sign in sign up
<<
Home , Conflict theories, Randall Collins

Functional and Theories of Educational Stratification Author(s): Source: American Sociological Review, Vol. 36, No. 6 (Dec., 1971), pp. 1002-1019 Published by: American Sociological Association Stable URL: http://www.jstor.org/stable/2093761 Accessed: 02/06/2009 08:21

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=asa.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to American Sociological Review.

http://www.jstor.org 1002 AMERICAN SOCIOLOGICALREVIEW status." AmericanSociological Review 20 cation."American Journal of 69 (June) :317-325. (September):185-193. Labovitz,Sanford Schnore,Leo F. and David W. Varley 1970 "The assignmentof numbersto rank order 1955 "Some concomitantsof metropolitansize." categories."The AmericanSociological Re- American Sociological Review 20 (Au- view 35 (June):515-524. gust) :408-414. Lane, Angela Schulze,Robert 0. 1968 "Occupational mobility in six cities." 1958 "The of economic dominantsin com- American Sociological Review 33 (Octo- munity power structure."American Socio- ber):740-749. logicalReview 23 (February):3-9. Lasswell,Thomas E. Trow, Donald B. 1959 "Social class and size of community." 1967 "Status equilibrationin the laboratory." American Journal of Sociology 64 PacificSociological Review 10:75-80. (March):505-508. Veldman,Donald J. Pfautz, Harold W. and Otis Dudley Duncan 1967 Fortran Programmingfor the Behavioral 1950 "A critical evaluation of Warner's work . New York: Holt, Rinehart and in community stratification." American Winston. SociologicalReview 15 (April):205-215. Vidich,Arthur J. and Joseph Bensman Reiss,Albert J. 1958 Small Town in Mass . New York: 1959 "The sociological study of communities." Doubleday. Rural Sociology24 (June): 118-130. Warner,W. Lloyd, Wilfred C. Bailey, et al. Schnore,Leo F. 1949 in Jonesville. New York: 1963 "Some correlatesof urban size: A replica- Harper.

FUNCTIONAL AND OF EDUCATIONAL STRATIFICATION *

RANDALLCOLLINS Universityof California,San Diego

American Sociological Review 1971, Vol. 36 (December):1002-1019

Two theoriesare consideredin accountingfor the increasedschooling requiredfor employ- ment in advancedindustrial society: (a) a technical-functiontheory, stating that educational requirementsreflect the demandsfor greaterskills on the job due to technologicalchange; and (b) a conflict theory, stating that employment requirementsreflect the efforts of competing status groups to monopolize or dominate jobs by imposing their cultural standardson the selectionprocess. A review of the evidenceindicates that the conflict theory is more strongly supported. The main dynamic of rising educational requirementsin the United States has been primarilythe expansionof mobility opportunitiesthrough the school system, rather than autonomouschanges in the structure of employment.It is argued that the effort to build a comprehensivetheory of stratificationis best advanced by viewing those effects of technologicalchange on educational requirementsthat are substantiated within the basic context of a conflict theory of stratification.

EDUCATION has becomehighly important social mobility. This paper attempts to as- in occupational attainment in modern sess the adequacy of two theories in account- America, and thus occupies a central ing for available evidence on the link be- place in the analysis of stratification and of tween education and stratification: a func- tional theory concerning trends in technical * I am indebted to Joseph Ben-David, Bennett skill requirementsin industrial ; and Berger, , Margaret S. Gordon, a conflict theory derived from the approach Joseph R. Gusfield, Stanford M. Lyman, Martin Weber, stating the determinants of A. Trow, and Harold L. Wilensky for advice and of Max comment; and to MargaretS. Gordon for making various outcomes in the struggles among available data collected by the Institute of Indus- status groups. It will be argued that the trial Relations of the University of California at evidence best supports the conflict theory, Berkeley, under grants from the U. S. Office of although technical requirements have im- Education and U. S. Departmentof Labor. Their endorsement of the views expressed here is not portant effects in particular contexts. It will implied. be further argued that the construction of a EDUCATIONALSTRATIFICATION 1003 general theory of the determinants of strati- tainment after the completion of education fication in its varying forms is best advanced (Blau and Duncan, 1967:163-205; Eckland, by incorporating elements of the functional 1965; Sewell et al., 1969; Duncan and analysis of technical requirementsof specific Hodge, 1963; Lipset and Bendix, 1959:189- jobs at appropriate points within the con- 192). There are differences in occupational flict model. The conclusion offers an inter- attainment independent of social origins be- pretation of historical change in education tween the graduates of more prominent and and stratification in industrial America, and less prominent secondary schools, colleges, suggests where further evidence is required graduate schools, and law schools (Smigel, for more precise tests and for further de- 1964:39, 73-74, 117; Havemann and West, velopment of a comprehensive explanatory 1952:179-181; Ladinsky, 1967; Hargens theory. and Hagstrom, 1967). Educational requirementsfor employment The Importance of Education have become increasingly widespread, not A number of studies have shown that the only in occupations but also at the number of years of education is a strong de- bottom of the occupational hierarchy (see terminant of occupational achievement in Table 1). In a 1967 of the San America with social origins constant. They Francisco, Oakland, and San Jose areas also show that social origins affect educa- (Collins, 1969), 17%oof the employers sur- tional attainment, and also occupational at- veyed requiredat least a high school diploma

Table 1. Percent of Employers Requiring Various Minimum Educational Levels -of__Employees ,by Occupational Level. National Survey, 1937-38 Un- Semi- Cleri- Mana- Profes- skilled skilled Skilled cal gerial sional

Less than high school 99% 97% 89% 33% 32% 9% High school diploma 1 3 11 63 S4 16 Some college 1 2 23 College degree 3 12 52 100% 100% 100% 100% 100% 100%

San Francisco Bay Area, 1967 Less than high school 83% 76% 62% 29% 27% 10% High School diploma 16 24 28 68 14 4 Vocational training beyond high school 1 1 10 2 2 4 Some college 2 12 7 College degree 41 70 Graduate degree 3 5 100% 100% 100% 101% 99% 100% (244) (237) (245) (306) (288) (240) Sources: H.M. Bell,-Matching Youth and Jobs (Washington: American Council on Education, p. 264, as analyzed in Lawrence Thomas, The Occu- pational Structure and Education (Englewood Cliffs: Prentice-Hall, 1956) P. and Randall Collins, "Education and Employment," unpublished 346uPA.D. dissertation, University of California at Berkeley, 1969, Table III-1. Bell does not report the number of employers in the sample, but it was apparently large. 1004 AMERICAN SOCIOLOGICALREVIEW

for employmentin even unskilledpositions; 1 of technological change. Two processes are a national survey (Bell, 1940) in 1937-1938 involved: (a) the proportion of jobs requir- found a comparable figure of 1%. At the ing low skill decreases and the proportion same time, educational requirements appear requiring high skill increases; and (b) the to have become more specialized, with 38% same jobs are upgradedin skill requirements. of the organizationsin the 1967 survey which (2) Formal education provides the training, required college degrees of managers pre- either in specific skills or in general capaci- ferring business administrationtraining, and ties, necessary for the more highly skilled an additional 15%o preferring engineering jobs. (3) Therefore, educational require- training; such requirements appear to have ments for employment constantly rise, and been virtually unknown in the 1920s (Pier- increasingly larger proportions of the popu- son, 1959:34-54). At the same time, the lation are requiredto spend longer and longer proportions of the American population at- periods in school. tending schools through the completion of The technical-function theory of educa- high school and advanced levels have risen tion may be seen as a particular application sharply during the last century (Table 2). of a more general functional approach. The Careers are thus increasingly shaped within functional theory of stratification (Davis the educational system. and Moore, 1945) rests on the premises (A) that occupationalpositions requireparticular The Technical-Function Theory of Educa- kinds of skilled performance; and (B) that tion positions must be filled with persons who have either the native ability, or who have A common explanation of the importance acquired the training, necessary for the of education in modern society may be performanceof the given occupational role.2 termed the technical-function theory. Its basic propositions, found in a number of 2The concern here is with these basic premises sources (see, for example, B. Clark, 1962; rather than with the theory elaborated by Davis Kerr et al., 1960), may be stated as follows: and Moore to account for the universality of (1) the skill requirements of jobs in in- stratification.This theory involves a few further propositions: (C) in any particular form of so- dustrial society constantly increase because ciety certain occupational positions are function- ally most central to the operation of the social 1 This survey covered 309 establishmentswith system; (D) the ability to fill these positions,and/ 100 or more employees, representing all major or the motivation to acquirethe necessarytraining, industry groups. is unequallydistributed in the population; (E) in-

Table 2. Percentage Educational Attainment in the United States, 1869-1965.

