of status incon- sistency" (1966). The first, referred to as the “cognitive balance-dissonance variant set“ stems from Heider's theory of balance (1958) and Festinger's theory of dissonance (1957): Balance theory, at its most general level, focuses on the tendency for the affect charac- terizing the individuals' responses to parts of an entity to be of like sign. Oil 01“ ar me ti 1? The no that by Zeldi: | model to defir “19 the origir rium: 9,111,001?» Essentiany t following: ( rows tend t1 imbalanmd r Anders“. ‘9 The gm Variant (1961) and - portlonal tributive 4 Dissonance theory, for another example,.focuses on cognitions which imply the obverse of one another and which appear to result from commit- ment to behavior which results in the frustra- tion of an important motive (cited in Kimberly, 1966:218). . ' The most cogent of the studies in this tradition is that by Zelditch and Anderson (1966) which used a "balance" model to define the socio-psychological processes concern— ing the originations of tensions due to status disequilib- rium: A rank is defined as any value onany criterion. An individual's ranks are balanced if they are all above, all equal to, or all below those of other individuals in the same system. The stratification structure is balanced if the ranks of all individuals within it are bal- anced (Zelditch and Anderson, 1966:246-248). Essentially then, this "balance“ formulation assumes the following: (a) balanced ranks are stable; (b) imbalanced ranks tend to change until they become balanced; and (c) imbalanced ranks produce a state of tension (Zelditch and Anderson, 1966:249)L The next tradition, which has been called the "cost- gain variant set" (Anderson, 1971) derives from Homans (1961) and Thibaut and Kelley (1959). According to Homans, tensions derive from situations in which persons in an exchange relationship expect the profits of each to be pro- portional to his investments. Under such a rule of "dis- tributive justice" A man in an exchange relation with another will expect that the rewards of each man be propor- tional to his costs, the greater the rewards, the greater the costs, and that the net rewards, or profits, of each man be proportional to his in‘ 9'" In the investments ar exchanged and {cited in Blau Thibau tion of this t 11 1:1 11 mmi‘ir—tko The n who dEVEIOpm DOSEd mainly Huber” adv reward- COS t (Kimbefly Add! t“ t“Miti pSYCh0109y matlve d 5 investments; the greater the investments, the greater the profit (Homans, 1961:75). In the following passage, Homans adds that when investments are almost equal, the value of the rewards exchanged and the costs involved is expected to be equal (cited in Blau, 1971:57). Thibaut and Kelley (1959) present a slight varia- tion of this tension formulation: This is the concept of comparison level, which they define as the standard against which an individual evaluates the attractiveness of a given relationship. Attractiveness is a func- tion of how close the outcomes, that is, the rewards and costs, of the relationship are to what the individual feels he deserves (Thibaut and Kelley, 1959: cited in Kimberly, 1970:84- 85). The most important Spokesman for this "cost-gain outcome variant set" is Kimberly (1962, 1966, 1967, 1970) who developed a social psychology of status-inconsistency, based mainly on the work of Thibaut and Kelley (1959). Kimberly advances the argument that tensions are not a result of status-inconsistency per se, but rather of poor reward-cost outcomes often associated with inconsistency (Kimberly, 1967). Adams (1965), whose work seems to span the first two traditions mentioned, argues that, in tracing the social psychology of "just distribution" of rewards to the work of relative deprivation: First, the dissatisfaction associated with relative deprivation results from felt injus- tice, rather than.directly from relative deprivation. Second, what is Just is based upon strong expectations. Third, a comparative i The related to t "expectation two major fi Lenski (1954 Geschwender ing in this 0fthis diss Both S1'0" assumpt a150 similar Ponent as a NMle Lehski The °fhis IdEa dild 6 process is inherent in the development of expectations and the perception of injustice. Fourth, and finally, felt injustice is a response to a discrepancy between what is per- ceived to be and what is perceived should be (Adams, 1965: cited in Kimberly, 1970:84). The final of the three theoretical positions related to tension-formation may be referred to as the "expectation-conflict variant set" (Anderson, 1971). The two major figures in this are Sampson (1963, 1969) and Lenski (1954, 1956, 1964, 1966, 1967) with Hughes (1944) Geschwender (1967) and Gibbs and Martin (1964) contribut- ing in this tradition. The primary theoretical emphasis of this dissertation is based on this final tradition. A Both the Sampson and Lenski views begin with ten- sion assumptions derived from Benoit-Smullyan (1944). They also similarly indicate an “expectation incongruity“ com- ponent as a cause of tension. However, as we shall see, Sampson's expectation set is more "task positional" (1969) while Lenski's is more "social response" oriented (1967). The "core assumption set" of Sampson (1969) consists of his idea of "positional" expectation incongruity and congruity; his rank variables are elements such as sex, age and class-level in college (Anderson, 1971a). In addition, in group situations tasks are characterized by a division of labor, since complex tasks that groups must accomplish consist in reality of a number of separate sub-tasks that must be coordinated (Brandon, 1965; Bunker, 1966; Sampson and Bunker, 1966; Kardush, 1968; cited in Sampson, 1969). An additional assumption is that these sub-tasks are given high or low S how important Davis and M00 sizes functio A maj tation of "co (sex, age, c1 importance as reason the st does not acco ual involved, if there is a assigned and individua] se Expectancy-” leads to tens COngl‘uence 16 (SampSOn‘ 196 Turni uCOl‘e" Ideas deals with a l'lCCLlpaHOr'a.I min and e1 core aSSUm I Dt~ eated them n 7 high or low status ranks by the participants in terms of how important they are. This position is similar to the Davis and Moore (1966) theory of stratification which empha- sizes functional importance and concomitant evaluation. A major assumption then, is that there is an expec- tation of "correspondence" between the state of the ranks (sex, age, claSS-level, et cetera) and the prestige and importance assumed in the sub-tasks involved. If for some reason the skill-level of the sub-task of an individual does not accord with the state of the ranks of the individ- ual involved, "expectation incongruity" results. However, if there is a match between the level of the sub-tasks assigned and the State of the individual's rank, then the individual sees himself and is seen by others as congruent. Expectancy-incongruity between expected and realized tasks leads to tensions for the individual, while expectancy- congruence leads to satisfaction with thetask-assignments (Sampson, 1969; cited in Anderson, 1971a). Turning next to Lenski, one might spell out the "core" ideas in his formulation (Anderson, 1971b). Lenski deals with a group of status-variables including income, occupational prestige, education and a combined measure of racial and ethnic group standing. Interestingly, Lenski's core assumptions were only implicit until he himself delin- eated them recently: Individuals strive to maXimize their satisfac- tions, even, if necessary, at the expense of others. This means that an individual with inconsistent statuses or ranks has a natural Assu defls with c twosseem b hepays atter in relevant hat when an vwght to the mther than i U". The sec anindividual flaumes, or tong hIQDEr indWidual ex his higher S t anindlvidua] Stat“S-Variab awe-#01130 8 tendency to think of himself in terms of that status or rank which is highest, and to expect others to do the same. Meanwhile, others who come in contact with him have a vested interest in doing just the opposite, that is, in treat- ing him in terms of his lowest status or rank (Lenski, 1966:87). . Assuming that the state of the ranks that Lenski deals with can be seen as high or low, the following assump- tions seem to follow. When an individual evaluates himself, he pays attention only to the average of the high states of his relevant ranks. A variant of this assumption might be that when an individual evaluates himself, he gives greater weight to that set of status-variables in which he is high, rather than to the set of status-variables in which he is low. The second assumption is that when others evaluate an individual, they evaluate him on the basis of his lower statuses, or give greater weight to his lower rather than to his higher statuses. The third assumption is that the individual expects others to evaluate him on the basis of his higher statuses. Finally, any discrepancy between how an individual and others evaluate the condition of his status-variables leads to tensions in the individual: One can see how this works, and the conse- quences of it, by imagining the interaction of a Negro doctor and a white laborer in a situa- tion where neither the racial nor occupational status-system alone is relevant. The former, motivated by self-interest, will strive to establish the relationship on the basis of occupation (or perhaps education or wealth), while the latter, similarly motivated, will strive to establish the relationship on the basis of race. Since each regards his own point of view as right and proper, and since neither is likely to view the problem in a detached, analytical fashion, one or both are In a tions, We 596 self, tensior self-evaluat' the other har expects to be such differer ”expectation- Snullyan's (‘ Point as MT 1 l i I 1 1 i S : A Val is the 1 integration, Set“ Status‘lnteg fun” of Soci barber) and tions of the (1) The inch-I (StatUses) s 1 ions (“3185' h 9 likely to be frustrated, and probably angered by the experience'(Lenski, 1966:87). - . In a comparison of the Sampson and Lenski formula- tions, we see that Lenski emphasizes that in evaluation of self, tensions arise because of the difference between self-evaluation and evaluation made by others. Sampson, on the other hand, talks of the incongruity between where one expects to be and where one actually is in a group. Despite such differences, both aspects are variants of a basic "expectation-conflict variant set" deriving from Benoit- Smullyan's (1944) tension model. Anderson summarizes this point as follows: ' Sampson's and Lenski's formulations, of course, both have a tension assumption. Sampson's theory is more limited in scope, in that it pertains to task oriented groups with a divi- sion of labor. Lenski appears to be thinking of 'social' expectations regarding reward objects in a wide sense; acceptance, rebuffs, social standing; place in a division of labor and reward mediated by such positions in a. Etatus structure for Sampson (Anderson, 1971b: A variant of this "expectationconflict variant' set" is the Gibbs and Martin (1964) theory of "statUs integration.“ The basic terms involved are roles, statuses, status-integration and role conflict. Status is seen as a form of social identification (such as married man, or barber) and the roles are the rights, duties, and obliga- tions of the status. The basic assumptions are as follows: (1) The individual has a series of social identifications, (statuses) which have attached rights, duties and obliga-- tions (roles). (2) The individual is expected to perform the roles if cess in C0“ individual I other status conforming t forming to t confronted w Martin, 1964 This the variant 1 certain behav individuals' roles it make sinultaneousl Simil “role conflic With interact societies, 5 SituatiOnS ra [511191. 195: the "marginal Caught betwee A A A A f". W'A. 10 the roles in each and every status he occupies. (3) Suc- cess in conforming to the roles of the statuses of an individual is dependent upon the nature of the roles of other statuses that a person occupies. Finally, (4), if conforming to the roles of one status interferes with con- forming to the roles of another status, an individual is confronted with an incompatibility in statuses (Gibbs and Martin, 1964:23). This, then, is the tension model that is posed, and the variant is seen as similar to Sampson and Lenski: certain behavior is expected of the roles in each of the individuals' statuses, and if two statuses have conflicting roles it makes these statuses incompatible when occupied simultaneously. I Similar to these formulationsis Hughes' (1944) "role conflict" position which stems from Simmel's concern with interaction within small groups instead of whole societies. Simmel was concerned with the instabilities of situations rather than with their monolithic structures (Simmel, 1950). Following in this tradition, Hughes sees the "marginal man" as one at odds with his contemporaries. Caught between worlds: . . he is not amenable to easy categorizations; his self-image and the image others have of him are at best ambiguous. Hughes sees much of this hinging upon 'Dilemmas and Contradictions of Status. He uses examples like the Negro doctor and the female scientist to point out that role- conflicts ensure, and that these in turn, create peculiar patterns and strate— gies of interaction. . . . Hughes has r+(+D-l'1 Hughe occurring whe personal attr isa: -m-a.n a ll restimulated concern for status-discrepancy and directed it more toward the individual than toward the problems of social structure that occupied Weber (Demerath, 1965: 131). Hughes then speaks of "status contradictions" occurring when individuals enter a status without all the personal attributes usually expected of incumbents. (This is a: . . clear recognition of the ambiguous and conflicting social position of the status inconsistent, but his concept differs from Benoit-Smullyan in that for him the term 'status' refers to all social positions with definite rights and duties, not just those involving relationships of superiority and sub- ordination (Jackson, l962:2). The final author in this variant set is Geschwender (1967) who seems to combine some "expectation" assump- tions of Sampson and some "distributive justice" assump- tions of Homans (1961). On the one hand he argues that it is inconsistency of expectations (and the dissonance they produce) that makes rank inconsistency stressful; on the other hand he views certain ranks as investments and cer- tain other ranks as rewards. Stress is produced when there are inadequate profits or insufficient rewards for the investments involved (Geschwender, 1967). Assumptions About Behavior Consequences of Status Inconsistency Following the discussion of the tension-assumptions of the three major variant sets, we turn to a discussion of the "auxiliary set of assumptions“. These are the 12 "auxiliary set" commonly questioned in the area of status- consistency (Anderson, 1971b). In the "cagnitive balance-dissonance variant set", Zelditch and Anderson (1966) discuss a number of modes of equilibration which include types of mobility, redefini- tion of the system of stratification, and withdrawal. The latter is delineated as follows:, Isolation involves simply ceasing to com- pare one's ranks with those of others. Insula- tion involves shifting to a comparison with others among whom one's own inconsistent ranks are not significant ranks. Role differentia- tion involves restricting comparison to statuses (as_opposed to the more general com- parison of actors) which do not involve inconsistent ranks (Zelditch and Anderson, 1966; cited in Kimberly, 1970:93). Mobility, as a means of equilibration, differs between causally and non-causally related ranks, the dependent rank changing in the direction of the independ- ent rank in the first case and the lower rank being elevated in the second case. In addition, expectations of mobility affect the intensity 0f reaction to rank— inconsistency. Reactions are less negative if mobility is anticipated. The authors then distinguish between individual, stratum and status, or, collectivity, mobility. The final model of equilibration becomes a redefinition 6f the system of stratification: Alsc variant get ailemetic“ Brehm and ( “is of re tion 0f ra of “War these fac inCrEaSe 13 Two kinds_of redefinition are identified. A left-wing protest involves an attempt to decrease the importance of an old, estab- lished rank which is seen as blocking mobil- ity, and an attempt to increase the importance of a new rank with respect to which the equilibrating unit has been rising recently. A right-wing protest involves an attempt to increase the importance of an old, estab- lished rank of the equilibrating unit and an attempt-to decrease the importance of.a new rank which is seen as replacing it (Zelditch and Anderson, 1966; cited in Kimberly, 1970: 94). Also within the "cognitive balance-dissonance variant set," are Adams's theory of "inequity" (1965), an adaptation of disonance theory (Festinger, 1957), and Brehm and Cohen's discussion of a number of different ways of reducing rank inconsistency (1962). The reduc- tion of rank-inconsistency involves changing the object of comparison, withdrawing from the situation, distorting these factors cognitively, or having the individual increase or decrease inputs and outcomes (Adams, 1965: 283-296). Proceeding to the "cost-gain outcome variant set", Kimberly (1967) defines two modes of equilibrating skill and position-status. The first, based on position- mobility, suggests that if status-aSpiration is high, and the type of inconsistency is a low skill-conflict posi- tion: incon51 tMs same distribut 1.nJollStice exPErienc b€havior 1 other ham °b59rvab13 (”01113115, 1 Tu ant Set" 1'! 