Class Origin, Class Destination, and Education: A Cross-National Study of Ten Industrial Nations

by Hiroshi Ishida, Walter Muller, John M. Ridge
Citation
Title:
Class Origin, Class Destination, and Education: A Cross-National Study of Ten Industrial Nations
Author:
Hiroshi Ishida, Walter Muller, John M. Ridge
Year: 
1995
Publication: 
The American Journal of Sociology
Volume: 
101
Issue: 
1
Start Page: 
145
End Page: 
193
Publisher: 
Language: 
English
URL: 
Select license: 
Select License
DOI: 
PMID: 
ISSN: 
Abstract:

 

Class Origin, Class Destination, and Education: A Cross-National Study of Ten Industrial ~ations'

Hiroshi Ishida

Columbia University

Walter Miiller

University of Mannheim

John M. Ridge

University of Oxford

This article examines three themes about the relationships among class origin, education, and class destination in 10 industrial na- tions: (1) differential access to education for different class origins,

(2) the allocation of class positions by education, and (3) the role of education in class reproduction and mobility. The patterns of association between class origin and education and between educa- tion and class destination are similar across the 10 nations. How- ever, the strength of these associations shows cross-national varia- tions. Class reproduction and mobility involve different social processes, which are differentially affected by education. However, a cross-national similarity emerges again in the way education me- diates the association between class origin and destination. The conclusion presents some implications of this analysis for the study of comparative macrosociology.

INTRODUCTION

Two diverging orientations to the study of social stratification and mobil- ity are apparent: social differentiation and mobility is either conceived as a vertical movement along status hierarchies or defined within the

An earlier version of this article was presented at the annual meetings of the American Sociological Association, Washington, D.C., August 1990. We thank the members of the CASMIN Project, the participants of the Nuffield Sociology Seminar, Oxford, in 1988, and the AJS reviewers for helpful comments. This research was in part supported by grants of the Stiftung Volkswagenwerk, Hannover, to the CASMIN project. The first author also acknowledges financial support of the Columbia University Council for Research in the Social Sciences. Correspondence may be addressed to Hiroshi Ishida, Department of Sociology, Columbia University, New York, New York 10027.

O 1995 by The University of Chicago. All rights reserved.

0002-9602196/10101-0004$01.50

AJS Volume 101 Number 1 (July 1995): 145-93 145 context of class structure (Goldthorpe 1987). The contrasting views have perhaps been most clearly formulated in the recent debate on the vertical or class-oriented approach to the study of social mobility (Hout and Hauser 1992; Sorensen 1992; Erikson and Goldthorpe 1992~). However, research on the role of education in the intergenerational transmission of advantage (e.g., Treiman and Terell 1975; Treiman and Yip 1989) does not seem to have advanced much beyond the refinements of the vertically oriented status attainment paradigm proposed by Blau and Duncan (1967). Our study focuses on the role of education in social mobility within the context of class structure. More specifically, we examine three related themes about the relationships among class origin, education, and class destination in industrial nations: (1) unequal access to education for different class origins, (2) the allocation of class positions by education, and (3) the role of education in class reproduction and mobility.

The first theme concerns the association between class origins and educational attainment. Important progress has been made in recent years in the empirical analysis of cross-cohort trends in educational at- tainment, in particular by Shavit and Blossfeld's (1993) volume, which includes individual studies of 13 countries with a wide range of politico- economic systems and types of educational institutions and policies (see also Featherman and Hauser 1978; Halsey, Heath, and Ridge 1980; Mare 1981; Smith and Cheung 1986; Raftery and Hout 1993). These studies overwhelmingly report a finding that is not consistent with the functional- ist expectation that the effect of particularistic criteria on educational and socioeconomic attainment declines in the course of industrial devel- opment (Parsons 1951; Levy 1966; Blau and Duncan 1967; Treiman 1970):~inequalities in educational transitions among children from differ- ent social origins have not diminished but rather have remained remark- ably stable since the early 20th century despite the increase in the level of industriali~ation.~

The findings appear to be more in line with the cultural capital theory, which predicts that the linkage between class origins and educational attainment is far from being diminished in industrial societies, because children from advantaged cultural backgrounds are already equipped with the linguistic and cultural competence neces- sary to succeed in school (Bourdieu 1973, 1974; Bourdieu and Passeron 1977; Bernstein 1977). However, this theory has not been tested directly in the Shavit and Blossfeld (1993) volume.