B.A.'s or M.A.'s or High School Resident lst prof. 2nd prof. Ph.D.'s graduates/ college degrees/ degrees/ 1/10 of pop. 17 yrs. students/ 1/10 of pop. 1/10 of pop. pop. Period old pop. 18-21 15-24 25-34 25-34

1869-1870 2.0 1.7 1879-1880 2.5 2.7 1889-1890 3.5 3.0 1899-1900 6.4 4.0 1.66 0.12 0.03 1909-1910 8.8 S.1 1.85 0.13 0.02 1919-1920 16.8 8.9 2.33 0.24 0.03 1929-1930 29.0 12.4 4.90 0.78 0.12 1939-1940 50.8 15.6 7.05 1.24 0.15 1949-1950 59.0 29.6 17.66 2.43 0.27 1959-1960 65.1 34.9 17.72 3.25 0.42 1963 76.3 38.0 1965 19.71 5.02 0.73 Sources: Historical Statistics of the United States, Series A-28-29, H 327- 338; Statistical Abstract of the United States 1966, Tables 3 and. 194; Digest of Educational Statistics (U. S. Office of Education, 1967), Tables 66 and 88. EDUCATIONAL STRATIFICATION 1005 The technical-function theory of education The only available evidence on this point may be viewed as a subtype of this form of consists of data collected by the U. S. De- analysis, since it shares the premises that partment of Labor in 1950 and 1960, which the occupational structure creates demands indicate the amount of change in skill re- for particular kinds of performance,and that quirements of specific jobs. Under the most training is one way of filling these demands. plausible assumptions as to the skills pro- In addition, it includes the more restrictive vided by various levels of education, it ap- premises (1 and 2 above) concerning the pears that the educational level of the U. S. way in which skill requirements of jobs labor force has changed in excess of that change with industrialization,and concerning which is necessary to keep up with skill re- the content of school experiences. quirements of jobs (Berg, 1970:38-60). The technical-function theory of educa- Over-education for available jobs is found tion may be tested by reviewing the evidence particularly among males who have gradu- for each of its propositions (la, lb, and 2).3 ated from college and females with high As will be seen, these propositions do not school degrees or some college, and appears adequately account for the evidence. In order to have increased between 1950 and 1960. to generate a more complete explanation, it Proposition (2): Formal education provides will be necessary to examine the evidence for required job skills. This proposition may be the underlying functional propositions, (A) tested in two ways: (a) Are better educated and (B). This analysis leads to a focus on employees more productive than less edu- the processesof stratification-notably group cated employees? (b) Are vocational skills conflict-not expressed in the functional learned in schools, or elsewhere? theory, and to the formalizationof a conflict (a) Are better educated employees more theory to account for the evidence. productive? The evidence most often cited Proposition (la): Educational requirements for the productive effects of education is of jobs in industrial society increase because indirect, consisting of relationships between the proportion of jobs requiring low skill aggregate levels of education in a society and decreases and the proportion requiring high its overall economic productivity. These are skill increases. Available evidence suggests of three types: that this process accounts for only a minor (i) The nationalgrowth approach involves part of educational upgrading, at least in a calculatingthe proportionof growth in the society that has passed the point of initial U. S. Gross National Productattributable to industrialization. Fifteen percent of the in- conventional inputs of capital and labor; these leave a large residual, which is at- crease in education of the U. S. labor force tributedto improvementsin skill of the labor during the twentieth century may be at- force based on increasededucation (Schultz, tributed to shifts in the occupational struc- 1961; Denison, 1965). This approachsuffers ture-a decrease in the proportion of jobs from difficultyin clearlydistinguishing among with low skill requirementsand an increase technologicalchange affectingproductive ar- rangements,changes in the abilities of work- in proportionof jobs with high skill require- ers acquiredby experienceat work with new ments (Folger and Nam, 1964). The bulk technologies, and changes in skills due to of educational upgrading (857%) has oc- formal education and motivational factors curred within job categories. associated with a competitive or achieve- ment-orientedsociety. The assignmentof a Proposition (lb): Educational requirements large proportionof the residualcategory to educationis arbitrary.Denison (1965) makes of jobs in industrial society rise because the this attributionon the basis of the increased same jobs are upgradedin skill requirements. income to personswith higherlevels of edu- cation interpretedas rewardsfor their con- equalitiesof rewardsin wealth and prestige evolve tributions to productivity.Although it is a to ensure that the supply of personswith the nec- common assumptionin economic argument essary ability or trainingmeshes with the structure that wage returns reflect output value, wage of demands for skilled performance.The problems returnscannot be used to prove the produc- of stating functional centrality in empirical terms tive contributionof educationwithout circu- have been subjects of much debate. lar reasoning. 3 Proposition 3 is supportedby Tables 1 and 2. (ii) Correlationsof educationand level of The issue here is whether this can be explained economicdevelopment for nations show that by the previous propositionsand premises. the higherthe level of economicdevelopment 1006 AMERICAN SOCIOLOGICALREVIEW of a country,the higherthe proportionof its on the job or casually (Clark and Sloan, population in elementary, secondary, and 1966:73). Retraining for important techno- higher education (Harbison and Myers, 1964). Such correlationsbeg the question of logical changes in industry has been carried causality. There are considerablevariations out largely informally on-the-job; in only a in school enrollmentsamong countries at the very small proportion of jobs affected by same economic level, and many of these technological change is formal retraining in variationsare explicablein terms of political educational used (Collins, 1969: demandsfor access to education(Ben-David, 1963-64). Also, the overproductionof edu- 147-158; Bright, 1958). cated personnelin countrieswhose level of The relevance of education for nonmanual economic developmentcannot absorb them occupational skills is more difficult to evalu- suggeststhe demandfor educationneed not ate. Training in specific , such as come directly from the economy, and may run counter to economic needs (Hoselitz, medicine, engineering, scientific or scholarly 1965). research, teaching, and law can plausibly be (iii) Time-lagcorrelations of educationand considered vocationally relevant, and possi- economicdevelopment show that increasesin bly essential. Evidences comparingparticular the proportionof populationin elementary schoolprecede increases in economicdevelop- degrees of educationalsuccess with particular ment after a takeoff point at approximately kinds of occupationalperformance or success 30-50% of the 7-14 years old age-groupin are not available, except for a few occupa- school. Similaranticipations of economicde- tions. For engineers, high college grades and velopment are suggested for increases in degree levels generally predict high levels of secondaryand higher education enrollment, althoughthe data do not clearlysupport this technical responsibility and high participa- conclusion(Peaslee, 1969). A pattern of ad- tion in professional activities, but not neces- vances in secondaryschool enrollmentspre- sarily high salary or supervisory responsi- ceding advancesin economicdevelopment is bility (Perrucci and Perrucci, 1970). At the found only in a small numberof cases (12 of 37 examinedin Peaslee, 1969). A pattern same time, a number of practicing engineers of growthof universityenrollments and sub- lack college degrees (about 40% of engineers sequenteconomic development is found in 21 in the early 1950s; see Soderberg, 1963: of 37 cases, but the exceptions(including the 213), suggesting that even such highly tech- United States, France, Sweden, Russia, and the job. For Japan) are of such importanceas to throw nical skills may be acquired on serious doubt on any necessarycontribution academic research scientists, educational of higher education to economic develop- quality has little effect on subsequent pro- ment. The main contributionof educationto ductivity (Hagstrom and Hargens, 1968). economicproductivity, then, appearsto occur For other professions, evidence is not availa- at the level of the transitionto mass literacy, and not significantlybeyond this level. ble on the degree to which actual skills are learned in school rather than in practice. Direct evidence of the contributionof edu- In professions such as medicine and law, cation to individual productivity is sum- where education is a legal requirement for marizedby Berg (1970:85-104, 143-176). It admission to practice, a comparison group indicates that the better educated employees of noneducated practitionersis not available, are not generally more productive, and in at least in the modern era. some cases are less productive, among sam- Outside of the traditional learned profes- ples of factory workers, maintenance men, sions, the plausibility of the vocational im- department store clerks, technicians, secre- portance of education is more questionable. taries, bank tellers, engineers, industrial re- Comparisons of the efforts of different oc- search scientists, military personnel, and cupations to achieve "professionalization" federal civil service employers. suggest that setting educational requirements (b) Are vocational skills learned in school, and bolstering them through licensing laws or elsewhere? Specifically vocational educa- is a common tactic in raising an occupation's tion in the schools for manual positions is prestige and autonomy (Wilensky, 1964). virtually independent of job fate, as gradu- The result has been the proliferation of nu- ates of vocational programs are not more merous pseudo-professions in modern so- likely to be employed than high school drop- ciety; nevertheless these fail to achieve outs (Plunkett, 1960; Duncan, 1964). Most strong professional organizationthrough lack skilled manual workers acquire their skills of a monpolizable (and hence teachable) EDUCATIONAL STRATIFICATION 1007 skill base. Business administration schools processes of social organization It may be represent such an effort. (See Pierson, 1959: suggested that the "demands" of any oc- 9, 55-95, 140; Gordon and Howell, 1959: 1- cupational position are not fixed, but repre- 18, 40, 324-337). Descriptions of general, sent whatever behavior is settled upon in nonvocational education do not support the bargaining between the persons who fill image of schools as places where skills are the positions and those who attempt to con- widely learned. Scattered studies suggest that trol them. Individuals want jobs primarily the knowledge imparted in particular courses for the rewards to themselves in material is retained only in small part through the goods, power, and prestige. The amount of next few years (Learned and Wood, 1938: productive skill they must demonstrate to 28), and indicate a dominant student culture hold their positions depends on how much concerned with nonacademic interests or clients, customers, or employers can suc- with achieving grades with a minimum of cessfully demand of them, and this in turn learning (Coleman, 1961; Becker et al., depends on the balance of power between 1968). workers and their employers. The technical-function theory of educa- Employers tend to have quite imprecise tion, then, does not give an adequate ac- conceptions of the skill requirementsof most count of the evidence. Economic evidence jobs, and operate on a strategy of "satisfic- indicates no clear contributions of education ing" rather than optimizing-that is, setting to economic development, beyond the provi- average levels of performanceas satisfactory, sions of mass literacy. Shifts in the propor- and making changes in procedures or per- tions of more skilled and less skilled jobs sonnel only when performance falls notice- do not account for the observed increase in ably below minimum standards (Dill et al., education of the American labor force. Edu- 1962; March and Simon, 1958:140-141). cation is often irrelevant to on-the-job pro- Efforts to predict work performance by ob- ductivity and is sometimes counter-produc- jective tests have foundereddue to difficulties tive; specifically vocational training seems in measuring performance (except on spe- to be derived more from work experience cific mechanical tasks) and the lack of con- than from formal school training. The qual- trol groups to validate the tests (Anastasi, ity of schools themselves, and the nature of 1967). Organizations do not force their em- dominant student cultures suggest that ployees to work at maximumefficiency; there schooling is very inefficient as a means of is considerable insulation of workers at all training for work skills. levels from demands for full use of their skills and efforts. Informal controls over out- Functional and Conflict Perspectives put are found not only among production workers in manufacturing but also among It may be suggested that the inadequacies sales and clerical personnel (Roy, 1952; of the technical-function theory of education Blau, 1955; Lombard, 1955). The existence derive from a more basic source: the func- of informal organization at the managerial tional approach to stratification. A funda- level, the widespread existence of bureau- mental assumption is that there is a gen- cratic pathologies such as evasion of responsi- erally fixed set of positions, whose various bility, empire-building, and displacement of requirements the labor force must satisfy. means by ends ("red tape"), and the fact The fixed demand for skills of various types, that administrative work is