14 Mobility, which would be downward, will not occur. In lieu of this, it is predicted that an attempt to 'despecialize' the system will be made. This involves 'decreasing differ- ences' in the difficulty of positions. This mode of equilibration 'decreases' the 'aver- age' difficulty of difficult positions (Kimberly, 1967:89). If status-aspiration is high and the type of inconsistency is a high skill-easy position: Mobility, which would be upward will occur. If this is blocked, it is predicted that an attempt to 'specialize'the system will be made. This involves 'increasing differences' in the difficulty of positions. This mode of equilibration 'decreases' the 'average' diffi- culty of 'easy' positions. Consequently, it will increase the stress to which the high skill-easy position inconsistent is subjected (Kimberly, 1967:89). The equilibration process is handled by Homans, in this same tradition, in terms of the relationship between distributive justice and satisfaction. If a state of injustice exists and it is to a man's disadvantage, he experiences deprivation, and will "display the emotional behavior we call anger" (Homans, 1961:75). If, on the other hand, "distributive justice" is not realized and is observably to a man's advantage, he will feel guilty (Homans, 1961). Turning finally to the "expectation-conflict vari- ant set" we look first at the Sampson (1963; 1969) variant of the princip 1953) 0 WOW“; 1957) 15 of the modes of equilibration. Following are Some of his principles of expectancy congruence: Individual tendencies towards expectancy- congruence give rise to social communicated, interpersonal pressures for one's self and others to present pictures of themselves which are congruent with what is expected. ‘ Deviation of a group member from the group's Opinion on an issue will produce pressures toward uniformity of opinion, only when this deviation is also a deviation from expecta- tion for that particular group member. To the extent that P presents multiple faces to these others (because he is.subject to multiple expectations), they too will be in an undesirable situation and will seek to change it towards a picture of P which is congruent. Therefore, P will be under pressures from others to present a picture of himself which is consistent (Sampson, 1963:157-159). Sampson then gives a specific illustration (Adams, 1953) of expectation-incongruity as it might affect the productivity of a group. Given the inconsistency in expectations for the other's behavior, this reduction (in pro— ductivity) may stem from each individual's inability easily to coordinate his behavior with the behavior of others of his group. Such a group may expect much of its effort on achieving a congruence or a consistency of expectations, or working to defend against the apparent inconsistencies which eXist, and thus have little energy remaining to devote to the task (Sampson, 1963:161). We now turn to Lenski (1954; 1956; 1964; 1966; 1967) whose formulation will be explored in considerable detail the f01 is with others. of an a he can ations i mitant 1 are cons economic ago he i change 0 50”. 197 Should b 13 added Obtain 0t (1966) 1; assumptjc equlibra to his Su and Socia I Gibbs and 16 detail. Lenski‘s fauxiliary set of assumptions" Suggest thefollowing. One response to tension for an individual is withdrawal from interaction with, and evaluation by, others. A second is that if an individual has a low state of an ascribed rank as part of his status configuration, he can promote changes designed to alter the social evalu- ations of the low states of his ascribed rank. Conco- mitant is the assumption that a person whose statusesets are consistent tends to support the political and economic status quo. If he is satisfied with the status gugghe is likely to vote Republican and, if he is social- change oriented, he is likely to vote Democratic (Ander- son, 197lb:l). I In the interplay of research and theory, one should be reminded that an "auxiliary set of assumptions" is added to the "core set of assumptions" in order_to obtain observable consequences (Lakatos, 1968). Lenski (1966) links his "core assumptions" with his "auxiliary assumptions" by relating the equilibration orlack of equilibration of tensions based on statusoinconsistency to his surveyfstatus indicators of political liberalism and social withdrawal (Lenski, l966). In this same “expectation-conflict variant set" Gibbs and Martin (l964) pose this hypothesis of tension resoluti vary ini that p0: tion is a series EXPECt 17. resolution: western societies exhibit suicide rates that vary inversely with the degree of status-integration in that population (Gibbs and Martin, 1964). This explana— tion is derived, through a lengthy reasoning process, from a series of postulates: (a) the suicide rate of a population varies inversely with the stability and durability of social relations in the population, which (b) varies directly with the extent to which the Y individuals in that population conform to the patterned and socially sanctioned demands and expectations placed upon them by others, which (c) varies inversely with the extent to which individuals in that population are confronted with role conflicts, which (d) varies directly with the extent to which individuals occupy incompatible statuses in that population, which (e) varies inversely with the degree of status integration in that population (Gibbs and Martin, 1964:27). ' The final writer to be discussed in the Fexpectation-conflict variant set" is Geschwender (1967) who postL simple ti of rank now be are re: are: are l8. who postulates that equilibration attempts move from simple to complex modes: He defines individual mobility as the easiest. Next in difficulty is striking out against individuals and categories of individuals; that is, prejudice and discrimination. The most complex mode is an attempt to alter' society. Withdrawal or social isolatiOn and suicide are also seen as ways of coping with rank inconsistency (Geschwender, 1967: cited in Kimberly, 1970:95). A complex set of possibilities of relating types of rank inconsistency to these modes of equilibration will now be developed. Persons with two types of inconsistency are referred to as under-rewarded inconsistents. These are: high ethnicity and low occupation and income, and high education and low occupation and income. The psychological reaction to these types of inconsistency is anger. Persons with the former type of inconsistency can be upwardly mobile on their lower ranks if their education is sufficient. If it is not, prejudice and discrimination are likely. If this fails, they are likely to join a racist social movement. Persons with the latter type of inconsistency can be mobile on their lower ranks if age permits. If it does not, prejudice and discrimination are likely. If this fails, they are likely to join an extrem- ist social movement (Geschwender, 1967:163- 168; cited in Kimberly, 1970:95). Persons with two additional types of inconsistency are referred to as over-rewarded inconsistents: These are low ethnicity and high occupation and income and low education and high occu- pation and income. The psychological tions: °l°9y u; A” emph. flexDECti nariiy t 5Uryey c eXperime balaanll ‘19 reaction to these forms of inconsistency is guilt. Persons with either of these types of inconsistency are unlikely to be mobile in their lower ranks. Ethnicity and education are difficult to change. Persons with both types of inconSistency may develop prejudice. If this failS, they may.advocate moderate lib- eral reforms within the society (Geschwender, l967:l634168; cited in Kimberly, 1920:96). To summarize this section, all of the formula- tions: (a) seem to follow Benoit-Smullyan by using a core assumption involving an "inconsistency- tension" assumption and a "tendency to change" assumption. (b) the formulations differ with respect to how tension is produced and in terms of the "auxiliary set" of assumptions about the behavioral consequences. Methodologies We now turn our attention to the specific method- ology used to test the "auxiliary set of assumptions." An emphasis will be placed on the Lenski tradition in the iexpectation-conflict variant set." Thisis done pri- marily because my own data, similar to Lenski's, are survey data, while the other programs are closely tied to experimental paradigms. In addition, "dissonance- balance" and the "costegain outcome“ formulations of status-c justice. at al., that th formula tions, the ter “dissor (Cited types 1 and et Duted the sq indica 08rsgr stfiYGc C0n51| Signi‘ tallj of Cr Prefe for c tani 20 status-consistency end up dealing with “distributive justice," a field that poses problems of its own (Berger, et al., 1968). One might clarify the issue by proposing that the term istatus conSistencyi be reserved for the formulations involving fexpectation conflictf formula- tions, e.g., Sampson (1969) and Lenski (1966); and that the term "distributive justice" be used for the "dissonance-balance" and "cost-gain outcome" formulations (cited in Anderson, 1972). Lenski originally (1954; 1956) examined four types of status indicators (income, occupation, education and ethnicity) as configurative elements. He then com- puted an index of crystallization (based on the sum of the squared deviations from the mean) across these status- indicators for a sample drawn from the Detroit area. A person whose percentile score on these different ranks stayed fairly close to the mean of all the ranks was considered "highly crystallized,“ while one who deviated significantly from the means was considered “low crys- tallized.9 Lenski then related the score on this index of crystallization to political behavior and attitudes. Preference for the Democratic Party indicated preference for change occasioned by the tensions in a "low crys- tallized" configuration. In another article based on the same Detroit that a "low alack of pa activities a The: degree, rat related to “non-direct tinguish be ranks in 3! interest w 121011 in a An concerning iSm ls fOL measures as Status in whl'ch fOr Chang ents were 21 same Detroit-area data, Lenski (1956) further suggested that a "low crystallized" configuration would result in a lack of participation in voluntary and other social activities and associations (Lenski, 1956). These original stUdies by Lenski asserted that the degree, rather than the direction, of inconsistency was related to political liberalism and withdrawal. This "non-directed" type of inconsistency-index did not dis- tinguish between different combinations of high and low ranks in an "uncrystallized set" of ranks. Primary interest was in the overall level or lack of crystalliza- tion in a particular configuration. An early corroboration of Lenski's hypothesis concerning status-crystallization and political liberal- ism is found in a study by Goffman (1956) who used measures of education, income, and occupational prestige as status-indicators. Goffman was interested in the ways in,which status-crystallization is related to preferences for change in power-distribution within society. Respond- ents were asked to ’check their perceptions of the amount of influence in the conduct of national.affairs presently held by, and the amount of influ- ence they prefer to be held by, state govern- ments, big business, labor unions, businesses that were not big, and the national govern- ment (Goffman, 1957; cited in Geschwender, 1967:502). Indiv showed a grea with a high c' ports the noi can lead to power within Anot results whi< cerning p01 In a commen tained that measure of employed. beCause at the“! tal ferent 1nd 22 Individuals with low-status crystallization showed a_greater preference for change than individuals with a high degree of crystallization. This study sup- ports the notion that a lack of status-crystallization can lead to a desire for change in the distribution of powerwithin a society. Another study, by Kenkel (1956), produced results which did not support Lenski's hypothesis con- cerning political liberalism and status-crystallization. In a comment by Lenski (1961), however, it was main- tained that Kenkel had used different indices, since a measure of ethniCity was not among the status-attributes employed. Lenski's rejoinder to Kenkel is inadequate, because at that point in its development, the Lenski theory talked about ranks but did not specify that dif- ferent indices should make a difference. ‘Jackson, in a crucial study in 1962 suggested that ethnicity be considered an aspect of ascribed status and that equilibration processes were more likely to occur where combinations of ascribed and achieved-status attributes were considered in measuring status- inconsistency. This study moved the "non-directed" study of status-crystallization to the study of more "directed" combinations of status-attributes. Em; with a stuc statistical study, in t expressing cratic, WEI Study of pc Kelly and . Crystalliz include a achieved-s Ir Stance in nations; reHgious The resu] ancieS b‘ 23 Emphasis on this more-directed approach continued with a study by Treiman (1966) which showed that the only statistically significant results from Lenski's 1954 study, in terms of the effects of inconsistency in expressing liberal economic attitudes and voting Demo- cratic, were those that included ethnicity. In another study of political extremism published in the same year, Kelly and Chambliss failed to substantiate Lenski's crystallization-formulation, but they too failed to include a measure of ethnicity and dealt only in achieved-status attributes. In 1967 Lenski moved from his non-directive stance in a study of the voting-patterns of four western nations; he included at least one ascribed status, religious behavior, with his achieved-status variables. The results bore out the contention that sharp discrep- ancies between achieved and ascribed status-attributes seem to lead to equilibrating processes more readily than status-discrepancy between two achieved-status attri- butes alone. Further substantiation of the plausibility of this ascribed-achieved status-inconsistency model is found in later studies by Segal and Knoke (1968), Segal (l969), Segal and Knoke (1969) and Smith (1969). Olsen and Tully (1? bus: "At 1 economic-ett important t' by itself“ Hhi ary set of status-ing extensiveI ing mecha. tion, Su (2) withd Symptoms chinge t1 examinEd equ'libr repreSen tradltic the Stai EXPECte' negativ| “positi 24 and Tully (1970) summarize this literature as fol- lows: “At least in regard to voting preferences, socio- economic-ethnic inconsistency is apparently much more important than is socio-economic status inconsistency by itself“ (Olsen and Tully, l970:2). While the literature substantiating the "auxili- ary set of assumptions" in regard to the relations of status-inconsistency and political liberalism is fairly extensive, other response-processes seen as equilibrat- ing mechanisms have not received equivalent substantia- tion. Such areas as: (l) enhanced-mobility striving, (2) withdrawal and social isolation, (3) psychosomatic symptoms of stress, (4) preference for and attempts to change the social order, and (5) prejudice, need to be examined for status-inconsistency effects of these equilibrating processes. In an attempt to answer these questions, a representative sample of the studies done in the Lenski tradition has been compiled, including a statement about the status-attributes used, the response-processes expected, as well as the results, either positive or negative. A general review of these studies indicates that ”positive" results are obtained. That is, equilibrating I UQFL—Hthfl‘a mu—ammm ucmucoume ucmumuqmnz~ ZO~H~Q_p_moa m>wummmc m>wummmc m>wpmmw= m>_gwmoa mupamwm mmwgumqoguxma mwcmcgaoncUm huwmowmwpmc mxm_-uumm ox?~-:uczco co_>m;mn m:_po> :owpummmwpwmmwv xuw::EEou cowuomcmacw _mwuom zcouUMmeummcs mucmELomcmn zucu mmpnmwcm> pcmucmama cowumazuuo compmuaum comumqauuo meoucw cowumuanm msoucw cowpmaauuo cowumuaum meoucw co_pmu=um :o_pma:uuo xucmacmcou .pomcmgu ucaocmxuma mmpnmwgm> ucmucmamuca onhHowcwcm ocmngmm mac__w;a awupabaa cwgcmz .2 .Emcmcao Emwpcoummuoca cobwcme< cw mmmpu —m_uom mom— HHH .a mez .gumgmemo m .mm mm< osm— mmcoc cmummucmg vcmcomg .Eoogm F .Ne .mmucou .uom mo¢~ .m _me .cmssmm A.uoov Nm .mm< mmm— ugmzum .mEmu< cocb=< m>wummmc .mwflmmwm mmnzuwuum _eucu+~oa mmammqudfi UCQUCvamQ mEOUCm cowumqauuo :OmHQUDUU mmhnmfiLm> «mmmcmx we .w nesea .: .mmeQV mwm~ .szu*m Loan:< 26 m>wuwmoa m>_u_moa m>_u_moa m>wummmc m>wuwmoa m>wuwmoq m>wummmc mupamwm muo_c cp cowumgPUwpcaa pmmcc: Fm:u_>wccw cowumpomw Pomuom mpcmsm>oe meuom aw cowumawuwucma games: .>mucw uwprmcmcmm _m3mcvsuwz mm_uwcocws vcmzog xuWPwumo; cowummmcmmmmu vcmzop mu=p_bpe mum—Fm: so» “Leanna _mu_b__oa mou:b_upa pauwp_Foa «one cowuuasuuo meoucw :o_umu=um »p_ow=;ua cowumazuoo meoucr cowumusum mEoucw cowgmuznm cowumaauuo apwuwcgum auwuwcgba cowumqauuo weoucw co_pmo=um wEoucp co_uma=uuo covumuaum meoocw cowumqauuo cowamuacm mm_amwgm> acmucmamo manmwcm> pcmncmamucm .umms szccm .<.m.< momp copmcwgcm: suwnaa cwmcwm .o cwEmncmm .< mmEmO .LmvcmZSUmww Amcaav e .mv mouse; meuom mmmp .< magma .Lmvcmzsumow .:.m.z .mmwo Nomp .