Only in the Netherlands (De Graaf and Ganzeboom 1993) and Sweden (Jonsson 1993) is clear evidence found of declining effects of social origins. Using a class-oriented framework, Miiller and Haun (1994) also find declining disadvantage for the working classes and the class of farmers in Germany that is not visible with the prestige measure used in the Shavit and Blossfeld (1993) volume. Simkus and Andorka (1982) also show that class differentials in the progression to completion of primary education have decreased across cohorts in Hungary; they ascribe these changes

The studies collected in Shavit and Blossfeld's (1993) volume, unlike earlier research, address the same research questions and employ a simi- lar analytical framework, allowing an overall comparison across nations. Their results, nonetheless, are not comparable in a strict sense, because the definition and measurement of educational outcomes and social ori- gins still vary among countries. This article, in contrast, examines the pattern of association between class origins and education in 10 industrial nations by ensuring comparable measurements of key variables and applying the same theoretically driven model to all nations. Furthermore, using the concept of inclusion and exclusion mechanisms, we identify the detailed pattern of association between class origins and education and assess the relative importance of these mechanisms within a society and across societies. The concept allows us to distinguish, for example, two forms of educational advantage of sons of the professional and manage- rial class: a better access to higher levels of education and an avoidance of the lowest levels.

Our approach to distinguishing inclusion and exclusion mechanisms, which are constructed to represent different social processes, is particu- larly effective in identifying the impact of various institutional forces and policy interventions on educational attainment. The typical policy approach in industrial societies has been to attempt to reduce the educa- tional disadvantages faced by the manual classes but not to change the educational participation of the upper classes. In this respect, the state socialist nations have probably made the most deliberate efforts to im- prove the educational prospects of the manual working class (Parkin 197 1; Simkus and Andorka 1982; Szczepanski 1978; Krymkowski 1991; Matgjs 1993). During the early period of constructing socialism, most of these countries adopted a policy of recruiting members of the manual working class into institutions of higher learning through a system of quotas and other preferential treatment (SzelCnyi and Aschaffenburg 1993; Heyns and Bialecki 1993). We therefore give special emphasis to assessing the differences in the relative importance of inclusion and exclu- sion mechanisms between capitalist and state socialist societies.

The second theme concerns the effect of education on allocation of class positions in industrial societies. The industrialism thesis (Kerr et al. 1960; Treiman 1970; Bell 1973; Hout 1989) predicts a strong correlation between achieved level of education and positions in the class structure: the higher the level of education, the more desirable and privileged the

mainly to effective social policies, not to the inherent logic of industrialism. However, SzelCnyi and Aschaffenburg (1993) report no noticeable trend of reduction in the effects of social origin on educational transition in Hungary.

positions. However, the detailed pattern of association between educa- tion and class positions is not as clear as the advocates of this thesis would claim. For example, highly qualified people are believed to have advantages in access to professional and managerial positions, but they may also be successful in avoiding manual positions. The distinction between these two forms of advantage, which we call inclusion and exclu- sion mechanisms, is often ignored. Similarly, the poorly qualified may be disadvantaged in two ways-excluded from professional and managerial positions and more likely to be recruited into manual positions. The detailed pattern is also likely to be influenced by a range of factors includ- ing the state's role in regulating the allocation of positions in the labor market and the strength and weakness of various classes in controlling the recruitment process (Parkin 1979).

The third theme discussed in this article is the role of education in intergenerational class reproduction and mobility. We bring together the relationship among class origin, class destination, and education by ad- dressing whether the patterns of class reproduction and mobility are me- diated by means of education. The status attainment approach provides a powerful tool for addressing this issue. The use of structural equation models enables researchers to decompose the zero-order correlation be- tween father's and son's occupational status into the direct effect and the indirect effect through education (Duncan 1975; Alwin and Hauser 1975). However, in this article we define mobility within the context of class structure rather than status hierarchy (Goldthorpe 1987). A different approach is required to address the role of education in class reproduction and mobility. Although a single measure (such as a path coefficient) is a useful summary, the association between class origin and class destina- tion is produced by different social processes, which may or may not involve the use of education. Previous studies report that the role of education in reproducing social relations of production depends upon the type of class relations being reproduced (Robinson and Kelly 1979; Kerckhoff, Campbell, and Trott 1982; Robinson 1984). The reproduction of ownership of the means of production takes place directly from one generation to the next without involving education, while the reproduc- tion of control over labor power is less direct and is in part facilitated by educational attainment. Similarly, different patterns of mobility are probably differentially affected by education (Griffin and Kalleberg 1981; Yamaguchi 1983; Erikson and Goldthorpe 1987a, 1987b).