only indirectly determinantof at any given time, is the basic related to the output of the organization, who will be selected for what positions. So- suggest that managers, too, are insulated cial change may then be explained by speci- from strong technological pressures for use fying how these functional demands change of technical skills. On all wherever with the process of modernization. In keep- levels, ing with the functional perspective in gen- informal organization exists, it appears that eral, the needs of society are seen as deter- standards of performance reflect the power mining the behavior and the rewards of the of the groups involved. individuals within it. In this light, it is possible to reinterpretthe However, this premise may be questioned body of evidence that ascriptive factors con- as an adequate picture of the fundamental tinue to be important in occupational success 1008 AMERICAN SOCIOLOGICALREVIEW even in advanced industrial society. The mobilization of particular minority groups social mobility data summarizedat the onset rather than by an increased economic need of this paper show that social origins have to select by achievement criteria. a direct effect on occupational success, even Goode (1967) has offereda modified func- after the completion of education. Both case tional model to account for these disparities: studies and cross-sectional samples amply that work groups always organize to pro- document widespread discrimination against tect their inept members from being judged Negroes. Case studies show that the opera- by outsiders' standards of productivity, and tion of ethnic and class standards in employ- that this self-protection is functional to the ment based not merely on skin color but on organizations, preventing a Hobbesian com- name, accent, style of dress, manners, and petitiveness and distrust of all against all. conversationalabilities (Noland and Bakke, This argument re-establishes a functional 1949; Turner, 1952; Taeuber et al., 1966; explanation, but only at the cost of under- Nosow, 1956). Cross-sectionalstudies, based mining the technological view of functional on both biographical and survey data, show requirements. Further, Goode's conclusions that approximately 60 to 70% of the Amer- can be put in other terms: it is to the ad- ican business elite come from upper-classand vantage of groups of employees to organize upper-middle-classfamilies, and fewer than so that they will not be judged by strict 15% from working-class families (Taussig performance standards; and it is at least and Joselyn, 1932:97; Warner and Abeg- minimally to the advantage of the employer glen, 1955:37-68; Newcomer, 1955:53; to let them do so, for if he presses them Bendix, 1956:198-253; Mills, 1963:110- harder he creates dissension and alienation. 139). These proportions are fairly constant Just how hard an employer can press his from the early 1800's through the 1950's. employees is not given in Goode's functional The business elite is overwhelminglyProtes- model. That is, his model has the disad- tant, male, and completely white, although vantage, common to functional analysis in there are some indications of a mild trend its most general form, of covering too many toward declining social origins and an in- alternative possibilities to provide testable crease of Catholics and Jews. Ethnic and explanations of specific outcomes. Functional class background have been found crucial analysis too easily operates as a justification for career advancement in the professions as for whatever particular pattern exists, as- well (Ladinsky, 1963; Hall, 1946). Sexual serting in effect that there is a proper reason stereotyping of jobs is extremely widespread for it to be so, but failing to the condi- (Collins, 1969:234-238). tions under which a particular pattern will In the traditional functionalist approach, hold rather than another. The technical ver- these forms of ascription are treated as re- sion of job requirementshas the advantage sidual categories: carry-overs from a less of specifying patterns, but it is this specific advanced period, or marks of the imperfec- form of functional explanation that is jet- tions of the functional mechanism of place- tisoned by a return to a more abstract func- ment. Yet available trend data suggest that tional analysis. the link between social class origins and oc- A second hypothesis may be suggested: cupational attainment has remained con- the power of "ascribed" groups may be the stant during the twentieth century in Amer- prime basis of selection in all organizations, ica (Blau and Duncan, 1967:81-113); the and technical skills are secondary considera- proportion of women in higher occupational tions depending on the balance of power. levels has changed little since the late nine- Education may thus be regarded as a mark teenth century (Epstein, 1970:7); and the of membershipin a particular group (possi- few available comparisons between elite bly at times its defining characteristic), not groups in traditional and modern societies a mark of technical skills or achievement. suggest comparable levels of mobility Educational requirements may thus reflect (Marsh, 1963). Declines in racial and ethnic the interests of whichever groups have power discrimination that appear to have occurred to set them. Weber (1968:1000) interpreted at periods in twentieth-centuryAmerica may educational requirements in bureaucracies, be plausibly explained as results of political drawing especially on the history of public EDUCATIONAL STRATIFICATION 1009 administration in Prussia, as the result of "breeding," "respectability," "propriety," efforts by university graduates to monopo- "cultivation," "good fellows," "plain folks," lize positions, raise their corporate status, etc. Thus the exclusion of persons who lack and thereby increase their own security and the ingroup culture is felt to be normatively power vis-h-vis both higher authorities and legitimated. clients. Gusfield (1958) has shown that edu- There is no a priori determination of the cational requirements in the British Civil number of status groups in a particular so- Service were set as the result of a power ciety, nor can the degree to which there is struggle between a victorious educated up- consensus on a rank order among them be per-middle-class and the traditional aristoc- stated in advance. These are not matters of racy. definition, but empirical variations, the To summarizethe argument to this point: causes of which are subjects of other devel- available evidence suggests that the techni- opments of the conflict theory of stratifica- cal-functional view of educational require- tion. Status groups should be regarded as ments for jobs leaves a large number of facts ideal types, without implication of neces- unexplained.Functional analysis on the more sarily distinct boundaries; the concepts re- abstract level does not provide a testable main useful even in the case where associa- explanation of which ascribed groups will tional groupings and their status cultures are be able to dominate which positions. To fluid and overlapping, as hypotheses about answer this question, one must leave the the conflicts among status groups may re- functional frame of reference and examine main fruitful even under these circumstances. the conditions of relative power of each Status groups may be derived from a num- group. ber of sources. Weber outlines three: (a) differences in life style based on economic A Conflict Theory of Stratification situation (i.e., class); (b) differences in life situation based on power position; (c) differ- The conditions under which educational ences in life situation deriving directly from requirementswill be set and changed may be cultural conditions or institutions, such as stated more generally, on the basis of a geographical origin, ethnicity, religion, edu- conflict theory of stratification derived from cation, or intellectual or aesthetic cultures. Weber (1968:926-939; see also Collins, Advantage. There is a con- 1968), and from advances in modern organi- B. Struggle for in society for various "goods" zation theory fitting the spirit of this ap- tinual struggle need proach. -wealth, power, or prestige. We no assumption that every individual A. Status groups. The basic units of so- make is motivated to maximize his rewards; how- ciety are associational groups sharing com- are inherently mon cultures (or ""). The core ever, since power and prestige wealth is often con- of such groups is families and friends, but scarce commodities, and of even a they may be etxended to religious, educa- tingent upon them, the ambition of for more than tional, or ethnic communities. In general, small proportion persons sets up an im- they comprises all persons who share a equal shares of these goods part of others sense of status equality based on participa- plicit counter-struggleon the and disesteem. Indi- tion in a commonculture: styles of language, to avoid subjection each but tastes in clothing and decor, manners and viduals may struggle with other, is derived primarily other ritual observances, conversational top- since individual identity in and be- ics and styles, opinions and values, and pre- from membership a , the cohesion of is a key ferences in sports, arts, and media. Participa- cause status groups the tion in such cultural groups gives individuals resource in the struggle against others, their fundamental sense of identity, espe- primary focus of struggle is between status cially in contrast with members of other as- groups rather than within them. sociational groups in whose everyday culture The struggle for wealth, power, and pres- they cannot participate comfortably. Sub- tige is carried out primarily through organi- jectively, status groups distinguish them- zations. There have been struggles through- selves from others in terms of categories of out history among organizations controlled moral evaluation such as "honor," "taste," by different status groups, for military con- 1010 AMERICAN SOCIOLOGICALREVIEW quest, business advantage, or cultural (e.g., Once groups of employees of different religious) hegemony, and intricate sorts of status groups are formed at various positions interorganizationalalliances are possible. In (middle, lower, or laterally differentiated) in the more complex societies, struggle between the organization, each of these groups may status groups is carried on in large part be expected to launch efforts to recruit more within organizations, as the status groups members of their own status group. This controlling an organization coerce, hire, or process is illustrated by conflicts among culturally manipulate others to carry out whites and blacks, Protestants and Catholics their wishes (as in, respectively, a conscript and Jews, Yankee, Irish and Italian, etc. army, a business, or a church). Organiza- found in American occupational life tional research shows that the success of (Hughes, 1949; Dalton, 1951). These con- organizationalelites in controlling their sub- flicts are based on ethnically or religiously ordinates is quite variable. Under particular founded status cultures; their intensity rises conditions, lower or middle members have and falls with processes increasing or de- considerable de facto power to avoid com- creasing the cultural distinctiveness of these pliance, and even to change the course of the groups, and with the succession of advan- organizations (see Etzioni, 1961). tages and disadvantagesset by previous out- This opposing power from below is comes of these struggles which determine the strengthened when subordinate members organizationalresources available for further constitute a cohesive status group of their struggle. Parallel processes of cultural con, own; it is weakened when subordinates ac- flict may be based on distinctive class as well quiesce in the values of the organization as ethnic cultures. elite. Coincidence of ethnic and class boun- C. Education As Status Culture.The main daries produces the sharpest cultural dis- activity of schools is to teach particular tinctions. Thus, Catholics of immigrant ori- status cultures, both in and outside the class- gins have been the bulwarks of informal room. In this light, any failure of schools to norms restricting work output in American impart technical knowledge (although it may firms run by WASPs, whereas Protestants also be successful in this) is not important; of native rural backgrounds are the main schools primarily teach vocabulary and in- "rate-busters" (O. Collins et al., 1946). flection, styles of dress, aesthetic tastes, val- Selection and manipulation of members in ues and manners. The emphasis on socia- terms of status groups is thus a key weapon bility and athletics found in many schools is in intraorganizationalstruggles. In general, not extraneous but may be at the core of the the organizationelite selects its new members status culture propagated by the schools. and key assistants from its own status group Where schools have a more academic or vo- and makes an effort to secure lower-level em- cational emphasis, this emphasis may itself ployees who are at least indoctrinated to be the content of a particular status culture, respect the cultural superiorityof their status providing sets of values, materials for con- culture.4 versation, and shared activities for an asso- ciational group making claims to a particular 4 It might be argued that the ethnic cultures basis for status. may differ in their functionality: that middle- con- class Protestant culture provides the self-discipline Insofar as a particular status group and other attributes necessary for higher organi- trols education, it may use it to foster con- zational positions in modern society. This version of functionaltheory is specific enough to be sub- cient managers are selected for survival. The oligo- ject to empirical test: are middle-class WASPs polistic situation in large-scale American business in fact better businessmenor governmentadminis- since the late 19th century does not seem to pro- trators than Italians, Irishmen, or Jews of patri- vide such a