< mused .gmucmzsummw A.bammv p .ke mmucom vauom momp summed .m .cm22mm Amcznv e .me mmugom meuom sump .m uw>mo .cm~umu Ammmcmx mo .3 .mmwav momp .m v_>mo .cmpru Losu=< 0>vuvmoa mMAflMMm mmcmcu pmu_u__oa Low mococmemcu HCQUCQQQQ Qcm mEoucm mmfiamhxm> HEwUCQOUCN cmkga cm mm N. .tmEkkou Locu=< 27 m>wuwmoa m>_pmumc m>wpmmmc m>wu_moa m>wummmc m>mpmmoa mupammm newscwmasw PmuwmoFowmxganogoxmn cowuumtmwumm new acmELquEw Pmmeo_o;uxmq xpcwmpcmu quwp__oa mpmum mmmcmmmFLmzoa m_cmemmm cowumcwamm quFwnoe ungQ: cowuumwm_pmm now mucosa Pmuwuwpoa go» mucocmmmcq mmpnmwcm> “cavemama »»_ow=;»w cowumnzuuo cowumuaum cowuaaauuo «souc_ mEoucw cowumoavw co_umq:uuo msoocw compmuaum wEoucp cowuauaum xuwcowcmm Ana cowpmoaum man man cowumaauuo mEoucw mmpnmwcm> pcmncmamuca A.m= Loc93< mm am a mm—nmem> ~ w team ca mfipawwm HCQUCDQDQ ucmb U 28 m>wuwmon m>wuwmon m>wpmmmc m>pumamc m>wpmmmc m>wummmc mppammm Fnzmeuguwz cowumawupucma Papoom mc_uo> _mcmnw_ cowpwawuwucwa new mmu=p_pbm Pmowbw_oa awsmcmmwucma _mo_pw_oa mmpnmwcm> uwsocoum -oowprog Emw_memn?, qupbw_oa mmpangm> acmucmamo apw0wcgpm coppmaauuo oeoucw cowumuaum aumuwcgum co_uma:ouo meoucw cowumuaum msowmwpmc -oczum cowpmuzum mmu_ucw uwxcms umeocoum mmwummca chF—wzu mzpm> mcwppmzu :owumqauuo :owumuaum cowumazuuo msoocw comumuscm mmpnmwcm> acoucmawucH A.m=un .o ugmzcm .ccmsamA AngE=mv m .N .mzuou .Fowuom mmmp um>mo .mwxocx Aacsav m ._N .m.m.< mmm_ .a se__FP= ._ax=m¥ Amczav m .Pm ..¢.m.< oom— aa_Fasmgu .e .3 avenue .x .sppax aogba< w>vuwmon muramwm mucmcmmmca mcwuo> .ucmao teenage om~c2q Aufiuficcum vomN :ofiumqauoo bxmcgmm .txmzmq cocuaq mw—erLm> HCDUCQQwQ mm~nmmLm> “cuntwqmfi:~ r“ 29 m>wpwmoa m>wummmc m>_pwmoa m>wpwmoq m>wummmc m>wu_moa m>vamon mupammm Ememcuxm mew: usmwc mucmcm_esw ucmzou zuwpwpmo; mmanwcw> _mwuom cmcpo ucm sp.em; _m0wbwpoa mmcmsu quwuw_oa to» mocmcmeoca mmmcum to msopasxm mucocmmmca mcwpo> wucmcmmmca mcwpo> cowumasuuo msoucw cowumuzum cowpmaauuo mEoucw co_»mu=vm xu_u_cnuo cowumqsuuo msoucw auwuwcgbm cowpmaauuo msoucw cowumuaum cowumazuuo cowpmozum :owmwpmg cowumaauuo xuwuwcsum comumaauuo mm_nmwcm> unoccmawa mmpnmwcm> ucmvcmamvcfi Acommco mo .av .mmwo mmmp .m Acme .smza Acmaemuammv m .m Amopowuom momp »m_mem .m .u .o cmp_m3 .cmewucsm A.=.o .amu ebaozv .mm_a mmm_ .m u_e=oo .guo_a m:_uams some .<.m.< camp »_P=» .0 snag .m :_>Lmz .cmmpo .gwm .uom a cbpam: co .a mom, ammo: .z a__m=a .4 caeguwm .m_wmz AF_LQ Lo : uEQUEQQQUCH Cu 1 Low>mcwn w>vurmoa mo_cowcmfl uthCQQwD murnmwm 30 m>_ummmc m>wuwmoa m>_ummm= m>wpwmoa m>wuwmoa mu_:mmm mwocmmz ucmzou mnzuwpum mucmcmwmcg mcwuo> Acowumcnw_w=amv =o_pmuecwbcauw mmmpu FmWUOm mcwpo> u_amcuosmo gwcmcmmwugmq co: Low>mgwn PMUwuwpoa mmFamwcm> pcmucmamo meoucw Ammzoamv cowumuzcm cowpm~w_paumxco chsuuacam .cm> umochmm .em> um>mwcum cowamasouo meouzw comumuaum xuwuwcgum xbw_waoe _m:o_uma:uuo meoucw cowumusnm cowumasuuo mEoucw cowumuavm mopgmwcm> ucmucoamvcH smzv Pk ..m.a.< comp .w .o .cmswmgp A.umov o .em ..m.m.< mom, .m mmsogp .;b_sm Acmnewoammv m .mm Acpmsowuom exocx .a Femom .3 .z .m uwsao ._mmwm Agwnsmumov me mmucou Fowuom mmmF oxocx vw>mo .m upsea .Pamam Amczcv m .em ..m.m.< mmop .m ew>ao ._amam aocba< processes 5 attributes atleast m hold, for tion to tn Ascribed-F W1 observed in a conf fest then Pattern u fu] d Su99e 31 processes seem to be present when achieved-status attributes are found discrepant in a configuration with at least one ascribed status. This pattern seems to hold, for several other equilibrating processes in addi- tion to that of political preference. Ascribed-Achieved Status Distinctions What explanations have been supportive of this observed need for at least one ascribed—status attribute in a configuration before equilibration-processes mani- fest themselves? Olsen and Tully (1970) trace the pattern (at least in political behavior) to the process first suggested by Goffman (1957): status inconsistency can be expected to result in preferences for political change only when the individual is prevented from reducing his inconsistency through upward mobility. Since achieved statuses are pre- sumably seen by most people as more or less changeable if desired, achieved status incon- sistency should be associated with individual social mobility rather than desires for political change. Only inconsistencies involving one or more low ascribed statuses that are essentially unchangeable within the existing socio-political system should pro- duce pressures to alter that system (Olsen and Tully, l970:3). In terms of analyzing the general notion of ascribed status in the sociological literature, some use- ful distinctions have been made by Smith (1969), who suggests that one can examine ascribed-status ranks for fixed or c by others' T able stat 0f ascril Processe tio“ (Sm aS frequ lack of Sega] ( his Stu SiStent critic and wh‘ He add butes “Quid "hese WhoSe 32 fixed or changeable character since they are determined by others' evaluations of one's personal qualities: . they are dependent on attributes (qualities) which are either unchanging (sex, pigmentation, etc.) or while vari- able undergo changes over which ego has no control (e.g., age). They are there- fore fixed by virtue of being permanent qualities and/or because they are, from ego's point of view, inelastic (Smith, 1969:909). The distinction, then, between fixed and change— able status-attributes, has been the basis of discussion of ascribed statuses and their effects on equilibration processes, especially in terms of political participa- tion (Smith, 1969). However, an additional factor, not as frequently cited, is the factor of visibility, or lack of visibility, of a particular status-attribute. Segal (1968) suggests the importance of this factor in his study of the voting preferences of status incon- sistents, stating that an individual will vote Demo— cratic when others define his status-attributes as low and when he views these same status attributes as high. He adds the qualification that these low-status attri- butes must be visible in some fashion. Accordingly, one would expect more equilibration effects among Blacks, whose ascribed state is visible, than among Catholics, whose ascribed state is socially less visible. Smith (i969). SU‘ ", . . inc generally of stress relativel; (Smith, 1 W different they affi Tully (l relate v one neec ever, 3. Where t such as Seem tc The in Whose ldent: Prob]. Pri. rank 33 (1969), summarizing the Segal study, suggests that ". . . inconsistencies between low visibility statuses generally lead to intra psychic (covert) experiences of stress which are resolved in a compatible (i.e., relatively non-visible) mode, namely non-partisanship" (Smith, 1969:919). What we seem to have then is a combination of different states of changeability and visibility as they affect ascribed-status characteristics. Olsen and Tully (1970), among others, have suggested that to relate voting behavior to status-equilibration effects, one needs a non-changeable rank which is visible. How- ever, Segal (1968) and Smith (1969) have suggested that where the non-changeable rank was of low visibility, such as Catholicism, the pressures on the individual seem to shift from the “public" to the "private" realm. The individual who is "cross-pressured" politically, but whose low-ascribed status is not readily publicly identifiable, may still suffer from "intra-punitive" problems which then force him to withdraw from the political scene and not vote at all (Segal, 1968). If one considers the less-visible ranks as fprivate", one may wonder about the forms of private ranks that are changeable. There is no formal 34 discussion of this kind of social-ranking in the liter- ature. However, the visible, although changeable status attribute may be related to the status-equilibration effects of status inconsistency. Status-ranks which are low in visibility and changeable may be considered as a "secondary" form of stratification. According to Blau, "Once major strati- fication structures have developed, secondary ones in which low status persons rank higher than high status persons may be developed in an effort to promote cohesion within the system“ (Blau, 1964:50). Accordingly, when an individual's "major", highly visible ranks are low, he may emphasize other "secondary," less visible ranks through which he can gain a measure of esteem. For example, the wife of a professional academic may have only a grade-school education. Although her educational accomplishments are seen as low in status-rank, in secondary-status areas such as cooking, housekeeping, voluntary ser- vice, she may be highly esteemed by others and may gain a great deal of satisfaction from this recogni- tion. However, the high state of her secondary rank is recognized by a limited membership group, in contrast to the recognition afforded the Ph.D. whose rank is publicll “Doctor“ "public' others. obvious tively p tion pro may with explain The indi hi9h sta 0f view the indi fication Validati Problema between wife Wit home, bu iiVing r 35 publicly acknowledged by such devises as the title of "Doctor“ in formal presentations. As distinguished from "public“ ranks which are immediately discernable to all others, "private” status-ranks are not immediately obvious and are discernable only to small groups. In these terms, then, Lenski (1956) very nega- tively poses the issue of withdrawal as an equilibra- tion process, when he says that the status-inconsistent may withdraw, or retreat into isolation. He does not explain either the'process or the intent of withdrawal. The individual may very well withdraw to activities and high states of status-ranks which from his own point of view constitute highly rewarding activities. When the individual who claims eminence in secondary strati- fication steps outside of the limited-membership group, validation of these secondary-status rankings is problematic. There is then a possibility of tension between the publicly and privately-arrived rank. The wife with little education may be a gourmet cook at home, but she may feel extremely inadequate in the living rooms of her husband's doctoral friends. The following set of assumptions, then, is relevant to these secondary-stratification ranks: L as an ess isolation defense w ranks. [ tion acti a9ree w11 for supe‘ in EXCha mOre "1‘ 0n "par ValuQS be that they a! reward: 36 (a) Secondary ranks are only emphasized by P when a highly visible rank is low. (b) Secondary rank is recognized only by a limited membership group; validation out- side of that group is problematic. Lenski's withdrawal-response to tension (1956), as an essentially negative phenomenon and a form of isolation, may be seen as a retreat to a second line of defense where the individual excels in low-visibility ranks. Despite the validation problems of stratifica- tion activities restricted to specific groups, we agree with Blau that "the members who cease to compete for superior status win social acceptance in the group in exchange for the contribution they thereby make to group solidarity" (Blau, 1964:50). Finally, in terms of an analysis of ascribed statuses, these non-visible secondary ranks should be more "local" (Merton, 1957) than "cosmopolitan", based on "particularistic" rather than "universalistic" values (Kimberly, 1970). The distinction here seems to be that universalistic values are less problematic in terms of social interaction because in many systems they are rewarded, while particularistic values are rewarded in fewer, smaller, systems. The sources of status-'1‘ar then be be less visit different forthcomin mm 'lenski Tr Tu dition, we the Lenski the "auxil assumption have not d "auxiliary tions to t CUSSion. tions" H” An (a) rank i 901 itic“ “Wm 11'. POlitic“ underSt00d 37 status-ranking for locals in specific communities should then be based on particularistic values and on private, less visible stratification levels not transferable to different communities or publics (iucker and Kimberly, forthcoming; cited in Kimberly, 1970). Methodological Challenges to the "Lenski Tradition" Research Turning next to challenges of the Lenski tra- dition, we see that most of the inferred questions of the Lenski position actually relate to the relevance of the "auxiliary set of assumptions“ to the "core set of assumptions" (Anderson, 1971). The "coreassumptions" have not drawn the extensive criticisms directed to the "auxiliary assumptions," however three specific excep- tions to them will be treated at the end of this dis- cussion. The major challenges to the "auxiliary assump- tions" will be treated in order of their appearance. Anderson and Zelditch (1964) suggest: ". (a) rank inconsistency does not invariably produce a political response and (b) that if a political response occurs it can be rightist or leftist or take the form of political apathy depending on certain not very well understood conditions" (Anderson and Zelditch, 1964:120). Rush 1 might additi ual Nil effects Zelditc tive rai only at individu another. in ambiV' rich Ital high inco accomplis his ethnic equal amon buship in Am C(Ude that “CHons f 0f the equ Choose, 38 This position is borne out by the findings of Rush (1965), suggesting that in any given study, tables might contain bgth left and right-wing responses. In addition, there is some question of the rank an individ- ual will emphasize or pay attention to in terms of the effects of inconsistent self-image. Anderson and Zelditch suggest that even if we assume that the alterna- tive ranks are mutually exclusive, this effect occurs only at one particular time, and that subsequently the individual having emphasized one rank, may switch to another. The suggestion then is that this might result in ambivilence, based on some process of "rebuff". A rich Italian who at one point in time emphasizes his high income to his fellow immigrants and is proud of his accomplishments, at another point in time may denigrate his ethnicity. He wonders why he is not accepted as an equal among rich non-Italians, when he applies for mem- bership in an exclusive country club. Anderson and Zelditch (1964) as a result con- clude that it is difficult to make any determinant pre- dictions from Lenski's formulation, or to determine which of the equilibration-processes a status-inconsistent will choose. point of status me "by itse' or the ac bear on 1 be surpr‘ between i rium" (RL warrant i "theory c from whic risks are Era] rule StétuS~cO pa“tive Step c105 39 Runciman (1967; 1970), from a "qualitative" point of view, argues that discrepancy between the status man enjoys in one field and another will never, "by itself", explain the tensions an individual faces, or the equilibration-processes which he might bring to bear on the problem. Runciman claims that it will not be surprising to find differences of some sort or other between those who are and those who are not “in equilib- rium" (Runciman and Bagley, 1970). But there is no warrant for citing such a finding in support of a "theory of status-congruence" unless specific premises from which it could have been predicted are shown. Two risks are particularly apparent: The first is that it may turn out that the correlation is to be explained not by the discrepancy of status as such, but by a par- ticular feature of the situation which may happen to coincide with a discrepancy of status. The second is that even where dis- crepancies of status are precisely matched with differences of attitude, the difference may be such as can be statistically accounted for by the cumulative independent influence of the two or more 'status factors' involved (Runciman and Bagley, 1970:177). Although Runciman claims that there is no gen- eral rule of relative deprivation emerging from lack of status-congruence, he does feel that the study of "com- parative reference groups" will take the researcher one step closer to an explanation of the relationship. between 1 nisms use same laps reference the claim scope of cause of Ti instead ir precisely explain oi attitude. Th "@101“ th disc”313cm the avail 40 between tensions produced and the equilibrating mecha- nisms used to manage them. ". . . there will be the same lapse into circularity if 'choice of comparative reference group' is cited post hoc as a vindication of the claim that the reference group which defines the scope of a person's resentment was at the same time the cause of it" (Runciman and Bagley, 1970:181). The authors feel its usefulness must lie instead in helping the investigator to formulate more precisely one or more particular generalizations which explain otherwise unexpected or puzzling differences in attitude. The first step must, of course, be to estab- lish a consistent association between feel- ings of 'relative deprivation' and variations in attitudes or circumstances which are logically independent of this feeling. Once given this the investigator will want to show that under specified historical condi- tions a specified reference group comparison will impinge upon a specified group and therefore, given the psychological generaliza- tions applicable in the particular cultural context, give rise to feelings of relative deprivation (Runciman and Bagley, 1970:181- 182). The useful question, for Runciman, is not how many of the person's multiplicity of status-ranks are discrepant, but ". . . which out of the multiplicity of the available comparisons he makes between himself and others, and what are the consequences of this for his other SOC 183). B Lenski pr one hand, one deals ables, ma inconsist interacti. social in‘ subtle, 1: problem." mode] inv inCon51 St 41 other social attitudes" (Runciman and Bagley, 1970: 183). Blalock's (1965;‘l966; 1967) challenge of the Lenski procedure amounts to two major assertions. On the one hand, the additive model of stratification, in which one deals with the main effects of separate-status vari- ables, may explain almost as much as the status- inconsistency model which poses the addition of complex— interaction terms into the development of a model of social interaction (Blalock, 1967). A second, more subtle, issue raised has to do with the "identification problem.