We will therefore represent the association between class origin and class destination by different effects and assess the impact of education upon these effects. By disaggregating the association between origin and destination into different components, we identify more precisely the role of education in class reproduction and mobility.

DATA AND VARIABLES

Our data set comes from the CASMIN (Comparative Analysis of Social Mobility in Industrial Nations) project data file, which is the recoding of the unit-record data of the national surveys conducted in the 1970s in Europe.4 To the nine European countries in the original set we have added data on ~a~an.'

Therefore, our study includes the following 10 nations: England and Wales, France, Hungary, the Republic of Ireland, Japan, Northern Ireland, Poland, Scotland, Sweden, and (the former) West Germany. We believe that it is important to ensure that our sample refers only to those who have completed the process of acquiring aca- demic and vocational qualifications. Therefore, we have restricted our analysis to those 30-64 years old. We also restricted our analysis to men, because women's data were not available in all the nations.

We use three variables: class of origin, class of destination, and educa- tion. Class of origin is defined as the respondent's father's class at the respondent's age of about 14. Destination refers to the current class of the respondent. In both cases we use a six-category version of the Erik- son-Goldthorpe-Portocarero class schema (Erikson, Goldthorpe, and Por- tocarero 1979). Our six classes are: the professional-managerial or service class (1+11),6the routine nonmanual class (111),the petty bourgeoisie (IVab), the farming class (IVc +VIIb), the skilled manual working class (V+VI), and the nonskilled manual working class (vII~)' (see app. A for details of the class schema and table 1 for the distributions of class origin and class destination).

Education refers primarily to qualifications, that is, certificates of vo- cational and academic training. We use the terms qual$cation and education interchangeably. There is, of course, great difficulty in achieving a comparable classification across countries. We have relied upon the work of Konig, Luttinger, and Muller (1988), and in the case of Japan we have endeavored to keep as close as possible to the spirit of their

We are indebted to the members of the CASMIN project for their help in creating the data set and constructing class and education variables. For details of the data sets, see Erikson and Goldthorpe (19926, chap. 2).

We are indebted to Professor Ken'ichi Tominaga and the 1975 Social Stratification and Mobility (SSM) Committee for making available the 1975 Japanese SSM data set. For details of the data set, see Tominaga (1979). However, we alone are responsible for the recoding and reanalysis of the data.

The term service class, which we will use throughout this article, refers to the professional, managerial, and administrative occupations by virtue of their role in servicing or maintaining societal institutions (Goldthorpe 1982).

'For the details of the class schema, see Goldthorpe (1987, chap. 2) and Erikson and Goldthorpe (19926, chap. 2). See Ganzeboom, Luijkx, and Treiman (1989, pp. 12-13) for justification of collapsing the full 10-category version.

scheme. We have limited ourselves to three levels of qualification in order to avoid a large number of empty cells. Our lowest level refers to a "social minimum" of elementary education, which can include some basic vocational qualifications. Higher levels of vocational qualification, such as apprenticeship, are included in our middle level.8 It also includes intermediate academic or general qualifications up to preuniversity or "maturity" level (such as A-level in England or Abitur in West Ger- many). Our third level includes all tertiary qualifications (see app. B for details of the categories included in each level by country and table 1 for the distribution of qualifications).

MODELS OF ASSOCIATIONS

Basis of the Origin to Qualification (OQ) and Qualification to Destination (QD) Models

The models we propose are designed to represent simply, and as far as possible symmetrically, the complex relations of origin (0) to qualification

(Q) and qualification to destination (D) in all 10 of the countries for which we have data. We are concerned throughout with relative, not absolute, patterns: that is, we look at associations net of structural change and assume that these are properly represented by log odds-ratios. The mod- els are "topological" in character (Hauser 1978; Hout 1983), similar in methods of design and interpretation to those used by Erikson and Goldthorpe (1987a, 1987b, 1992a, 1992b).~

We develop models for the two-way association both of 0 with Q and of Q with D. For the OD association, we will use a variant of the model of social fluidity proposed by Erikson and Goldthorpe (1987a, 19876, 1992b). Each transition is modeled as a set of effects specified for the various cells of the OQ, QD, and OD transitions.1° In analogy with a

The scarcity of higher grade apprenticeship in France led us to classify,all vocational qualifications as middle level in our educational scale (see app. B).