mechanism; nor does monial or working class cultural backgrounds? employment. Schumpeter (1951), the leading ex- Weber suggested that they were in the initial positor of the importance of managerial talent constructionof the capitalist economy within the in business, confined his emphasis to the formative confines of traditional society; he also argued period of business expansion, and regarded the that once the new economicsystem was established, large, oligopolistic corporation as an arena where the original ethic was no longer necessary to run advancement came to be based on skills in organi- it (Weber, 1930:180-183). Moreover, the func- zational (1951:122-124); these personalistic tional explanation also requires some feedback skills are arguably more characteristic of the patri- mechanismwhereby organizationswith more effi- monial cultures than of WASP culture. EDUCATIONAL STRATIFICATION 1011 trol within work organizations. Educational century, this rivalry was an important basis requirementsfor employment can serve both for the founding of large numbers of colleges to select new membersfor elite positions who in the U. S., and of the Catholic and Lu- share the elite culture and, at a lower level of theran school systems. The public school education, to hire lower and middle em- system in the U. S. was founded mainly ployees who have acquired a general respect under the impetus of WASP with the for these elite values and styles. purpose of teaching respect for Protestant and middle-class standards of cultural and Tests of the Conflict Theory of Educational religious propriety, especially in the face of Stratification Catholic, working-class immigration from Europe (Cremin, 1961; Curti, 1935). The The conflict theory in its general form is content of public school education has con- supported by evidence (1) that there are sisted especially of middle-class, WASP cul- distinctions among status group cultures- ture (Waller, 1932:15-131; Becker, 1961; based both on class and on ethnicity-in Hess and Torney, 1967). modern societies (Kahl, 1957:127-156, 184- At the elite level, private secondary schools 220); (2) that status groups tend to occupy for children of the WASP upper class were different occupational positions within orga- founded from the 1880s, when the mass in- nizations (see data on ascription cited doctrination function of the growing public above); and (3) that occupants of different schools made them unsuitable as means of organizationalpositions struggle over power maintaining cohesion of the elite culture it- (Dalton, 1959; Crozier, 1964). The more self (Baltzell, 1958:327-372). These elite specific tests called for here, however, are of schools produce a distinctive personality the adequacy of conflict theory to explain type, characterizedby adherenceto a distinc- the link between education and occupational tive set of upper-class values and manners stratification. Such tests may focus either (McArthur, 1955). The cultural role of on the proposed mechanism of occupational schools has been more closely studied in placement, or on the conditions for strong Britain (Bernstein, 1961; Weinberg, 1967), or weak links between education and occupa- and in France (Bourdieu and Passeron, tion. 1964), although Riesman and his colleagues Education As a Mechanism of Occupa- (Riesman, 1958; Jencks and Riesman, 1968) tional Placement. The mechanism proposed have shown some of the cultural differences is that employers use education to select among prestige levels of colleges and uni- persons who have been socialized into the versities in the United States. dominant status culture: for entrants to (b) Evidence that education has been their own managerial ranks, into elite cul- used as a means of cultural selection may be ture; for lower-level employees, into an at- found in several sources. Hollingshead's titude of respect for the dominant culture (1949:360-388) study of Elmtown school and the elite which carries it. This requires children, school dropouts, and community evidence that: (a) schools provide either attitudes toward them suggests that em- training for the elite culture, or respect for ployers use education as a means of selecting it; and (b) employers use education as a employees with middle-class attributes. A means of selection for cultural attributes. 1945-1946 survey of 240 employers in New (a) Historical and descriptive studies of Haven and Charlotte, N. C. indicated that schools support the generalization that they they regardededucation as a screeningdevice are places where particular status cultures for employees with desirable (middle-class) are acquired, either from the teachers, from character and demeanor; white-collar posi- other students, or both. Schools are usually founded by powerful or autonomous status tions particularly emphasized educational groups, either to provide an exclusive educa- selection because these employees were con- tion for their own children, or to propagate sidered most visible to outsiders (Noland respect for their cultural values. Until re- and Bakke, 1949:20-63). cently most schools were founded by re- A survey of employers in nationally prom- ligions, often in opposition to those founded inent corporations indicated that they re- by rival religions; throughout the 19th garded college degrees as important in hiring 1012 AMERICAN SOCIOLOGICALREVIEW potential managers, not because they were Variations in Linkage between Education thought to ensure technical skills, but rather and Occupation to indicate "motivation" and "social experi- ence" (Gordonand Howell, 1959:121). Busi- The conflict model may also be tested by ness school trainingis similarly regarded,less examining the cases in which it predicts edu- as evidence of necessary training (as em- cation will be relatively important or unim- ployers have been widely skeptical of the portant in occupational attainment. Educa- utility of this curriculumfor most positions) tion should be most important where two than as an indication that the college gradu- conditions hold simultaneously: (1) the ate is committed to business attitudes. Thus, type of education most closely reflects employers are more likely to refuse to hire membership in a particular status group, liberal arts graduates if they come from a and (2) that group controls employment college which has a business school than if in particular organizational contexts. Thus, their college is without a business school education will be most important where (Gordon and Howell, 1959:84-87; see also the fit is greatest between the culture Pierson, 1959:90-99). In the latter case, of the status groups emerging from schools, the students could be said not to have had and the status group doing the hiring; it will a choice; but when both business and liberal be least importantwhere there is the greatest arts courses are offered and the student disparity between the culture of the school chooses liberal arts, employers appear to and of the employers. take this as a rejection of business values. This fit between school-group culture and Finally, a 1967 survey of 309 California employer culture may be conceptualized as organizations (Collins, 1971) found that a continuum. The importance of elite educa- educational requirements for white-collar tion is highest where it is involved in selec- workers were highest in organizationswhich tion of new members of organizationalelites, placed the strongest emphasis on normative and should fade off where jobs are less elite control over their employees.s Normative (either lower level jobs in these organiza- control emphasis was indicated by (i) rela- tions, or jobs in other organizationsnot con- tive emphasis on the absence of police record trolled by the cultural elite). Similarly, for job applicants; (ii) relative emphasis on schools which produce the most elite gradu- a recordof job loyalty; (iii) Etzioni's (1961) ates will be most closely linked to elite oc- classification of organizationsinto those with cupations; schools whose products are less high normative control emphasis (financial, well socialized into elite culture are selected professional services, government, and other for jobs correspondingly less close to elite organizationallevels. public services organizations) and those with In the United States, the schools which remunerativecontrol emphasis (manufactur- produce culturally elite groups, either by ing, construction, and trade). These three virtue of explicit training or by selection of are highly interrelated, thus mutu- indicators students from elite backgrounds,or both, are their conceptualization as in- ally validating the private prep schools at the secondary of normative control emphasis. The dicators level; at the higher level, the elite colleges between normative control em- relationship (the Ivy league, and to a lesser degree the educational requirements holds phasis and major state universities); at the profes- requirementsand white-collar for managerial sional training level, those professional both including and requirements generally, schools attached to the elite colleges and and technical posi- excluding professional universities. At the secondary level, schools control emphasis does not tions. Normative which produce respectably socialized, non- education requirements. affect blue-collar elite persons are the public high schools (especially those in middle-class residential 5Sample consisted of approximately one-third areas); from the point of view of the culture of all organizationswith 100 or more employees of WASP employers, Catholic schools (and in the San Francisco, Oakland, and San Jose all-black schools) are less acceptable. At the metropolitan areas. See Gordon and Thal-Larsen (1969) for a descriptionof proceduresand other level of higher education, Catholic and black findings. colleges and professional schools are less EDUCATIONAL STRATIFICATION 1013 elite, and commercialtraining schools are the lective in this regard, choosing not only from least elite form of education. Ivy League law schools but from a group In the United States, the organizations whose background includes attendance at most clearly dominated by the WASP upper elite prep schools and colleges (Smigel, 1964: class are large, nationally organized business 39, 73-74, 117). There are also indications corporations,and the largest law firms (Dom- that graduates of ethnically-dominatedpro- hoff, 1967:38-62). Those organizationsmore fessional schools are most likely to practice likely to be dominated by members of mi- within the ethnic community; this is clearly nority ethnic cultures are the smaller and the case among black professionals. In gen- local businesses in manufacturing, construc- eral, the evidence that graduates of black tion, and retail trade; in legal practice, solo colleges (Sharp, 1970:64-67) and of Catho- rather than firm employment.In government lic colleges (Jencks and Riesman, 1968:357- employment, local governmentsappear to be 366) have attained lower occupational posi- more heavily dominated by ethnic groups, tions in business than graduates of white whereas particular branches of the national Protestant schools (at least until recent government (notably the State Department years) also bolsters this interpretation. and the Treasury) are dominated by WASP It is possible to interpret this evidence ac- elites (Domhoff, 1967: 84-114, 132-137). cording to the technical-function theory of Evidence on the fit between education and education, arguing that the elite schools employment is available for only some of provide the best technical training, and that these organizations. In a broad sample of the major national organizations require the organizational types (Collins, 1971) educa- greatest degree of technical talent. What is tional requirementswere higher in the bigger necessary is to test simultaneously for tech- organizations,which also tended to be orga- nical and status-conflict conditions. The most nized on a national scale, than in smaller direct evidence on this point is the California and more localistic organizations.6The find- employer study (Collins, 1971), which ex- ing of Perrucci and Perrucci (1970) that amined the effects of normative control em- upper-class social origins were important in phasis and of organizational prominence, career success precisely within the group of while holding constant the organization's engineerswho graduated from the most pres- technological modernity, as measured by the tigious engineering schools with the highest number of technological and organizational grades may also bear on this question; since changes in the previous six years. Techno- the big national corporationsare most likely logical change was found to affect educa- to hire this academically elite group, the tional requirementsat managerialand white- importanceof social origins within this group collar (but not blue-collar) levels, thus tends to corroborate the interpretation of giving some support to the technical-func- education as part of a process of elite cul- tion theory of education. The three variables tural selection in those organizations. -normative control emphasis, organizational Among lawyers, the predicted differences prominence, and technological change-each are clear: graduates of the law schools at- tached to elite colleges and universities are 7Similar processes may be found in other so- more likely to be employed in firms, whereas cieties, where the kinds of organizations linked to graduates of Catholic or commercial law particular types of schools may differ. In England, schools are more likely to be found in solo the elite "public schools" are linked especially to practice (Ladinsky, 1967). The elite Wall the higher levels of the national civil service (Wein- berg, 1967:139-143). In France, the elite Ecole Street law firms are most educationally se- Polytechnique is linked to both government and industrial administrative positions (Crozier, 1964: 238-244). In Germany, universities have been 8 Again, these relationshipshold for managerial linked principally with government administration, requirementsand white-collar requirements gen- and business executives are drawn from elsewhere erally, both including and excluding professional (Ben-David and Zloczower, 1962). Comparative and technical positions, but not for blue-collar analysis of the kinds of education of government requirements.Noland and Bakke (1949:78) also officials, business executives, and other groups in report that larger organizationshave higher educa- contexts where the status group links of schools tional requirements for administrative positions differ is a promising area for further tests of con- than smaller organizations. flict and technical-functional explanations. 