“ Blalock suggests: If one takes status inconsistency as a per- fect mathematical function of a difference between two statuses, and if he also wishes to estimate the separate main effects of these statuses, the three 'independent' variables will be confounded together. This can be seen empirically by noting that one cannot vary inconsistency while holding con- stant the other status and inconsistency. Without a priori assumptions, the structural parameters cannot be estimated by least squares or any other purely empirical pro- cedure (Blalock, 1967:69-70). In other words, one cannot take a regression model involving both an interaction-term related to an inconsistency-effect and the main effect of several status variables, and suggest the possibility of 42 theoretical interpretation. A hopeful note is posed in the suggestion that It is true, however, that the existence of statistical interaction or nonadditivity, may be taken as a clue that a status inconsistency effect may be present, though rather definite a priori assumptions must be made if one is to make stronger statements about the magni- tude, or even directions, of the various component effects (Blalock, 1967:70). A more recent challenge to the Lenski formulation (Laumann and Segal, 1971) suggests that more important than the inconsistency-effects due to combinations of ascribed-achieved status configurations, is the per- sistence of subcultural factors as explanations of the social interaction and political behavior of ethno- religious members in urban communities. In a study of Detroit, the authors state that While there are substantial differences among ethno-religious groups on a number of politi- cal and social attitudes and characteristic modes of social participation, net of group differences in educational composition, we have also suggested that the theory of status crystallization affords little if any explanatory power in accounting for the pattern of differences among groups (Laumann and Segal, 1971:55). In others words, there is no overall pattern of statistical interaction between education and ethno- religious group-membership with regard to either social participation or political attitudes. However, specifi. low asc status a status-' effects SUggests effects, interact that Sln Visjb]e reporteC achieVeC gr°“Ps. 43 . there are significant interaction effects involving particular ethno- religious groups that indicate subcultural differences in political orientations and sociability patterns (Laumann and Segal, 1971:55). Interaction-effects in this sample are seen specifically in Jews, who, because of their moderately low ascribed-ethnic status and generally high-achieved status are especially prone to the known effects of status-inconsistency. However, very similar interaction effects are found among German Presbyterians, which suggests that the stability of high achieved status in these groups across generations, and the relative homogeneity of these groups, may be more parsimonious explanations of their similarities than is the theory of status- inconsistency, which in fact would lead us to expect differences in the very areas in which similarities have been observed (Laumann and Segal, 1971:55). Since Jews in this study show interaction effects, and non-whites in an earlier study also show interaction effects (Segal, 1968), the authors conclude that similar behavior on the part of the two highly- visible ethnic groups may be the basis of previously reported inconsistencies between high states of achieved-status and low visibility ascribed ethnic groups. dire inco seve assu (Him Sega effe expe; avere resul latlc 44 It is possible for apparent effects of inconsistencies between low-ascribed and high-achieved statuses to emerge on the basis of social dynamics other than those assumed by the theory of status incon- sistency. Specifically, it is the per- sistence of traits characteristic of ethnic subcultures that leads to these results (Laumann and Segal, 1971:55). Turning next to challenges that have been directed toward the "core assumptions" themselves, there is developing evidence that people will judge the status- inconsistent individual in terms of the average of his several statuses rather than in terms of the Lenski assumption of judgment in terms of his lower status (Himmelfarb and Senn, l969; Berger and Fisek, 1970; and Segal, Segal and Knoke, 1970). Himmelfarb and Senn (1969) suggest that the effects of stimulus-inconsistency were examined in two experiments testing the applicability of a simple averaging-model to impressions of social class. The results of both studies supported an "averaging" formu- lation of impression formulation. Implicitly, it was assumed that people hold expectations that a person's ranking on each of the dimensions of income, educa- tion and occupation tends to be the same or at least. correlated. [Former research has emphasized the] mechanisms people use to avoid presenting themselves in an incongru- ous manner. The present studies, in a somewhat simplified fashion, are relevant to the way judges respond to information is res diffel wise, would farb E the ti multi- formul mechai opera' comes in C0. to co» Perfo reSpo that (BErg SltUa 45 indicative of status incongruity (Himmelfarb and Senn, 1969:50). The results suggest that stimulus-inconsistency is resolved by a simple-averaging process only when the different dimensions are of equal importance. Other- wise, they suggest that a weighted-averaging formulation would provide a better model of status judgments (Himmel- farb and Senn, l969). Berger and Fisek (l970) address themselves to the theoretical issue of how expectations are formed in multi-characteristic task situations. The authors formulate two alternative mechanisms, a "balancing" mechanism and a "combining" mechanism, which may be operative in such situations. The balancing mechanism comes originally from some of the earlier ideas found in cognitive-consistency theories. "Actor p will tend to cognitively balance his situation so as to form performance expectations for self and others that cor- respond with a distribution of states of characteristics that is consistent or univalent for each individual" (Berger and Fisek, l970:29l). The second mode of cognitive definition of the situation is one called a “combining" mechanism; The ideas involved here are loosely associ- ated with those from information and decision-making theories. According to (l9 we we Len: mecl ses tha' ind ind' ter heac the The l'hu are Sega 46 this mechanism. the actor essentially operates as an information available process system, taking into account all information available to him as regards the relevant status characteristics and the task in the situation (Berger and Fisek, l970:292). The results of the experiment and the Paul Tress (1971) replication which resulted similarly, confirm the predictions made from the "combining" argument, thus yielding evidence against the "core assumptions" in the Lenski formulation that suggest some kind of "balancing" mechanism (Lenski, l966). Segal, Segal and Knoke (1969) pose two hypothe- ses in terms of Lenski's "core assumptions." One is that the best predictor of a status-inconsistent individual's evaluation of his social class will be the average of his several objective statuses. As status- indicators Segal, Segal and Knoke use such achieved cri- teria as education, family income and occupation of the head of household. As the dependent variable they use the individual's subjective social-class identification. The first hypothesis is not supported for status- inconsistent individuals, in that maximizing processes are used in evaluating social class position (Segal, Segal and Knoke, l969). These three research studies indicate that: H0 t0 be cons both the . “Sing Sim inc°“Slst tions w” the lndi Se]f (La expeCtai Eflglfli majOr EQui\ 4.7. (a) people may not judge an individual on the basis of his lowest status-rank (Himmelfarb and Senn, l969) (b) the individual evaluates his own status- configuration based on some 9combinationf of his status ranks (Berger and Fisek, l970) (c) that some kind of "averaging" process of status ranks for ego may be involved (Segal, Segal and Knoke, l970)‘ However, in order for Lenski's “core" assumptions to be considered false, it would have to be shown that both the individual and others, at the same time, were using similar "averaging processes" to evaluate the inconsistent individual. Otherwise the "core" assump- tions will be maintained since others will attribute to the individual lower status than he attributes to him- self (Laumann and Segal, l97l) and thus his deference- expectations will be frustrated (Lenski, 1966). Conclusion In this chapter we have reviewed the "status- consistency" tradition and suggested that there are three major theoretical positions stemming from Benoit- Smullyan's (1944) study of social status and problems of equilibration. The first of these, the w‘r expectan< works of (l963; lE as the "c the works are the < Anderson on the 01 was witii POSition. able fOr data, an< done in , other f0! JustiCeu apt topit 48 expectancy-conflict position, includes primarily the works of Lenski (1954; 1956; 1964; 1962) and Sampson (1963; 1969). The second tradition, which we referred to as the "cost-gain outcome" formula is found primarily in the works of Kimberly (1962; 1966; 1970). Finally, there are the dissonance-balance formulations of Zelditch and Anderson (1966) on the one hand and J. Stacy Adams (1965) on the other. The primary consideration of the chapter was with the Lenski tradition in the expectancy-conflict position. Emphasis was on the fact that the data avail- able for the dissertation were similar to Lenski's survey data, and while other formulations derived from research done in experimental situations. Secondly, two of the other formulations were concerned with "distributive justice" rather than "status consistency" and therefore an apt t0pic for separate consideration. Within the expectancy-conflict tradition, Lenski's "core" assumptions deal with the lack of consistency between thigh and low" states of achieved and ascribed variables. The implicit assumptions suggest that when an individual evaluates himself, he gives greater weight to his higher states on these status-variables, and that others give greater weight to his low states. However, the individual expects others to evaluate him on the basis of his highe the two e Effort to ary" set Most crit attempts rather th R the Lensk ranks is there wi] 0f others drawal fr Closer 10 Cate diff ity, PTT' inconsist This is e the 10W S I they affe ab1e and «secondar tT'Ons he] 49 his higher ranks. When there is a discrepancy between the two evaluations, tensions in the individual result. Effort to manage these tensions gives rise to an "auxili- ary" set of assumptions observable as response-processes. Most critical reviews of this literature have been attempts to test the "auxiliary" set of assumptions rather than the "core" (Anderson, 1971; Lakatos, 1968). Review of a representative sample of studies in the Lenski tradition suggests that if at least one of the ranks is a low state of an ascribed status-variable, then there will be either an attempt to change the evaluations of others via such means as political extremism, or with- drawal from these evaluations into personal isolation. A closer look at the nature of ascription shows it to indi- cate different states of both changeability and visibil- ity. Prior discussions have emphasized that reactions to inconsistency are greater where ranks are non-changeable. This is especially true when the non-changeable rank is the low state of a status-variable. In terms of different states of visibility as they affect status consistency, ranks which are change- able and lack visibility may be seen as aspects of "secondary stratification." The following set of assump- tions relates to these secondary stratification ranks. implied form of retreat excellin Status-y methOdol aSsumpti ever, th ration 1 StatUS-r "loin“ this Cas «expecta "Core" 6 cal]y by area has 50 (a) Secondary ranks are emphasized by P only when a highly visible rank is low. (b) A secondary rank is recognized by only a limited membership group and whenever the individual steps outside that group, the validation is a problem. Lenski's withdrawal-response to tension, which is implied to be an essentially negative phenomenon and a form of isolation, may instead be interpreted as a retreat to second-line defense, the individual then excelling in low visibility ranks even at the expense of status-validation problems outside specific groups. In addition, we saw that there were some serious methodological challenges, primarily to the "auxiliary" assumptions but to the "core" assumptions as well. How- ever, the general conclusion reached was that a configu- ration including at least one low state of an ascribed status-rank seems to lead to the equilibrating process originally elucidated by Benoit-Smullyan (1944), and in this case specified by Lenski, as a variant of the fexpectation-conflict variant set.f Although Lenski's "core” assumptions seem to have been challenged theoreti- cally by the studies discussed, the research done in this area has not basically altered the plausibility of these assumptic do not 5L 51 assumptions, nor shown that the "auxiliary" assumptions do not support the "core" assumptions (Anderson, 1971b). The Bat consist the pre analysi Probabi Detroit year 19 Area St 9a" (Fr “\Udy 816 837 843 849 855 CHAPTER II DATA The Data to be Used In terms of the previous studies of status- consistency, we will now discuss the data dealt with in the present study. The empirical data upon which this analysis is based were gathered from five cross-sectional probability-sample surveys of the adult population in the Detroit metropolitan area, in 1953-1958, exclusive of the year 1954. These surveys were conducted by the Detroit Area Study, a research group of the University of Michi- gan (Freedman, 1953). The Detroit Area Studies Study Year Title of Study 816 1953 "Ideal family size in Detroit" 837 1955 "Orientation of moral issues in a metropo- lis: the meaning of work" 843 1956 "Party leadership and political behavior: intra-class correlation of attitudes in Detroit" 849 1957 "Religion in the metropolitan community“ 855 1958 "The vitality of supernatural experience" 52 stage. C Detroit in Wayne samples all adul metropol (tracts, selectic Proport' its Chat unit Cor Chance . errors is base: annua] anan] Samples and the gathere. AithougI appeare‘ 53 The sampling procedure employed was a three- stage, cluster-area sample. The area embraced in the Detroit Area Study surveys are the census-tracted area in Wayne, Macomb and Oakland Counties of Michigan. The samples used were designed to yield a cross-section of all adults living in private households in the Detroit metropolitan area. In the sampling technique employed (tracts, blocks, and dwelling units), the probability of selection of every dwelling unit is known. Probabilities proportional to the size of the selection-unit governed its chances of being used at each stage. Each dwelling- unit contained within the tracted area possesses an equal chance of entering the sample (Kish, 1951). In order to maximize sample-size and minimize errors in sampling—variations, the current study, then, is based on a sample of cases drawn from five different annual sUrveys. Such combining of data from several annual studies is possible since the Detroit Area Study samples have been designed to facilitate such procedure, and the data necessary for this particular study were gathered in comparable form in all five surveys used. Although the probability that the same people could have appeared more than once is greater than zero, we will 54 not discuss the issue of "automatic correlation? (Galtung, 1967). Data used for analysis were collected by an established research-organization. Both advantages and disadvantages resulted from the fact that the primary research-objectives in collection differed from the central goals of this study. We can deal with what per- tains to major educational and income levels of an urban community because of the study's sampling of a major metropolis. However. data relevant to the condition of secondary analysis were not available for our study, groups of pe0p1e interested in theoretical study being very small in a random sample. If a sample were designed expressly for the present research problem, Blacks and Jews would be included. In addition, efforts to test Lenski's "auxiliary assumptions" (that tensions lead to attempts to change the political system, and withdrawal from social participation) were restricted to political preference on the one hand and frequency of contact with relatives and other friends, on the other hand (Axelrod, 1956). Additional attitudinal as well as subjective measures might have provided more direct indications of the dependent variables. 55 Following are the indicators used, the first column referring to the column-number of the summary-deck for data common to all DSA surveys. 