We used this type of model because we believe (1) that it is important to disaggregate the pattern of association into different components and (2) that theoretical ideas (such as cultural capital and the functionalist theory of stratification) are most easily represented by different effect matrices. As one of the reviewers pointed out, our model constrains some effects to be the same and leaves three degrees of freedom, unlike the (saturated) continuation ratio model, which has no degrees of freedom. We do not employ the saturated model, because, as shown below, our model already captures over 98% of the association in the table and the saturated model lacks a clear theoretical rationale.

lo In contrast to the experience of Erikson and Goldthorpe (1987b, 19926) in fitting their core model of fluidity to the various national OD tables, we have not needed to incorporate specific national variants in our models of OQ and QD: the same design, with the same degrees of freedom, is fitted in each case.

TABLE 2

MODELOF ASSOCIATIONBETWEEN CLASS ORIGIN (0)AND QUALIFICATION

(Q)

CLASSORIGIN High Medium Low

I +I1 (service class) ............................................. SO1 ... SO2
I11 (routine nonmanual class) ..................................... ... ...
IVab (petty bourgeoisie) ......................................... ... M02P
IVc +VIIb (farming class) ........................................M01 ... MOZF
V+ VI (skilled working class) ................................. M01 ... MOZS
VIIa (unskilled working class) ................................. M01 ... M02U

NOTE.-For explanation of ellipses, see n. 11.

TABLE 3

MODELOF ASSOCIATION (Q) AND CLASS DESTINATION

BETWEEN QUALIFICATION (D)

CLASSDESTINATION

QUALIFICATION I +I1 I11 IVab IVc +VIIb V+VI VIIa

High .......... HQ1 ... ... HQ2 HQ2 HQ2
Medium ......... ... ... ... ... ...
Low ........... LQ1 ... LQZP LQZ F LQZS LQ2U

NOTE.-For explanation of ellipses, see n. 11.

path model of direct and indirect effects (Alwin and Hauser 1975), we then derive measures of the partial QD association in a model in which we control for the effects of origin. Similarly, we ask to what extent the gross or total OD effects, as modeled in the OD transition, are actually mediated through qualification. The models for the OQ and QD associa- tions are summarized in tables 2 and 3.l1 We now explain the theoretical rationale of the models before discussing the results of the fitting process.

Inclusion and Exclusion Mechanisms

We assume that service-class positions give enhanced access to and con- trol over resources, while manual positions do not. High qualifications

"The cells in tables 2 and 3 that are marked by ellipses indicate that our model does not postulate any effects. In other words, these cells will be determined by the mar- ginal distributions of the tables and form a neutral fluidity region. The parameter estimates of different effects, therefore, indicate the distance in log form from the neutral fluidity region. Comparing a parameter for a particular effect across nations implies comparing log distance from the neutral fluidity region. In other words, we do cross-national comparison using the neutral fluidity region as the base (for details of the parameterization, see Erikson and Goldthorpe 19923, pp. 182-84).

similarly are associated with enhanced access to and control over re- sources, while low qualifications are not. No special effects are assumed for intermediate class positions and qualifications. From this first princi- ple we can derive a simple model in which we represent service-class origin as conveying educational advantages in two forms. First, there is a tendency to include people from the service class in the highly qualified category (parameter Sol). Second, people from the service class tend to avoid or be excluded from the poorly qualified category (parameter S02). People of manual-class origin, in contrast, tend to be excluded from the highly qualified category (parameter M01) and included in the poorly qualified category (parameter M02).

The model of the transition from educational outcome to class position is similar. The ends of the main diagonal represent inclusion: the highly qualified tend to enter service-class positions (HQl), and the poorly quali- fied tend to enter manual positions (LQ2). The minor diagonal represents exclusion: the highly qualified avoid manual destinations (HQ2), and the poorly qualified tend to be excluded from service-class positions (LQ1). There are two aspects to the empirical evaluation of this model. First, of course, we aim to get an acceptable fit to the data for each country. But we also need to evaluate the parameter estimates themselves. In this way we can compare the strength of the different effects within countries and measure variations in the strength of particular effects across coun- tries. Therefore, in our model fitting we allow parameter estimates to vary both within and between countries.

Hypotheses on Policy Aims

In general, the goal of public policy in many countries has been to loosen the links between origin and qualification, while tightening those between qualification and destination. These expectations can be further specified.