1014 AMERICAN SOCIOLOGICALREVIEW independently affected educational require- Historical Change ments, in particular contexts. Technological requirements for change produced significantly higher educa- The rise in educational throughoutthe last century may tional requirementsonly in smaller, localistic employment the conflict theory, and organizations, and in organizational sectors be explained using elements of the technical-func- not emphasizing normative control. Organi- incorporating into it at appropriate points. zational prominence produced significantly tional theory principal dynamic has centered on higher educational requirementsin organiza- The in the supply of educated persons tions with low technological change, and in changes by the expansion of the school sys- sectors de-emphasizing normative control. caused tem, which was in turn shaped by three Normative control emphasis produced sig- nificantly higher educational requirements conditions: associated with in organizations with low technological (1) Education has been high economic and status position from the change, and in less prominent organizations. colonial period on through the twentieth Thus, technical and normative status condi- century. The result was a popular demand tions all affect educational requirements; for education as mobility opportunity. This measures of association indicated that the has not been for vocational educa- latter conditions were stronger in this sam- demand tion at a terminal or commerciallevel, short ple. of full university certification; the demand Other evidence bearing on this point con- has rather focused on education giving entry cerns business executives only. A study of the into the elite status culture, and usually only top executives in nationally prominent busi- those technically-orientedschools have pros- nesses indicated that the most highly edu- pered which have most closely associated cated managers were not found in the most themselves with the sequence of education rapidly developing companies, but rather in leading to (or from) the classical Bachelor's the least economically vigorous ones, with degree (Collins, 1969:68-70, 86-87, 89, 96- highest education found in the traditionalistic 101). financial and utility firms (Warner and (2) Political decentralization, separation Abegglen, 1955:141-143, 148). The business of church and state, and competition among elite has always been highly educated in rela- religious denominationshave made founding tion to the Americanpopulace, but education schools and colleges in America relatively seems to be a correlate of their social origins easy, and provided initial motivations of rather than the determinant of their success competition among communities and reli- (Mills, 1963:128; Taussig and Joslyn, 1932: gious groups that moved them to do so. As 200; Newcomer, 1955:76). Those members a result, education at all levels expanded of the business elite who entered its ranks faster in America than anywhere else in from lower social origins had less educa- the world. At the, time of the , tion than the businessmen of upper and there were nine colleges in the colonies; in upper-middle-classorigins, and those busi- all of Europe, with a population forty times nessmen who inherited their companies were that of America, there were approximately much more likely to be college educated than sixty colleges. By 1880 there were 811 Amer- those who achieved their positions by entre- ican colleges and universities; by 1966, there preneurship (Bendix, 1956:230; Newcomer, were 2,337. The United States not only 1955:80). began with the highest ratio of institutions indicates that edu- In general, the evidence of higher education to population in the cational requirementsfor employment reflect world, but increased this lead steadily, for employers' concerns for acquiring respecta- the number of European universities was not ble and well-socialized employees; their con- much greater by the twentieth century than cern for the provision of technical skills in the eighteenth (Ben-David and Zloczower, through education enters to a lesser degree. 1962). The higher the normative control concerns (3) Technical changes also entered into of the employer, and the more elite the the expansion of American education. As organization's status, the higher his educa- the evidence summarized above indicates: tional requirements. (a) mass literacy is crucial for beginnings of EDUCATIONAL STRATIFICATION 1015 full-scale industrialization,although demand degree was displacing the high school degree for literacy could not have been important as the minimal standard of respectability; in the expansion of education beyond ele- in the late 1960s, graduate school or special- mentary levels. More importantly, (b) there ized professional degrees were becoming is a mild trend toward the reduction in the necessary for initial entry to many middle- proportion of unskilled jobs and an increase class positions, and high school graduation in the promotion of highly skilled (profes- was becoming a standard for entry to sional and technical) jobs as industrialism manual laboring positions. Education has proceeds, accounting for 15%oof the shift thus gradually become part of the status in educationallevels in the twentieth century culture of classes far below the level of the (Folger and Nam, 1964). (c) Technological original business and professional elites. change also brings about some upgrading in The increasing supply of educated per- skill requirements of some continuing job sons (Table 2) has made education a rising positions, although the available evidence requirement of jobs (Table 1). Led by the (Berg, 1970:38-60) refers only to the dec- biggest and most prestigious organizations, ade 1950-1960. Nevertheless, as Wilensky employers have raised their educational re- (1964) points out, there is no "professionali- quirements to maintain both the relative zation of everyone," as most jobs do not prestige of their own managerial ranks and require considerable technical knowledge on the relative respectability of middle ranks.8 the order of that required of the engineer Education has become a legitimate standard or the research scientist. in terms of which employers select employ- The existence of a relatively small group ees, and employees compete with each other of experts in high-status positions, however, for promotion opportunities or for raised can have important effects on the structure prestige in their continuing positions. With of competition for mobility chances. In the the attainment of a mass (now approaching United States, where democratic decentrali- universal) higher education system in mod- zation favors the use of schools (as well as ern America, the ideal or image of technical government employment) as a kind of pa- skill becomes the legitimating culture in tronage for voter interests, the existence of terms of which the struggle for position even a small number of elite jobs fosters goes on. a demand for large-scale opportunities to Higher educational requirements,and the acquire these positions. We thus have a higher level of educational credentials of- "contest mobility" school system (Turner, fered by individuals competing for position 1960); it produced a widely educated popu- in organizations, have in turn increased the lace because of the many dropouts who demand for education by the populace. The never achieve the elite level of schooling at which expert skills and/or high cultural 8 It appears that employers may have raised status are acquired.In the process, the status their wage costs in the process. Their behavior is value of American education has become nevertheless plausible, in view of these considera- diluted. Standards of respectability are al- tions: (a) the thrust of organizational research ways relative to the existing range of cultural since Mayo and Barnard has indicated that ques- tions of internal organizational power and control, differences. Once higher levels of education of which cultural dominance is a main feature, become recognized as an objective mark of take precedence over purely economic considera- elite status, and a moderate level of educa- tions; (b) the large American corporations, which tion as a mark of respectable middle-level have led in educational requirements, have held positions of oligopolistic advantage since the late status, increases in the supply of educated 19th century, and thus could afford a large persons at given levels result in yet higher internal "welfare" cost of maintaining a well- levels, becoming recognized as superior, and socialized work force; (c) there are inter-organi- previously superior levels become only aver- zational wage differentials in local labor markets, corresponding to relative organizational prestige, age. and a "wage-escalator" process by which the Thus, before the end of the nineteenth wages of the leading organizations are gradually century, an elementary school or home edu- emulated by others according to their rank (Reynolds, 1951); a parallel structure of "educa- cation was no longer satisfactory for a mid- tional status escalators" could plausibly be expected dle-class gentleman; by the 1930s, a college to operate. 1016 AMERICAN SOCIOLOGICAL REVIEW interaction between formal job requirements background of employers varies in its fit and informal status cultures has resulted in with the educational culture of prospective a spiral in which educational requirements employees. Such analysis of "old school tie" and educational attainments become ever networks may also simultaneously test for higher. As the struggle for mass educational the independent effect of the technical re- opportunities enters new phases in the uni- quirements of different sorts of jobs on versities of today and perhaps in the gradu- the importance of education. Inter-nation ate schools of the future, we may expect a comparisons provide variations here in the further upgrading of educational require- fit between types of education and particu- ments for employment. The mobilization of lar kinds of jobs which may not be available demands by minority groups for mobility within any particular country. opportunities through schooling can only The full elaboration of such analysis contribute an extension of the prevailing would give a more precise answer to the pattern. historical question of assigning weight to various factors in the changing place of Conclusion education in the stratification of modern societies. At the same time, to state the It has been argued that conflict theory conditions under which status groups vary provides an explanation of the principal dy- in organizationalpower, including the power namics of rising educational requirements to emphasize or limit the importance of for employment in America. Changes in the technical skills, would be to state the basic technical requirements of jobs have caused elements of a comprehensive explanatory more limited changes in particular jobs. theory of the forms of stratification. The conditions of the interaction of these two determinants may be more closely REFERENCES studied. Anastasi, Anne Precise measures of changes in the actual 1967 ", psychologists, and psycho- technical skill requirements of jobs are as logical testing." American Psychologist 22 yet available only in rudimentary form. (April): 297-306. Few systematic studies show how much of Baltzell, E. Digby 1958 An American Business Aristocracy. New particular job skills may be learned in prac- York: Macmillan. tice, and how much must be acquired Becker, Howard S. through school background. Close studies of 1961 "Schools and systems of stratification." Pp. what is actually learned in school, and how 93-104 in A. H. Halsey, Jean Floud, and long it is retained, are rare. Organizational C. Arnold Anderson (eds), Education, Economy, and Society. New York: Free studies of how employers rate performance Press. and decide upon promotions give a picture Becker, Howard S., Blanche Geer, and Everett C. of relatively loose controls over the technical Hughes quality of employee performance, but this 1968 Making the Grade: The Academic Side of College Life. New York: Wiley. no doubt varies in particular types of jobs. Bell, H. M. The most central line of analysis for 1940 Matching Youth and Jobs. Washington: assessing the joint effects of status group American Council on Education. conflict and technical requirementsare those Ben-David, Joseph 1963- "Professions in the class systems of present- which compare the relative importance of 64 day Societies." Current Sociology 12:247- education in different contexts. One such 330. approach may take organization as the unit Ben-David, Joseph and Awraham Floczower of analysis, comparing the educational re- 1962 "Universities and academic systems in mod- quirements of organizations both to organi- ern societies." European Journal of So- ciology 31:45-85. zational technologies and to the status (in- Bendix, Reinhard cluding educational) background of organi- 1956 Work and Authority in Industry. New zational elites. Such analysis may also be York: Wiley. applied to surveys of individual mobility, Berg, Ivar comparing the effects of education on mo- 1970 Education and Jobs. New York: Praeger. Bernstein, Basil bility in different employment contexts, 1961 "Social class and linguistic development." where the status group (and educational) Pp. 288-314 in A. H. Halsey, Jean Floud, EDUCATIONAL STRATIFICATION 1017