17 30 40 67 68 80 Relative ranking on annual income of family head (above or below the median) Political preference (Democratic) Age (forty-four and younger, or forty- five and older) Education of ego (above or below the median) Frequency of contact with relatives (a few times a month or more often) Frequency of contact with other friends (a few times a month or more often) Religious preference (Protestant or Catholic) The N of the sample consists of 1,340 Protestants 'and 1,027 Catholics. Concerning frequency of contact with relatives and other friends, the sample is restricted to two of the studies, 843 and 855, and the N for Protes- tants in these columns is 603 and for Catholics is 485. Methodological Challenges As already shown in Chapter I, there are several methodological and substantive problems that have to be dealt with when we try to test the Lenski formulation with sample-survey data. 56 It has been suggested that in the absence of explicit criteria of consistency, most studies of status- discrepancy employ procedures that implicitly adopt a frequency-definition of norms. Typically, status- consistent and status-discrepant subgroups are identified in one of two ways: . either the component status variables are cross-tabulated, with the cells of the cross-tabulation on or near the main diagonal defined as consistent and the cells in the extreme of the table defined as discrepant; or, alternatively, each component variable is partitioned into percentile-classes, each class scored with its mean percentile, and a measure of the dispersion of scores over status variables--a status discrepancy score --computed for each individual. Whatever the procedure, the sample is normally divided into groups varying in degree of status dis- crepancy, and comparisons on the criterion variable are made between these groups (Treiman, 1966:652). Blalock (1965; 1966; 1967) agrees that in order to determine the effects of rank-inconsistency on a dependent variable, the effects of the rank-hierarchies on the dependent variable must be taken into account. How- ever, he demonstrates that it is not possible, in the light of current statistical theory, to empirically dis- tinguish rank-hierarchy from inconsistency-effects (cited in Hyman, 1967:386). Blalock (1966) suggests then that although educa- tion and income are probably causally linked, wi be mi 91‘ ti "111 an 57 . . we can imagine varying education while holding income constant (or vice versa). This is true because we can presumably manipu- late other causes, say of income. These could be manipulated so as to counteract the changes in education, in much the same way that a physicist might hold temperature constant (while increasing pressure) by introducing a compensatory cooling agent. But if status- inconsistency is defined as the difference between education and income, it is manifestly impossible to vary one factor while holding the other two constant (Blalock, 1966:56). However, this model implies that a liberal vote or lack of social participation is caused by inconsistency, per se. But Lenski posits that inconsistency produces "strains," which in turn lead to liberalism and social withdrawal. If there are other factors affecting strain besides status-inconsistency, it may be possible to per- mit identification. Unfortunately, however, there is a practical complication that re-introduces the identifica- tion problem through the back door: Lenski and other students of status- inconsistency have not been able to measure the strain factor directly, and have had to use status-inconsistency as an 'indicator' of strain . . . and we are back where we started (Blalock, 1966:56). An obvious implication, then, is that attempts must be made to obtain independent measures of strain, and then to: relate these separately to inconsistency on the one hand, and to the dependent variable on the other. But careful conceptualization 58 will be necessary, since the theory calls for a special type of strain, due primarily to status inconsistency rather than 'generalized' strains (Blalock, 1966:56). Thus Blalock has spelled out the problem of sep- arating the effects of status-inconsistency from those of individual-status variables. An "identification" problem is presented by the existence of too many unknowns, when inconsistency is defined as ". . . a perfect mathematical function of the difference between two (or more) statuses" (Blalock, 1967:305). If one is trying to explain the variation in some dependent variable by a combination of status-variables plus inconsistency, it will not be possible to estimate the coefficients without making a priori assump- tions that place restrictions on the model. The necessity of making some restrictive assumptions in order to identify the various component forces one to face up to the problem of clarifying the theory to the point where this is possible (Blalock, 1967:305). This then is the problem, and it can be seen as quite formidable because there are incredibly many pos- sible parametric values that might have produced the same empirical results. In some instances, then: . one particular set of values might seem more plausible than others, on either theo- retical grounds or because the interpretation is much simpler. But if one wishes to select that most plausible set of values, he should 59 do so with the full realization that implicit assumptions must be made which rule out the remaining alternatives (Blalock, 1967:306). To choose between alternatives, one then would have to use criteria such as the theoretically simplest or most plausible. However one is able to accomplish this, the caveat remains: POne should always evaluate the evidence in favor of any particular theory against all plausible rival hypotheses“ (Blalock, 1967:308). Lenski's (1954) original assumptions amounted to the presumption that inconsistency has the same impact on the dependent variable, regardless of the direction of the inconsistency. In other words, the contributions of the inconsistent cells are both positive and equal. How- ever, it is equally possible to get results when: (a) There is an inconsistency-effect in one direction only. (b) The effect is positive in one direction but negative in the other. (c) One or the other of the main effects of the status-variable is zero. (d) The main effects of the status-variables are in opposite directions and the magnitudes of the inconsistency are quite large (albeit in opposite directions). 60 (e) The composite table shows no interaction in spite of the fact that there are incon- sistency effects (cited in Blalock, 1967: 306). Jackson's (1962) assumptions are that neither status-variable has a main effect, and there is an inconsistency-effect only when ascribed status is high and achieved is low. Although the alternatives are some— what more restricted than those of Lenski's assumptions, the alternatives to Jackson's (1962) directional assump- tions are: (a) Inconsistency effects appear only when the achieved status is "higher" than the ascribed. This would require that the two main effects be opposed, however. (b) Among other alternatives, a situation where inconsistency effects are equal but in oppo- site directions (cited in Blalock, 1967:305). The major alternative assumptions made in this study will be to presume that inconsistency might have opposite effects, depending on direction (Hyman, 1966; Mitchell, 1964). However, Blalock suggests: One might raise the objection that the single variable 'inconsistency' could hardly be expected to have Opposite effects. It would 61 perhaps be better to reconceptualize 'incon- sistency' in terms of two (Or more) distinct variables, e.g., (l) the degree to which (income) exceeds education, (2) the degree to which education exceeds (income) (Blalock, 1966:60). A second alternative assumption made in this study is to suppose that inconsistency-effects in a given .fi 7". direction are taken as proportional to the main effects of whichever status variable has the lower value. Blalock suggests, in support of this assumption, . one might argue that it is the lower of the two statuses that produces strains in the individual and that his behavior due to the strain component ought therefore to be essen- tially similar in nature to that produced by the 'main' effect of this particular status variable (Blalock, 1967:314). I 72:; In this case assume that we are dealing with the portion of a table in which income is lower than educa- tion. We would then take the inconsistent effect to be proportional to the main effect of the income variable. Similarly, if education is lower than income, then incon- sistency is assumed to produce effects proportional to the education-effects (cited in Blalock, 1967:314). These have been referred to by Hyman (1967) as assumptions based on "status effects.P Notwithstanding all this, Blalock suggests that . . there is a sense in which the existence of interaction can be helpful in appraising 62 the utility of the inconsistency formulation when predicting to specific dependent vari- ables. . . . Provided one assumes that the individual status variables produce effects that are additive, then the amount of inter- action represents a kind of minimum 'net' inconsistency effect. Thus, whenever one finds interaction, he might suspect an incon- sistency effect, though it may be difficult to pin this down to specific cells. Empiri- cal support for or against status incon- sistency models would therefore seem to hinge on the interaction term (Blalock, 1967:308). The existence of interaction might then be taken as weak evidence in favor of an inconsistency-effect, provided that alternative explanations for interaction can be eliminated (cited in Blalock, 1967:305). Finally, there are a number of alternative expla- nations for interaction, including: the possibility of sampling error, differential measurement errors, approaches to upper (or lower) limits due to measurement artifacts, non-linearity combined with multi- collinearity, and other specific non-additive models (e.g., multiplicative effects). They also include the possibility that the incon- sistency effects are spurious (e.g., due to age or ethnic differences that have not been controlled) (Blalock, 1967:308). We will attempt to look at some of these problems and suggest how our data based on the Lenski formulation respond to these challenges. Among Blalock's challenges there is the problem of "alternative explanations for interaction." As a form of "specification error," there 63 may be errors due to missing variables, in which case the interaction component would disappear if these disturbing influences could be controlled (Blalock, 1967a:72). Blalock uses, as an example, ethnic groups with different economic and cultural (religious) backgrounds being con- sidered in the same category of low ethnicity. One of the major purposes of this study is to control for eth- nicity by separating the sample into Catholics and white Protestants, and investigating the degree of interaction between income and education in these different groups. Another example of spurious interaction would be that ". . . whenever income is considerably lower than expected on the basis of education or occupation, it is quite likely that age needs to be controlled" (Blalock, 1967a:72). The assumption is that the investment in education needs several years for a commensurate income, until a degree of proportionality is reached (Homans, 1961). In order to control for this factor. we have differentiated between those above and below forty-five years of age. Although age, as an aspect of "structural crystallization“ (Smith, 1969), does seem to have a definite effect onl 64 interaction, the relationship should not be primarily an aspect of the frustrations of youth, as suggested by Blalock. Age has been controlled in several studies in the Lenski tradition, but very few, other than Smith (1969), have treated it theoretically as a method of expanding the problems of inconsistency and placing them in a wider social context. Smith's "Structural Crystallization" as a Methodological Contribution Smith (1969) has made an important contribution to this problem by distinguishing between “status incon- sistency" and "structural crystallization." Crystalliza- tion seems to be a contextual property of the social structure which has something to do with what are con- sidered ascribed or achieved statuses. The contribution of Smith's position derives from his idea that what is considered an ascribed or an achieved status is a function of the broader social context. As an indicator of crystallization, Smith uses age, explaining: Because of its personal and social correlates, aging tends to superimpose an effect upon the variability of achievement, a phenomenon reflected in the fact that change in achieved status usually slows down and often stops in later middle age. In effect, aging tends to fix achieved rankings, making their elastiCi- ties more equal to the elasticities of ascribed statuses (Smith, 1969:909). 65 We can look at statuses in terms of how subject to change they are on the part of the individual, such as their elasticity and inelasticity. At the same time one can look at how ambiguous to the individual are the relation- ships between the status-attributes in his status-set and how well he can control the visibility of these status- relationships to others. Smith suggests that As crystallization progresses, ego's efforts contribute decreasingly to any changes in his status. The second development, invariable relationships among statuses, suggests that crystallization is likely to reduce the ambiguities about status generally (Smith, 1969:910). This reasoning then has implications for refining predictions of the effects of status on selected dependent variables. Implied are the decreased relevance of achievement and increased relevance of ascription with increasing crystallization. The reasoning seems to be that individuals will distribute their energies into activities which produce rewards: those arenas of activities in which investments in achievement pay off in status returns, will be more 'salient' to an individual's assessment of his reward- position than those in which his rankings are fixed. Thus, we may infer that the salience of achieved rankings exceeds that of ascribed rankings as long as crystallization is low. With advances in crystallization, conversely, the individual's capacity to change status 66 rankings decreases, and in consequence, the salience of all his rankings to the determina- tion of his reward position should'become more nearly equal (Smith, 1969:911). ' ' Smith then suggests the following hypotheses: (l) Inconsistencies among achieved rankings will (2) be stressful in pr0portion to the salience of achievement. Thus, we shall expect incon- sistent achieved rankings to produce support for liberal parties when crystallization is low but not when it is high. Inconsistencies between paired ascribed and achieved rankings will be stressful in propor- tion to the degree of overall crystallization. In other words, as crystallization increases, so should the stressfulness and overall parti- san liberalism produced by this pattern (cited in Smith, 1969:914). In the main, Smith's hypotheses are borne out. speculation However, one major discrepancy is that inconsistencies between religious status and achieved status produce an effect only when the overall ensemble of statuses is relatively uncrystallized. The prediction had been that interaction effects would occur only between achieved- ascribed groups when crystallization was advanced. The explaining this result is: 67 . . . we are observing an inconsistency between statuses that are not partiCularly viSible in a social sense. Although social (i.e., others') evaluations of status dimen- sions do proceed with greater independence under low degrees of crystallization, intra- psychic awareness of inconsistencies among them, nevertheless, occurs and is stressful (Smith, 1969:919). These phenomena of "social awareness of visibil- ity“ and "intra-psychic" stress related to inconsistency need to be clarified and the discussion of “secondary stratification characteristics" (Blau, 1964) which was introduced in the theory chapter may provide some insights in this area. Further discussion will resume in the analysis chapters where an attempt will be made to clarify this divergence from the Smith hypothesis, while at the same time presenting rival and alternative hypotheses. Additional Blalock Methodological Challenges Another type of "specification error" posed by Blalock (1967a) is a multiplicative formulation, a . . natural extension of the notion that in order for Y to be present, both XI and X2 must also be present. Then it would require a combination of low ethnic status ang_high education in order for this felt need (to change the political system) to be translated into liberal voting behavior (Blalock, 1967a: 73 . 68 The suggestion is that it is not strain due to inconsistency that causes liberal voting behavior or lack of social participation,_but rather a strain due to ethnicity which needs an additional factor, education, to make people aware of the need for political change. Blalock's multiplicative model, in the specific example cited above, would say nothing about the second order- interactive effects of income and education. As a result, both high income and high education, as well as low income and high education, would be considered equal in terms of their interactive effects when combined with low ethnicity. However, as we will show in the follow- ing analysis chapters, alternative theoretical positions might predict that different combinations of high and low states of income and education might have interaction effects when paired with Protestant or Catholic back- ground. A final kind of spurious interaction measured by Blalock is the suggestion that persons with like number of years of education might not be equal, since persons listed as having 'high' educa- tion but realtively low occupations or income may very well have graduated from low status community colleges, trade schools, or two year business schools. In effect, they will have been misclassified as to educational status throuqh the use of a rough indicator such as 69 number of years of schooling (Blalock, 1967a: 72). According to Hauser, however, the existence of grade-specific differentials in academic performance by ethnic background or socio-economic origin is not a suf- ficient demonstration of an academic argument. Years of schooling as an indicator of educational quality, in terms of available evidence suggests that educational attainment is a rather good indicator of educational quality. We already know that educational attainment accounts for most of the influence of social background on adult achievements; refinements of intervening variable cannot add to the importance of education in that respect (Hauser, 1970:112). Although Hauser makes these statements on the basis of his experiences, it is admittedly difficult to evaluate his argument. The strongest support for the use of education as a status-indicator based on years of schooling, is the fact that this is the kind of data which I happen to have. An "Ideal Test" of the Lenski Formulation This section contains a discussion of what the data for an ideal test of the Lenski formulation should look like (Anderson, 1972). Lenski's theory in fact pre- dicts the following: tion (1) (2) (3) (6) 70 If a person P (primarily) evaluates himself on the basis of his high rank states; and, If P's significant others evaluate him (pri- marily) on the basis of his low rank states; and, If P expects the significant others to evalu- ate him on the basis of his high rank states; and, If P perceives the discrepancy between the expected and received evaluations from the significant others; and, If P perceives that voting Democratic con- tributes to social change that would affect his low ranks; then, P is more likely to vote Democratic than a person whose ranks are consistent (Anderson, l972zl). An ideal survey would have to contain informa- relevant to each one of the 'if- statements' above. For the conclusion (6) does not follow unless each of the five 'if- statements' is satisfied. In available sur- veys the practice seems to be to assume the validity of the five 'if-statements.