Origin-quali$cation association (equality of opportunity).-The major obstacle to equality of educational opportunity probably comes from the resistance of those of service-class origin. From an "outflow" perspective we could say that the service class tries to ensure their sons an advantage in educational attainment by having them avoid low qualifications and have access to high qualifications. But it could also be argued that from an "inflow" perspective (that is, from the point of view of gatekeeping institutions) sons of the service class are better equipped with those "cul- tural traits" necessary for entry into higher levels of education (Bourdieu and Passeron 197 7; Bernstein 197 7). These advantages of service-class origin are captured by our SO1 and SO2 parameters.

Insofar as modern industrial states have tried to increase equality of educational opportunity, the typical policy approach has been to attempt to reduce t.he disadvantages of the manual classes in access to high quali- fications rather than directly to reduce the advantages of the service class. In socialist countries, deliberate efforts were made to create a LLpeople's intelligentsia" by promoting members of the manual classes into the bureaucracies, either directly, or indirectly through the educational sys- tem (Erikson and Goldthorpe 1987b, p. 156). If these efforts were effec- tive, we would expect that Hungary and, in particular, Poland, would show manual exclusion lower than in nonsocialist societies. To the extent that the deliberate inclusion of those of manual origins into high qualifi- cations-"positive discrimination" (Wesolowski and Mach 1986)overcompensates for disadvantages, manual exclusion could even be pos- itive. Similarly, the relative advantage of the service class may also be reduced, although social policies are not always effective in taking away privileges enjoyed by the advantaged nor, at times, are they intended to achieve such a reduction. Generally, we would expect manual exclusion (MO1) to be low in egalitarian societies; however, the relative advantages of the service class (Sol; inclusion) will not necessarily be reduced. Our strategy of disaggregating the OQ association into different components thus allows us to test these detailed hypotheses about the effects of social policies.

Different factors and mechanisms are responsible for the fact that edu- cational outcomes differ among classes. We must therefore distinguish the disadvantages of the farm sector and the petty bourgeoisie from those of the industrial working class. We hypothesize that sons of the farming class and the petty bourgeoisie are likely to be disadvantaged in financial resources and cultural traits (such as the "culture of independence"). Further, education may have a limited instrumental value, especially for those children who expect to follow in their father's footsteps; working on the farm or in the family business may have been viewed as a more efficient and rational investment for the future than going to school (Gam- betta 1987). We also expect that, at least in the period to which our data refer, the difficulty of organizing education in rural areas will have acted as an additional handicap to farm sons. These factors imply, in general, a high value for M02F (i.e., a relatively high chance of those from farm origins having only low-level qualifications).

The social conditions of the class of farmers, however, vary across societies, perhaps more so than for any other class. The composition of the agricultural classes also varies across societies as a consequence of institutional regulations (e.g., land ownership legislation) or agricultural policies. Therefore, the relative disadvantage of the farm sector may be greatly reduced either because the farm sector includes a relatively high proportion of prosperous families, as in England, or because policy efforts have been made to incorporate rural areas into a system of education leading to at least intermediate qualification levels, as in Sweden (Jonsson 1987, 1988, 1993). The likelihood that sons of the petty bourgeoisie have the lowest level of qualifications (M02P) is again dependent on the com- position of the petty bourgeois class and on social policies governing the entry to and maintenance of such a class in each society. We expect that the value of M02P will vary across our nations (Bechhofer and Elliott 1981, 1985).

We also recognize that it is necessary to divide the industrial working class into its skilled and unskilled components. The former is, as an origin, likely to reduce the chances of low qualifications, although we do not expect it to have any impact on chances of access to high qualifica- tions. We therefore divide the M02 parameter into four parts: M02F for the farm sector, M02P for the petty bourgeoisie, and one each for the skilled and unskilled industrial working class, M02S and M02U, respectively.

Quali$cation-destination association (exploitation of human capi-tal).-The parameters of the QD association are broadly consistent with a functionalist theory of stratification (Davis and Moore 1945).12 How- ever, our own interpretation of a high value for HQ1 (the inclusion of the highly qualified in the service class) would refer to the enhanced capacity of the highly qualified to gain what are generally regarded as desirable positions, as well as to the restriction of entry to those with "suitable," certificated skills (Collins 1979).