and C. Arnold Anderson (eds), Education, Domhoff, G. William Economy, and Society. New York: Free 1967 Who Rules America? Englewood Cliffs, Press. New Jersey: Prentice-Hall. Blau, Peter M. Duncan, Beverly 1955 The Dynamics of Bureaucracy. Chicago: 1964 "Dropouts and the unemployed." Journal Press. of Political Economy 73 (April):121-134. Blau, Peter M. and Otis Dudley Duncan Duncan, Otis Dudley and Robert W. Hodge 1967 The American Occupational Structure. 1963 "Education and occupational mobility: A New York: Wiley. regression analysis." American Journal of Bourdieu, Pierre and Jean-Claude Passeron Sociology 68:629-644. 1964 Les Heritiers: Les Etudiants et la Culture. Eckland, Bruce K. Paris: Les Editions de Minuit. 1965 "Academic ability, higher education, and Bright, James R. occupational mobility." American So- 1958 "Does automation raise skill require- ciological Review 30:735-746. ments?" Harvard Business Review 36 Epstein, Cynthia Fuchs (July-August): 85-97. 1970 Woman's Place: Options and Limits in Clark, Burton R. Professional Careers. Berkeley: University 1962 Educating the Expert Society. San Fran- of California Press. cisco: Chandler. Etzioni, Amitai Clark, Harold F. and Harold S. Sloan 1961 A Comparative Analysis of Complex Or- 1966 Classrooms on Main Street. New York: ganizations. New York: Free Press. Teachers College Press. Folger, John K. and Charles B. Nam Coleman, James S. 1964 "Trends in education in relation to the 1961 The Adolescent Society. New York: Free occupational structure." Sociology of Edu- Press. cation 38:19-33. Collins, Orvis, Melville Dalton, and Donald Roy Goode, William J. 1946 "Restriction of output and social cleavage 1967 "The protection of the inept." American in industry." Applied Anthropology 5 Sociological Review 32:5-19. (Summer): 1-14. Gordon, Margaret S. and Margaret Thal-Larsen Collins, Randall 1969 Employer Policies in a Changing Labor 1968 "A comparative approach to political soci- Market. Berkeley: Institute of Industrial ology." Pp. 42-67 in Reinhard Bendix et Relations, University of California. al. (eds.), State and Society. Boston: Gordon, Robert A. and James E. Howell Little, Brown. 1959 Higher Education for Business. New York: 1969 Education and Employment. Unpublished Columbia University Press. Ph.D. dissertation, University of California Gusfield, Joseph R. at Berkeley. 1958 "Equalitarianism and bureaucratic re- 1971 "Educational requirements for employ- cruitment." Administrative Quar- ment: A comparative organizational terly 2 (March):521-541. study." Unpublished manuscript. Hagstrom, Warren 0. and Lowell L. Hargens Cremin, Lawrence A. 1968 "Mobility theory in the sociology of sci- 1961 The Transformation of the School. New ence." Paper delivered at Cornell Con- York: Knopf. ference on Human Mobility, Ithaca, N.Y. Crozier, Michel (October 31). 1964 The Bureaucratic Phenomenon. Chicago: Hall, Oswald University of Chicago Press. 1946 "The informal organization of the medi- Curti, Merle cal ." Canadian Journal of Eco- 1935 The Social Ideas of American Educators. nomic and Political Science 12 (February): New York: Scribners. 30-44. Dalton, Melville Harbison, Frederick and Charles A. Myers 1951 "Informal factors in career achievement." 1964 Education, Manpower, and Economic American Journal of Sociology 56 Growth. New York: McGraw-Hill. (March) :407-415. Hargens, Lowell and Warren 0. Hagstrom 1959 Men Who Manage. New York: Wiley. 1967 "Sponsored and contest mobility of Ameri- Davis, Kingsley and Wilbert Moore can academic scientists." Sociology of Edu- 1945 "Some principles of stratification." Ameri- cation 40:24-38. can Sociological Review 10:242-249. Havemann, Ernest and Patricia Salter West Denison, Edward F. 1952 They Went to College. New York: Har- 1965 "Education and economic productivity." court, Brace. Pp. 328-340 in Seymour Harris (ed.), Hess, Robert D. and Judith V. Torney Education and Public Policy. Berkeley: 1967 The Development of Political Attitudes in McCutchen. Children. Chicago: Aldine. Dill, William R., Thomas L. Hilton, and Walter R. Hollingshead, August B. Reitman 1949 Elmtown's Youth. New York: Wiley. 1962 The New Managers. Englewood Cliffs: Hoselitz, Bert F. Prentice-Hall. 1965 "Investment in education and its political 1018 AMERICAN SOCIOLOGICAL REVIEW