‘ A sur- vey that settles the five'if-statements' by assumption is however a very weak test of the 71 theory, compared with one that could check them out empirically (Anderson, l972:2). In terms of empirical tests of these assumptions, only the first if-statement has been empirically tested. Segal, Segal and Knoke (1970) hypothesized: ". . . the best predictor of a status-inconsistent individual's evaluation of his social class will be his highest objec- tive status" (Segal, Segal and Knoke, 1970:350). However, results of their study of a national sample of American adults suggests that status- inconsistent people average their several statuses in defining their social positions, rather than maximizing their claims by disregarding low statuses (cited in Segal, Segal and Knoke, 1970:347). In addition, the authors implicitly question assumptions 2-4 in speculating about alternative assump- tions derived from "symbolic interaction" theory (Mead, 1934). Other pe0p1e's evaluation of one's status will be incorporated into the self through the objective 'we' (Mead, 1934). On the basis of this principle we would expect a person to come to view himself as others view him. Similarly, in terms of Cooley's (1902) 'looking-glass self', we would expect an individual to bring his own definition of his social status into agreement with the evalua- tions of others. He will not maximize the evaluation of his status unless others do so. In any case, there will not be a gross 72 discrepancy between an individual's evalua- tion Of his sotial position and the evalua- tidn of others with whom he comes in contact (Segal, segal and Knoke, 1970: 348). In terms of the fifth assumption that voting Democratic contributes to social change that would affect low ranking, it is by no means clear that: . all electors voting Democratic support programs of social change. No doubt differ- ent people vote Democratic for different reasons, even within a single city, and it is their conceptualization of their intention that defines their action, and whether that action can be taken as a case of political liberalism (Dorein and Stockman, 1969:50). Anderson, finally, has suggested: it would be very hard to get the infor- mation in question in a survey like the Detroit Area Study. But that means that Lenski's theory may be, in an important sense, untestable through survey research (Anderson, l972:2). Conclusion and Review In summary, then, we have looked at some of the methodological difficulties related to alternative models to the additive function of component-status variables. The "mean value theorem in sociology" (Treiman, 1970), suggests some of the forms that the additive model of stratification may take: At the aggregate level. at least, the behavior associated with any given status configuration will be a weighted average of the behaviors associated with each of the 73 component statuses in the configuration. Individuals may strike a balance between the behaviors appropriate to each status; or may choose to act in terms of one status on some occasions and in terms of the other on other occasions, depending upon which status is more salient at the time; or may permanently resolve any conflict by always acting in a manner appropriate to whatever status is most salient in general. All of these alterna- tives imply an additive effect of status variables on behavior at the aggregate level (Treiman, 1970:225). In contrast to the different forms that the addi- tive model may take, there are several alternative models of status inconsistency, one of which suggests that status inconsistency, per se, has a disorganizing effect on individuals. There are also more directed approaches relating achieved and ascribed status variables in any number of combinations. It is important to remember, however, . . . that the two alternatives are not contradictory, but that the social pathology theory requires that in addition to additive effects of the status variables on the behavior of interest there be interaction effects as well (Treiman, 1970:226). If one feels, theoretically, that the interaction effects should explain a large degree of the variance, then "inconsistency effects" will have to be predicted and explained. CHAPTER 111 VOTING DEMOCRATIC AND STATUS INCONSISTENCY Introduction Studies using voting behavior as a dependent variable for inconsistency-effects have been rather extensive since the 1950's. In his earliest study of status-crystallization, Lenski suggested that a lack of status-consistency is as important in explaining politi- cal attitudes as the more conventional linear models relating social class and voting behavior. His findings indicate that a lack of status-consistency pushed one in the direction of voting Democratic to an extent which could not have been predicted from the linear, additive model (Lenski, 1954). In a later (1967) study, he sug- gests that voting Democratic and liberal attitudes in economic areas are indicative of a political reaction against the social order. In other words, as one indi- cator of inconsistency-effects, voting Democratic is a reaction against the system, an attempt to alter the 74 75 political status quo through the support of Pliberal“ movements. Lenski himself summarizes the dynamics involved in this relationship in the following paragraph: For some years voting behavior has been regarded as a highly relevant variable in studies of the effects of status- inconsistency. It has been hypothesized that status inconsistency is a source of stress far individuals, especially when the incon- sistencies are substantial, since the individual prefers to think of himself in terms of his higher status or statuses while others have a tendency to treat him in terms of the lower. (This is what one would expect if he assumes men are strongly moti- vated by self interest.) The resulting con- flict between expectations and experiences is bound to be disturbing, and it has been hypothesized that one common pattern of response is to react against the social order, by supporting a political party advocating change. If this reasoning is correct, one would then expect persons occupying incon- sistent statuses to be more likely than per- sons occupying more consistent statuses to support liberal and socialist parties (Lenski, 1967:298-299). A base-line model of voting behavior, then, would assume that there is a push toward the Democratic party if one is of low educational. financial or occupational status; and there is also a push in this direction for members of ethnic and religious minorities. At the same time there is a push in the Republican direction for per- sons with high social status in general, and for members of certain "core" Protestant churches (Segal, 1969:352). 76 Soczial and Economic Formulations of Votim B'eha’vi'or' ' ' What might be the link between these political anti social variables? In other words, what assumptions Lumierly these relationships? One might suggest that there are two basic assumptions underlying the relationship of socrial status, political behavior and attitudes. One can use: a model of rational self-interest: "Each citizen in our~ model votes for the party he believes will provide hini with a higher utility income than any other party dur"ing the coming election period" (Downs, 1957:38). This economic interpretation as the major deter- TDlnant of political behavior has been criticized, however. It has been argued that the economic model itself does not explain American political behavior and research seeking to relate economic self-interest to political invovement has been contradictory (Lindenfeld, 1964; Segan and Knoke, 1967). Social processes rather than economic pressures, can the other hand, are often cited as an intervening n: echanism: The higher the identification of the individ- ual with the group, the higher the probability that he will think and behave in ways which distinguish members of his group from non- members (Campbell, Converse, Miller and Stokes, 1960:307). .GZHII '1 77. The simple determinism of the economic-market priiiciple in class action is seen as complicated by the existence of status groups (Knoke, 1969). Lenski suggests that religious and racial minori- tieas tend to be status-groups in the sense in which Weber employed the term. they are groups which are differenti- ated in terms of social honor, and where honor , and respect are denied to a particular group, i its members tend to react critically toward ‘ the social system as a whole, its key insti- tutions, and their leaders. Thus it appears that American radicalism derives at least as is much from the 'status group struggle' as it does from the more familiar class struggle. In other words, the denial of equal honor and respect to all socio-religious groups may be as powerful a factor in stimulating discon- tent as the denial of economic advantages and political authority (Lenski, 1961:173). razmna Hi One might, then, consider two sets of assumptions based on the question of whether class or status variables seem most important in their effect on the dependent vari- able. The class model might consider such variables as income and education as important. Here a straight Homans (1954) "investment-reward exchange relationship" may be seen to hold: This assertion is consistent with views of social behavior as an exchange process, which suggests that people will attempt to achieve a condition of status congruence at the level of their highest status (Homans, 1962:15). 78 Investments will have to be related to rewards in some relevant way. The individual's evaluation of his own social position will not be in accord with the social position given him by others with whom he comes in con- tact. The observable response-processes, as they might affect political behavior and attitudes, are processes attempting to raise one's lower ranks. This model then would be concerned with the rigidity or changeability of J. these lower ranks. The more fixed and less changeable, the greater the likelihood of direct political action to change the system. Age may play a part here, the lower ranks of youth being considered more changeable and individual mobility seen as possible. The individual is less likely to suffer from inconsistency effects. As summarized by Anderson and Zelditch: If Ego finds his rank inconsistent in compari- son with alter, he will get upset to the extent that he perceives that upward mobil- ity in his loWer'ranks is blocked. If the situation is perceived to be transient then Ego is not likely to get upset (Anderson and Zelditch, 1964:118). ' In regard to which factors will influence Ego's reactions, the authors continue: If a person experiences deprivation associ- ated with his present status then the stronger his expectation is that he himself will be able to move to a higher status with less depriVation, the less likely he is to Combine 79 with others to remove the causes of depriva- tion in his present status (Anderson and Zelditch, 1964:118-119). Since status-differences are the basis of our second model, ascribed rather than achieved differences become more important. In the evaluation of esteem and honor, the factor of visibility and lack of ambiguity about status-relationships become crucial variables. In considering the effects of status-consistency on self image, one of the major factors may be the visibility of ascribed statuses on the blocking of social acceptance as they might be involved in some "rebuff process" (Zelditch and Anderson, 1966). The lack of social acceptance of a nouveau riche ethnic by his new peers obviously depends upon his “tainted" visibility, especially when his educational attainments are obviously below his income. Homans refers to a lack of "social certitude" as the . . simultaneous holding of high and low ranks creating doubt in others as to the reality and legitimacy of the individual's higher ranks. A certain minimum amount of . social visibility is necessary before rank- inconsistency can have the painful conse- quences described by Homans (Hyman, 1967: 395). Ascribed statuses are significantly based on two factors. These are: first, the degree of visibility of the individual's statuses to significant others, and 80 second, the degree of ambiguity of the relationship between statuses in an individual's configuration to that individual. In terms of observable response-processes, the individual may on one hand be reacting to blocked social mobility related to achieved variables, and on the other hand to blocked social acceptance related to the visibility of his lower ranks, especially if they are seen as ascribed.» As blockages in social acceptance are considered an aggregate, rather than an individual phenomenon, (both to ego and to alter), collective action may be seen as one kind of appropriate response, and voting Democratic can be seen, in this case, as an aspect of collective reproach of the social system thrOUgh its political structure. However, one might question, at this point, why the reaction to society should lead in a liberal rather than a conservative direction. The reasoning, derived from Smith (1969) suggests that To the degree that statuses attributed to ego are fixed, we may speak of the person as being in a condition analogous to a dependency relationship vis-a-vis the larger society (Smith, 1969:910). Smith sees minority groups, because of the nature of their ascribed statuses, as being "dependent groups" 81 and suggests that such groups generally vote Democratic. This derives from: . the apparent appeal of the Democratic Party to dependent groups generally, and from the apparent ideolOgical distinction between Democrats and Republicans-~i.e., with the Republican Party generally reputed to attract individuals favoring a kind of 'laissez- faire' individualism, and the Democratic Party, those favoring the idea of a 'Welfare t state' (Smith, 1969:913). - Ti( Structural Crystallization and Voting Behavior What seems to be at stake here, then, are the partisan responses to two kinds of "blocking." On the one hand, is the class-oriented mobility-blocking related to the degree of changeability of achieved variables? On the other hand, is the blocking of social acceptance of a status-oriented model where the visibility of the ascribed status characteristics would determine the responses to inconsistency-effects? In both cases it might be suggested that structural crystallization (as measured by age) affects the degree of changeableness and visibility in these two models and would add in both cases to the saliency of response related to voting- preference (Smith, 1969). In other words, with increas- ing age, both the individual's ability to change his statuses through effort decreases, and the individual's liv— ’ . I 82 status—relations become less ambiguous to himself and more visible to others. At this point it is apprOpriate to see whether Smith's (l969) hypotheses generally hold for our data, in order that we may determine whether to include it in the more formal presentation of the two earlier models. One might predict that due to the salience of achieved statuses for the young, one would expect incon- sistencies in statuses such as income and education to cause a problem with a status-uncrystallized population. The reasoning, according to Smith, is that: the degree to which the total effect (partisanship) is expressed varies with the salience of the given status dimension. The questions involved become easier to deal with when we consider achieved and ascribed dimen- sions separately. Since we anticipate that achievement is salient while subject to vari- ation, we expect achieved statuses to be more salient than ascribed statuses when crystalli- zation is low. (Below age 45) As long as ego can change achieved rankings but is unsuccessful at resolving inconsistencies among them, he will experience frustration at his inability. Thus, we expect incon- sistent achieved rankings to produce support for liberal parties when crystallization is low)but not when it is high (Smith, 1969: 912 . From this one could predict that when comparing income and education-inconsistency effects on voting Democratic, for a Protestant sample under age forty-five, 83 this status-uncrystallized group would exhibit inconsistency-effects. Referring to Table I, the inconsistency-effect will be computed by taking a "difference of differences" in the cell pr0portions, namely by subtracting the sum of the proportions in the consistent cells (high-high, low- 3 low) from the comparable sum in the inconsistent cells (high-low, low-high). The resulting measure may be either positive or negative, with a positive difference indicating a surplus in the proportion of respondents favoring the Democratic party in the inconsistent cells over the consistent cells (Smith, 1969). The data in Table I shows a strong inconsistency effect for young Protestants (+33) and in the predicted direction. In considering advancing status-crystallization and its effect on achievement-oriented inconsistencies, one would expect a lessening of inconsistency effects with increasing age. The reasoning here follows from the statement that: When achieved rankings have become set in a crystallized ensemble and are not subject to further variation, he will not experience frustration even if some inconsistencies still survive. 