impact." Pp. 541-565 in James S. Coleman Pierson, Frank C. (ed.), Education and Political Develop- 1959 The Education of American Businessmen. ment. Princeton: Princeton University New York: McGraw-Hill. Press. Plunkett, M. Hughes, Everett C. 1960 "School and early work experience of 1949 "Queries concerning industry and society youth." Occupational Outlook Quarterly growing out of the study of ethnic relations 4:22-27. in industry." American Sociological Review Reynolds, Lloyd 14:211-220. 1951 The Structure of Labor Markets. New Jencks, Christopher and David Riesman York: Harper. 1968 The Academic Revolution. New York: Riesman, David Doubleday. 1958 Constraint and Variety in American Edu- Kahl, Joseph A. cation. New York: Doubleday. 1957 The American Class Structure. New York: Roy, Donald Rinehart. 1952 "Quota restriction and goldbricking in a Kerr, Clark, John T. Dunlop, Frederick H. Har- machine shop." American Journal of So- bison, and Charles A. Myers ciology 57 (March):427-442. 1960 Industrialism and Industrial Man. Cam- Schultz, Theodore W. bridge: Press. 1961 "Investment in human capital." American Ladinsky, Jack Economic Review 51 (March):1-16. 1963 "Careers of lawyers, law practice, and Schumpeter, Joseph legal institutions." American Sociological 1951 Imperialism and Social Classes. New York: Review 28 (February):47-54. Augustus M. Kelley. 1967 "Higher education and work achievement Sewell, William H., Archibald 0. Haller, and among lawyers." Sociological Quarterly 8 (Spring):222-232. 1969 "The educational and early occupational process." American Sociological Learned, W. S. and B. D. Wood attainment Review 34 (February) :82-92. 1938 The Student and His Knowledge. New York: Carnegie Foundation for the Ad- Sharp, Laure M. vancement of Teaching. 1970 Education and Employment: The Early of Graduates. Baltimore: Lipset, Seymour Martin and Reinhard Bendix Careers College Hopkins Press. 1959 Social Mobility in Industrial Society. Johns Berkeley: University of California Press. Smigel, Erwin 0. The Wall Street Lawyer. New York: Free Lombard, George F. 1964 1955 Behavior in a Selling Group. Cambridge: Press. Harvard University Press. Soderberg, C. Richard "The American engineer." Pp. 203-230 in March, James G. and Herbert A. Simon 1963 Kenneth S. Lynn, The Professions in Amer- 1958 Organizations. New York: Wiley. ica. Boston: Beacon Press. Marsh, Robert M. E. and Glen G. 1963 "Values, demand, and social mobility." Taeuber, Alma F., Karl Taeuber, American Sociological Review 28 (August): Cain assimilation and the com- 567-575. 1966 "Occupational petitive process: A reanalysis." American McArthur, C. Journal of Sociology 72:278-285. 1955 "Personality differences between middle Taussig, Frank W. and C. S. Joslyn and upper classes." Journal of Abnormal 1932 American Business Leaders. New York: and 50:247-254. Macmillan. Mills, C. Wright Turner, Ralph H. 1963 Power, Politics, and People. New York: 1952 "Foci of discrimination in the employment Press. Oxford University of nonwhites." American Journal of Soci- Newcomer, Mabel ology 58:247-256. 1955 The Big Business . New York: 1960 "Sponsored and contest mobility and the Columbia University Press. school system." American Sociological Re- Noland, E. William and E. Wight Bakke view 25 (October):855-867. 1949 Workers Wanted. New York: Harper. Waller, Willard Nosow, Sigmund 1932 The Sociology of Teaching. New York: 1956 "Labor distribution and the normative sys- Russell and Russell. tem." Social Forces 30:25-33. Warner, W. Lloyd and James C. Abegglen Peaslee, Alexander L. 1955 Occupational Mobility in American Busi- 1969 "Education's role in development." Eco- ness and Industry, 1928-1952. Minneapolis: nomic Development and Cultural Change University of Minnesota Press. 17 (April):293-318. Weber, Max Perrucci, Carolyn Cummings and Robert Perrucci 1930 The Protestant Ethic and the Spirit of 1970 "Social origins, educational contexts, and . New York: Scribner's. career mobility." American Sociological Re- 1968 Economy and Society. New York: Bed- view 35 (June):451-463, minster Press, MOBILITY AND FERTILITY 1019 Weinberg,Ian Wilensky,Harold L. 1967 The English Public Schools: the Sociology 1964 "The professionalization of everyone?" of Elite Education. New York: Atherton AmericanJournal of Sociology70 (Septem- Press. ber):137-158.

SOCIAL MOBILITY AND FERTILITY *

KEITH HOPE NuffieldCollege, Oxford, England

American Sociological Review 1971, Vol. 36 (December):1019-1032

In several recent studies the effects of mobility or status inconsistencyon a dependentvari- able have been quantifiedby means of an an additive model in which sets of constants have been fitted to two principlesof classification.In examininga particularapplication of this model, the followingpaper begins by suggestingthe possibilitythat the underlyinghypothesis may be more adequatelyrepresented by a symmetricalmodel which fits one and the same set of constantsto both principlesof classification. The second purposeof the paper is to show that, whether or not the symmetricalmodel is deemedto be the more appropriate,the basic hypothesiscan be adequatelytested only by the formulationof likely alternativesand the employment of tests which are specific to those alternatives. Thirdly, a considerationof two alternativesto the basic mode-one of which is simply a linear transformationof the other-implicitly demonstratesthat some of the problems (of multicollinearityor identification)which are associatedwith quantitativestudies of dif- ference variablessuch as inconsistency or mobility are analogous to the pseudo-problems generatedby the concept of rotation in factor analysis. The generalizationof the methods employed to more than two principlesof classification and to more than one dependentvariable is obvious.

Preamble** hypothesis," particularly to the form1 in which it their work on The AmericanOccupa- was advanced by R. A. Fisher in IN The Genetical Theory of Natural Selection. tional Structure Blau and Duncan (19- Various formulations of the hypothesis are 67) devote a number of pages to a dis- cited. It is claimed that the hypothesis is re- cussion of what they call "the mobility futed if the data exemplify a particular pat- tern, which they term "the additive hy- * This paper is one of a number of working pothesis." In this paper data which have papers prepared for the Oxford Social Mobility previously been held to satisfy the additive Project which is financedby the Social ScienceRe- search Council. This work will appear from time hypothesis are re-examinedto see whether in to time in volumes published by the Oxford Uni- fact they satisfy that hypothesis, either in versity Press under the generaltitle OxfordStudies its original form or in a modified form. in Social Mobility. ** This preamble grew out of comments and criticismson the following sections of the paper 1In consideringtheir argument,it is important which were made by Mrs. Jean Floud and Professor to note that the mobility which Blau and Duncan 0. D. Duncan. As a reward for my attack on his subject to empiricaltest is mobility of the present hypothesis, Professor Duncan has, with his usual generation. They make only passing reference to generosity,supplied me with data on which further the Galton-Fisherhypothesis of the inheritanceof studies of fertility and mobility may be carried (voluntary or involuntary) infertility, which is a out. Although we appear to disagree on several mechanismwhereby the mobility of an antecedent points, he and I are in entire agreementon the need generationmight affect the fertility of the following to replicate findings such as those reported here. generation. This restriction is apparent in their The additive hypothesis,in an approximate form, argument that if differential fertility were com- has already stood up to several replicationsand is pletely explained by social mobility then there to that extent on a surer footing than the mobility would be no differentialfertility by class among effect which I claim to detect, personswho do not change their class,