'As achieved rankings become inelas- tic, they lose part of their salience, and continued striving thereafter becomes unre- warding and is abandoned. The removal of e4 Ammuznwcuum magnum esp cmvvmz m>oa< cmwumz 3o_om cowpmuzum carve: w>on< cmpvmz sopmm osoucm ewe: 204 so; mewsouocuvv pco>m~mmv m_+ Ammpv “Ne hamv ewe Amm_v Nmm Aaov umm maammmmw Pm- A¢m_v we“ Ammv ems Ammv Rom Ame_v sow “geomaooca caa_o age we "mufl N + Ammv Ram Aoomv New Aom_v emm Ammmv wok uvpocceo mm+ Amemu New Ampmv New ANm_V New ._mmv Noe Beaumauoaa we cove: "mm< 304 saw: 204 saw: cowpauanm 204 so; now: now: mEoucH mmucmgmmmwo Loom cm>mwzo< cm>mwgu< ucmspww< mo mucmgmwewo cave: cm>o Aammiumuidcw_Focpcouv Li muHoncuzmpmHmzoqu onh zo mcuuaam _ m4m<~ 85 the effort consequently eliminates the frus- tration. Only under low crystallization, therefore, would we expect inconsistencies in achievement to produce a disposition to favor social change (Smith, 1969:914). ' From the above one would expect that Protestants, over age forty-five and compared once again for income and education, would exhibit little if any inconsistency effects. Although there are some inconsistency effects, they are in a negative direction (-21) and thus gives some slight support for the hypothesis. (See Table 1.) Considering next paired achieved-ascribed differ- ences among the structurally uncrystallized, the sugges- tion is that prior to crystallization of one's ascribed status, its salience is at a low level and not likely to appear visible or aid in reducing the ambiguity of status relations to the individual. Income and education com- parisons for Catholics under age forty-five, it would be expected, would not lead to "inconsistency effects." The data suggest little if any inconsistency effects (+2). In an uncrystallized status-structure composed of achieved and ascribed rankings, little if any inconsistency- effects result. Finally, considering the paired achieved-ascribed differences among the structurally crystallized, the sug- gestion would be that: 86 . with the fixing of achievement and the disappearance of the saliency differential, the entire status-set acquires a more visible, global evaluation, and the various contingen- cies or interdependencies among ascription and achievement are acknowledged (Smith, 1969: 914). , Comparing income and education of Catholics above age forty-five, one would expect, then, inconsistency effects with increasing crystallization. The data sup— port this prediction (+19). Methodological Challenges In general, the data so far seem to support Smith's (1969) hypotheses. However, the notion of status-consistency problems has been undergoing a rather radical evolution from its original conceptions, methodo- logically as well as theoretically. One of the conclu- sions reached is that: A . . only on a priori grounds can one unequivocally allocate variation to either the effects of statuses taken independently or to their particular combination. Some headway can be made in untangling the 'inde- pendent' effects of status variables from the effects of their particular configurations at the cost of more rigorous specification of the problem. We have seen that the only meaning that 'status inconsistency effects' can have which is distinct from the meaning of 'status effects' is that the relation between status and attitudes of behavior depends upon (varies with) the conjunction of particular levels on the several dimen- sions of social status (Laumann, 1970:519). 87. This would indicate that there is a need for more rigorous theoretical specificationdf the kinds of inter- action between status-varjables and behavior that would count as statuseinconsistency effects. (Accordingly, we have to make predictions, based on different sets of assumptions. In other words we have to be able to pre- dict, from the models developed earlier, which of the two inconsistent combinations in our data has the greater effect on the dependent variable. In looking at Table l earlier, we used a "differ- ence of differences" between the sums of the inconsistent and consistent cells. A positive response was seen to indicate inconsistency-effects and a negative result (or no result at all) was seen to indicate a lack of inconsistency-effects. A methodological dilemma is uncovered here because although it is true that if a positive result is obtained one probably has some kind of inconsistency-effect, it is not necessarily true that if one fails to get a positive number one does not have an inconsistency-effect. This may be illustrated by looking at one of the inconsistent cells. Among those who are "high income and low education" whose indicator is voting Democratic, one of the responses may be, "I make pretty good money, but 88 they still won't accept me. Obviously the system needs changing." This seems to be a radicalizing response and may be related to an identification with low-prestige ethnic groups. However, another response might be that "the system is pretty good and considering where I started from, I've done very well, in terms of my income." This might be considered a system-supporting response. If a high income-low education cell is seen as inconsistent but the two responses are considered equally probable since some of the people in that cell will go in one direction while the rest will go in the other direc- tion, then the results may cancel each other. This is a crucial problem in using the "difference of differences" test for inconsistency-effects. If it works out that one obtains a positive response, results may be in the predicted direction. If one does not get a positive response (or any response at all), one may still be getting hidden inconsistency-effects. This problem has to do with Blalock's "identification" problem: to be able to solve questions, one must have enough informa- tion. In this situation one has more unknowns than rela- tionships (Blalock, 1965; 1966; 1967). The only way to get around this identification-problem, in the absence 89 of information in data, is to make some reasonably plausible assumptions. Alternative "Ascription" . Assumption (Model 1) The first assumption is that for many purposes education can be treated as an ascribed-status. This assumption seems most plausible after a certain age when advancing one's education becomes less realistic. Continuing with the development of assumptions, if one is to be concerned with which way the inconsistent individual will respond, radicalizing or system- maintaining, one must make assumptions regarding which of these ranks he will pay most attention to in terms of its effect on self-image. This is based on the idea that when inconsistency effects are observed, they emerge because the individual experiences a "self“ problem of some kind (Mead, 1934; Kolb, 1944; Meltzer, 1964). The individual feels dissatisfied with self-evaluation in terms of others' reaction to him. If one is low in some rank and it has nothing to do with how he feels, his self-image, then it should not affect him. If this kind of assumption is not included, then everyone can be considered rank- inconsistent to some degree. This leads to the question of which of these ranks affects self-image, and which is 90 an accident or can be rationalized as an unfortunate cir- cumstance. An assumption is needed concerning which of the two ranks, income or education, will be considered more salient to the individual. One assumption might be that the most salient rank is always the high state of an ascribed status. If one has available a high state of an ascribed status, then he may pay a lot of attention to it because he considers it most salient to his self-image. The least salient rank is a low state of an ascribed status. Achieved characteristics will fall somewhere in between in terms of salience. Where the individual is low in income but high in education, he is likely to suggest that although it is unfortunate that he has so little income, he ended up fairly well as far as his education is concerned. This circumstance is likely to lead to a lowering of the Demo- cratic vote. On the other hand, a person who has high income and low education, because he does not have a high ascribed-status, should be upset at the lack of social acceptance of his high achieved-status. This response would most likely lead to greater Democratic voting. If these last two assumptions are correct, then there is less likelihood that the two responses will can- cel each other out. If the salient status has a low 91 rank, then the individual is more likely to respond with a radicalizing approach and less likely to support the system. If the salient status has a high rank he is more likely to support the system maintaining response and is less likely to be radicalized. With these assumptions, then, one may be able to overcome the indeterminance related to the "identification" problem. One should be able to predict which of these ranks the individual cen- ters his self image around, and in what direction the results are likely to move. Alternative "Achievement" Assumption (Model 2) Laumann (1970) has commented on the relative significance of income and education on inconsistency- effects, and following Homans, supports an alternative assumption (Model 2). These two distinct types of status incon- sistency need hardly have the same conse- quences for behavior. The person who gains little income relative to his educational standing is likely to be depressed, frus- trated, or otherwise disenchanted by the low rate of return upon his education investment, while the person who gains much income despite an inferior level of educational attainment is unlikely to experience similar feelings of psychological malaise. Thus, once a distinction between the two types of status inconsistency is introduced, one finds ample grounds for postulating a positive cor- relation between status inconsistency and 92 (the dependent variable) when income is low relative to education and a negative correla- tion between status inconsistency and (the dependent variable) when incOme is high rela- tive to education (Laumann, 1970:517). This alternative set of assumptions reasons that in an achievement-oriented society, the most salient rank is always the high state of an achieved-status. At the same time the least salient rank is always the low state of an achieved status. The two ascribed statuses should fall somewhere between the high and low states of the achieved statuses. The results that follow this alterna- tive set of assumptions are opposite those that follow from the first set of assumptions. While the first model suggests that high ascribed-status would have the greatest salience for the individual's self-image, the second model predicts that high achieved-status would have the greatest salience for the individual in an incon- sistent situation. One might predict, then, from our previous dis- cussions, that Model 1 which emphasizes ascribed statuses would be most appropriate of Catholics, andModel 2 which emphasizes achieved statuses would be most appropriate to Protestants. One can infer from this that the incon- sistent cell of high income-low education would be most salient for Catholics and that the inconsistent cell of ' 93 high education-low incOme would be most salient for Protestants, in terms of inconsistency-effects. Analysis of Data The interpretation of the data in the Table stems from our earlier discussion where it was proposed that: the determination as to whether rank inconsistency, per se, as distinct from the two hierarchies from which it is generated, has an effect upon a dependent variable, requires information as to the direction of any association existing between each of these hierarchies and the dependent vari- able. Once this information is at hand a judgment can be made as to whether the data depart from what would be predicted merely on the basis of knowledge about the association of each of these hierarchies with the dependent variable (Hyman, 1967:386). For example, suppose that the rank-hierarchies of education and income are each positively associated with a dependent variable d. If education-income incon- sistency has no effect on d, then the value of d for education-income inconsistents should be somewhere between the value of d for those high on both education and income and those low on both education and income. Only if the data deviate from this null hypothesis can it be said that education-income rank-inconsistency has an effect upon d (Hyman, T967). 94 From the data in Table l, we see that for Protes- tants under age forty-five, the cell that contains those who have high education relative to income exhibits a score of 54 percent, or l percent above themean for the two consistent cells. Accordingly, this score, because it falls within the two consistent cells and close to the mean of these two consistent cells, does not substanti- ally contribute to the inconsistency-effects. We see that for Protestants under age forty-five, the cell in which income is low relative to education (54 percent), which should be the major contributor to "inconsistency effects" for model 2, does not substantiate the hypothe- sis. One might suggest, however, that for those under forty-five years of age, low income relative to educa- tion may be seen as temporary and not an indicator of blocked social mobility. How then can we explain the finding that the cell which contains individuals having high income relative to education has a score of 84 per- cent, or 32 percent above the mean for the two consistent cells? One might question this unexpected finding as a status effect of low education, rather than a status- inconsistency effect. However, the cell of low education relative to income is 28 percent above the consistently 95 low score, which should be the highest percentage if there is only a 7 percent spread from the high consistent cell to the low consistent cell, suggesting that for young Protestants, voting Democratic is not highly correlated with social status. Contrary to the hypotheses, it is obviously a combination of high income and low education that is determining the inconsistency-effects for this Table. If one suggests that high education relative to income might indicate blocked social mobility (at least in more crystal- lized structures--Smith, 1969), what might a high income relative to education indicate for young Protestants? The possibility of some aspect of visibility and blocked social acceptance is worth investigating. Contrary to our assumptions, low education relative to income may be more visible among the young than had been predicted. In other words, such an individual seems to have less control over the visibility of his low educational status and may not be acceptable to other high income-high education people with whom he may want to involve himself. What may be happening to this young Protestant may be part of some “rebuff process" (Zelditch and Anderson, 1966) related to the increasing autonomy of education as it relates to other facets of social life in America (Hauser, l970). 9.6 However, with Anderson and Zelditch: fWe suspect that conformity or, for that matter, resentment due to rebuff or anticipated rebuff would be hard to diagnose with con- ventional survey interviews (Anderson and Zelditch, l964: l24). For Protestants forty-five years and over, Model 2 would predict that if there are inconsistency- effects, they would be monitored in the high education- low income cell. However, one would eXpect that with advancing "structural crystallization", income-education disparities would become less of a problem as the sali- ence of achieved ranks decreases. What we discover is that both inconsistent cells are below the mean of the consistent cells which indicates that individuals in the inconsistent cells vote in a Republican direction (High income-low education: -l4 percent; high education-low income: -7 percent). The greatest deviation from the mean, however, is in the "high income-low education? cell, countering the prediction of Model 2. In addition, per- centages in both inconsistent cells fall between the percentages of the consistent cells, suggesting that no finconsistency effects" are operative. In other words, although achieved ranks do lose some salience with advancing age, remaining response-mechanisms seem to 9? shift from the left to the right of the political spec- trum. This would be similar to what Wiley (l920) predicts: "People with inconsistent class (achieved) attributes are especially prone to support right wing groups" (Wiley, 1970:529). In addition, Lipset reviewed four major polls on McCarthyism and concluded that among manual workers, especially the Republican ones, support for McCarthy increased with income. He suggested that: Perhaps the higher income people with lower occupational or educational strata were pre- cisely those who were most drawn to an ideology that attacked as pro-Communist both liberal lower-class based politics, and moderate, conservative 'old upper class' elitist groups (Wiley, 1970:530). One might also include in this high income-low education group small businessmen, retirees and even some salaried middle-class who identify with the "old middle class." One must conclude from the results at hand that, for Protestants of all ages, "class" (achieved) oriented explanations in terms of Model 2 could not be verified, and that the alternative "status" (ascribed) oriented Model seemed to explain the data more effectively. How- ever, the precise dynamics of the process seem inde- terminant from the data available. 98 In regard to status-uncrystallized Catholics, one would expect that even though we are comparing achievement-oriented statuses, the factor of Catholicism would allow this to be perceived as an ascribed-achieved inconsistency. We would expect then, that inconsisten- cies between ascribed and achieved rankings will be stressful in proportion to the degree of overall crystal- lization. If inconsistency-effects are to be found, one would expect that the ”high income—low education" cell would predominate, according to Model 1. However, the two inconsistent cells for Catholics under age forty- five differ by l percent, and no overall inconsistency- effects are evident. Once again we have to ask for an explanation of the differences between young Protestants, who evince distinct inconsistency-effects in the "high income-low education" cell, and young Catholics who evince no inconsistency-effects while no difference in the two inconsistent cells was discovered. This indetermi- nancy, once again, cannot be enlightened by the present survey data. I Finally, with increasing structural crystalliza- tion, Catholics are expected to show greater inconsistency-effects. When we examine Catholics above age forty-five, and compare the two inconsistent cells, 99 it is the high income-low education cell that is 18 per— cent above the mean of the two consistent cells; the high education-low income cell is only 1 percent above the mean of the consistent cells. This conforms with our prediction in Model l. In addition, the high income-low education cell (at 85 percent) was 3 percent larger than the consistently low cell (at 82 percent), although the latter should have been the largest of the four cells if no inconsistency-effects were present. This indicates then, according to the criteria set down, that inconsistency-effects are present in the high income-low education cell. However, another methodological dilemma is inherent in the data. When education but not income, or income but not education, is associated with a dependent variable d, then another technique is recom- mended to determine the difference between status-effects and status-inconsistency effects. Hyman (l967) suggests that we: consider 'income high-education low' inconsistency where education but not wealth is associated with d. If this kind of incon- sistency has no effect on d, then we would predict that 'wealth high-education low' inconsistents will show the same d percent- age as will consistents with about the same educational rank. Thus we can compare the d percentage of the 'education-wealth con- sistently low' group with that of these inconsistents, and only if there is a ‘00 substantial difference will we Judge that a rank- -in¢onsisten¢y effect is present (Hyman, 1967: 387) ‘ However, we find only a 3 percent difference between the consistently low cell (at 82 percent) and the inconsistent high income-low education cell (at 85 per- cent), findings that cast grave doubts on the presence of inconsistency effects. Review and Conclusion Interestingly, Smith has predicted that for ascribed-achieved differences among the structurally crystallized, it would be the high education-low income cell that would register the inconsistency-effect. This is implied in his statement that a young Negro is not as likely to resign himself to the contingency of race and income as long as increases are possible or appear possible. For with advancing age, as his income becomes set, and unchanging in response to his investment in achievement, the contingency is likely to take on greater salience as a source of stress (Smith, 1969: 914). From this, one might expect the high education- low income cell to be the larger in this case, as the contingencies among statuses would be most apparent from a Homans' (1962) investment-rewards perSpective. What seems to occur, as we have seen, is some kind of reaction to low education on the part of older lOl Catholics, somewhat closer to the prediction of Model l. The implication is that there is increasing visibility of ascribed statuses with advancing age as well as decreasing ambiguity about status-relationships; further, that the results of this visibility are especially stressful for older Catholics with little education. However, i inconsistency-effects were not convincingly demonstrated for this group, which makes the results doubtful. ; Although aging seems to have different effects on Protestants and Catholics in terms of the relationship between status-inconsistency and voting behavior, when we examine the separate inconsistent cells, the literature based on available survey-data results and our own pre- dictions do not adequately explain the differences. Of the two models provided, the ascribed-status model Operates, for some unexplained reason, to the exclusion _of the achieved-class model. In addition, when incon- sistent effects were demonstrated, the only cell that pro- vided the inconsistent effect was the high income-low edu- cation cell and this only among an unpredicted group, young Protestants, contrary to eXpectations based on Homans' investment-reward exchange theory (Homans, l962). Obviously other theoretical and methodological perspectives will have to be developed to demonstrate lOZ inconsistency effects on some dependent variable. Judging from the results of our research to this point, and most others in the same tradition, it is probable that survey-data may not be a sufficiently powerful instrument to handle the extremely hairy and complex dilemmas involved in the study of status-inconsistency effects. CHAPTER IV SOCIAL PARTICIPATION AND STATUS INCONSISTENCY Introduction In the former chapter, when we considered status- inconsistency and voting behavior, a base-line model of the relation of different status-factors and voting behavior was indicated. The expectation was that with increasing wealth there is a tendency to vote Republican, and with decreasing wealth a greater likelihood to vote Democratic. Within an additive model one would posit differences in terms of particular status-inconsistent configurations. Status-inconsistency, especially the kind that pairs high-achieved and low-ascribed status, accord- ing to Lenski (l966; l967), would lead to more liberal voting behavior than would be predicted from the additive model. The theoretical assumptions were that the individual attempts to raise his lower ranks and if the attempts are unsuccessful, the individual may, in one kind Of response, try to change the system itself. 103 l04 Continuing to look at the research done in the "Lenski tradition" and trying to follow some of the moves in this research program, we turn to contact- frequencies of social participation as another kind of "observable response process" (Zelditch and Anderson, 1966). In order to build a base-line model Of social participation and its relations to social status, the direction Of the relationships would have to be pre- dicted. In other words, higher status should give a push in the direction of social participation as a measure Of social integration (Axelrod, l956). At the same time, lower social status should give an Opposite push, serving to prevent social participation. That is, the low-status individual with fewer financial resources, for instance, might restrict his involvement to one group, such as relatives (Dotson, l95l). It has been suggested that: . . the highly discrepant are not only likely to be politically left. In l956, Lenski published evidence that they are also less frequently committed to secular volun- tary organizations. Even among those who are members, there is an inordinately low rate of interpersonal relations and a strong tendency to token affiliation (Demerath, l965:l32). The implicit theoretical perspective, while simi- lar to role-theory used by Hughes (l944), suggests that the status-inconsistents 105 are unable to define their position and are uncomfortable under the prevailing. status-system. They seek to change the system and, in the meantime, they withdraw from voluntary organizations where status judgments are most blatant (Demerath, l965: l32). Jackson's "Isolation" Assumptions Jackson (1962) has improved on the withdrawal hypothesis by finding that among inconsistents whose ethnic status is high but whose economic status is low, political liberalism is supplanted by a high rate Of psychosomatic symptoms of stress (Jackson, 1962). He goes on to use a distinction based on the differences between "ascribed" and "achieved" statuses as first defined by Linton (1950): Ascribed statuses are those which are assigned to individuals without reference to their innate differences or abilities. They can be predicted and trained for, from the moment of birth. The achieved statuses are, as a mini— mum, those requiring special qualities, although they are not limited to these. They are not assigned to individuals from birth, but are left Often to be filled through compe- tition and individual effort (Jackson, l96l: 43). Using this distinction between ascribed and achieved statuses, Jackson suggests that . a deficiency in the former (ascribed statuses) is likely to be blamed against the society, whereas a deficiency in the latter (achieved statuses) is more likely to be 106 blamed upon oneself. Thus, low ascribed status Should conduce to political radical- ism, while low achieved status should result in intrapunitive symptoms such as psycho- somatic stress (and less social participa- tion) (Jackson, l962; cited in Demerath, 1965:l33). Jackson suggests that the rationale for this assumption relates to the differences in choice Of response to status-inconsistency. In terms Of how incon- sistents, who respond politically, differ from those who respond by withdrawing into social isolation, three factors are (l) (2) (3) evident: higher achieved status lower ascribed status different positions Of their achieved and ascribed ranks relative to one another. Thus, the first response to status-inconsistency is related to achievement status in two ways: A high achievement-inconsistent, because of his greater ability to see his problems, in a social context and to make an active cop- ing response to them, is likely to respond to his stress with attitudes and actions favoring changes in the social system. Con- versely, the inconsistent of low achieved- status is likely to make a more passive, less coping response, and this response is likely to be directed at himself because he is less able to see the social sources of his status dilemma (Jackson, l960:7l). lOZ According to Jackson, the second possible explana- tory factor is ascribed status. However, he tends to discount this factor by suggesting that It seems unlikely that the choice of an internal, passive, psychosomatic response to status-inconsistency is due principally to high (ascribed) status, while low (ascribed) status produces the more actively COping, socially directed, political response (Jack- son, 1960:71). The third explanatory factor may influence the response to status-inconsistency by affecting the way in which the inconsistent individual defines his problems. A person whose achievements ranks are inferior to his ascribed ranks is likely to define his difficulties in terms of personal failure: The success-values of American society tend to evaluate any low achiever as a failure, and the high racial-ethnic rank of this type of inconsistent increases the likelihood of feelings of personal deficiency and self blame. Unlike the low-status consistent, he cannot justify his lack of success in terms of his ascribed handicaps. He is therefore likely to experience feelings of guilt, increasing the likelihood of a response turned against the self. . . . This factor, then, is an additional force preventing low-achievement inconsistents from focusing upon the social system as the source of their difficulties (Jackson, 1960:72). ' ' On the other hand, the inconsistent whose achieve- ment ranks exceed his ascribed ranks is evaluated as a success: 108 . . . since he has won his way (or maintained his position) despite the handicap of his low ascribed-status. If he experiences stress due to conflicting expectations he is unlikely to blame himself because according to American success values his behavior has been completely laudatory. He is therefore more likely to see his problems as stemming from the actions of others, and hence the social system, thus predisposing him to respond by favoring social change (Jackson, l960:72). The suggestion in this third explanation, then, is that finding the problem in oneself is more likely to lead to withdrawal and isolation than finding the problem in society. Speculating on how this might be applied to levels of social participation, Jackson suggests that inconsistency-patterns vary in the extent to which the individual feels that his status-attributes can be made consistent through upward mobility. Such equilibration should be possible if the pattern is one of low stand- ing on achievement-dimensions and high standing on ascriptive-dimensions: According to Benoit-Smullyan's equilibration hypothesis, persons with this 'remediable' kind of status-inconsistency would be expected to try to equalize their status- ranks through upward mobility. Or such striving might at least take place in fantasy. In either case, such a response could be expected to reduce the likelihood of any other response, especially such a compara- tively unrewarding response as (withdrawal into isolation) (Jackson, 1960:16-17). l09 Jackson suggests then, that mobility effects would probably not continue far into middle age, ". since only the young men have any real possibility of upward mobility to motivate their actual or fantasied strivings" (Jackson, l960:l7). On the basis of these assumptions, one can pre- dict that among persons with inconsistent status-patterns remediable through upward mobility: . the younger person will report (higher levels Of social participation) than the older person because the younger persons are likely to be occupied with upward mobility strivings and thus are less likely to make another response (lower levels of social participation) (Jackson, l960:l7). The alternative assumption would be that among persons of inconsistent status-patterns not remediable by upward social mobility (such as high achieved-low ascribed status): . . we would expect, on these grounds at least, little difference between age groups in the level of (social participation) reported, since not even the younger people in these patterns can hope to equalize their status-ranks through upward mobility (Jackson, l960:l7). The arguments above outlined may be summarized in the following assumption: I. Where upward mobility is possible, an individual whose status-ranks are inconsistent llO will respond to the strains produced by attempting to equilibrate his status-ranks through upward mobility in preference to other forms of response. The following hypothesis derives from the above stated assumption: l. In a situation where upward mobility would have an equilibrating effect on patterns of status-inconsistency (high ascribed-low achieved), younger individuals will show a higher level of social participation, and Older persons will show a reduced level of social participation. Alternative "Insulating" Assumptions As in our earlier discussion of voting behavior, additional assumptions concerning which of the two ranks, income or education, is going to be considered salient to the individual, are needed for prediction of reSponses to status-inconsistency. A combination Of high income and low education might present a greater problem for Catholics than for Protestants, based on earlier dis- cussions in which we suggested a formulation where the most salient rank was always the high state of an ascribed rank. Where the salient rank is a low state, lll then the individual is more likely to respond by an " i risulatedf response (Zelditch and Anderson, 1966), or to i r|1:eract to agreater degree with others like himself, such as relatives. If the salient status has a high rank, he is 'l ilon< cuwumz zo—mm cowumuauu eewuaz.a>on< .mw_= cargo: sopam . :04 «soucH Ammpznwcupm magnum mg» cow mmwsouozuvu pco>opmc mzhv om- ANNV ems Amwv Nee a_w A_FV amm Acmv Rem so; sac: 304 so; mmucmcmmwwo coca cm>mwgu< co mucmcommwo smug: «mmv awn Aoev uom cepoeuau cau_o see we "aa< ANNV emu Aou_v Ram u._oepau ma cave: ”mm< 304 saw: cocueusem got: ;m_x msoucc cmsacgu< p=a=_cc< Lm>o E monHozmhmHmzoqu Hzmmmmmno mo mm>~k<4mm u~40=pmwgu< Lava: Assv “we Anew RN“ 304 sac: cm>mwcu< cm>o Aomv uom beapmauoca caupo use ma "wm< AFNPV new acuumouoea ma cave: ”mo< sac: coccausnw gap: msoucu pawspm$< all; umzoqu no Aam< toe mecc_ozpcouv mkzuzmhmmmzoozu zo~kHh<4mm quHumwcu< c0>0wgu< ucmapce< m0 mucmcmwcwo Lens: cm>o Amm< 20$ mcw_pocpcouv muH4oxkuzMHmHmzoozH onh