The tendency to include the poorly qualified in farming (LQ2F) will in part reflect the local standing of agricultural occupations: where farms are small and farmwork traditional and not yet technologically advanced, a low level of education will appear sufficient for farmwork. But the parameter will surely also capture career decisions and cultural prefer- ences that we have mentioned above as contributing to a high value of M02F: at least in some societies those with low qualifications will have come disproportionately from farm backgrounds, and the preference for low qualifications may well reflect an intention to work in agriculture later in life. In any case, we expect that in general LQ2F will have distinctively high values, confirming that entry to the farm sector does not require other than minimal qualifications.

We expect that the chances of inclusion of the poorly qualified into the unskilled working class (LQ2U) will be greater than into the skilled (LQ2S). This result reflects both the fact that skilled and technician jobs

l2 The following effects are identified within the context of the QD two-way table, but we believe that similar interpretations of the effects hold within the context of the three-way table. As will be apparent in the analysis section, QD effect parameters change little after controlling for class origins.

require technical qualifications and also the degree to which the unskilled working class forms a distinct category of relatively undesirable positions that the poorly qualified are least able to avoid.

We also include in the QD design a parameter, LQZP, that represents the peculiar association of low qualifications and petty bourgeois posi- tions. Our initial hypothesis is that at least in some countries (such as England and Japan) a relatively high proportion of petty bourgeois desti- nations are in fact proletarian in character and may therefore suggest a parameter close in value and interpretation to LQZU, discussed above. However, in other countries, it is claimed (e,g., Miiller et al. 1989), access to petty bourgeois positions is regulated by qualification mechanisms (craft apprenticeships and technical examinations) on a considerable scale. In these cases, the parameter may have a much reduced value, or even in extreme cases become negative, indicating that low qualifications are actually a barrier to petty bourgeois destinations.

The exclusion of the poorly qualified from service-class positions (LQ1) will presumably be greatest where the state controls access via a qualifi- cation mechanism. This exclusion we would expect to find most markedly in Poland and Hungary, of the countries in our sample. At the same time, the tendency to include the poorly qualified in manual positions (LQ2) should also be relatively high. In contrast, in.societies where there is in general relatively little emphasis on the acquisition of appropriate skills and training before entry to the job market, both LQ parameters will be low. In summary, we would expect these parameters to have a high value in societies that can be seen as "credentialist" (Collins 1979; Parkin 1979).

The tendency of those with high qualifications to avoid manual desti- nations (HQ2) is also open to alternative, perhaps complementary, inter- pretations. On the one hand, the highly qualified have, of course, wider occupational choices. On the other, this parameter may also reflect the state's wish to exploit to the maximum the investment it has made in its educational system: the highly qualified would in a human-capital sense be wasted if they were allowed to take manual occupations. We might therefore expect a high value for HQ2 in those countries where the state exercises closest control over the occupational allocation process.

ANALYSIS

Class Origin and Educational Attainment

We begin by examining the association between origin and qualification. Table 4 presents the estimates of OQ effect matrix parameters in the origin-qualification two-way table and the goodness-of-fit statistics for our 10 nations.13 Two general points are worth noting in table 4. First, for all 10 nations the G2 (S, 1,483) values-G2's standardized to the smallest sample size of 1,483-have a P-value greater than .05 (see rows 6 and 7).l4 Furthermore, the reduction in G2 over the independence model is greater than 98% and the number of cases misclassified by the model is less than or equal to 1% in all nations (see rows 4 and 5). These results suggest that our OQ effect matrices capture most of the association between origin and qualification. Second, the parameter estimates appear to vary cross-nationally. Eliminating any interaction effect between an effect matrix and nation (e.g., a cross-national variation in SO1 parame- ter) from a model that contains all the possible interactions between an effect matrix and nation in the three-way table of 0 by Q by nation was significant (e.g., deleting a term representing cross-national variation in SO1 effect yielded G2 = 34.36, df = 9, and P < .001).15 In other words, although the pattern of OQ association represented by our OQ matrices is cross-nationally similar, the extent or the strength of the effects varies across nations.

Let us next examine the detailed parameter estimates. l6 The advan- tages of service-class origin in educational attainment are generally

l3 The estimates of effect matrix parameters and the goodness-of-fit statistics are com-

puted by GLIM 3.77. The quantity 0.1 is added to zero cells in the three-way table. l4 The formula for standardized G2 is [(G2 -df)lN] X 1,483 + df (see Erikson and Goldthorpe 19876, p. 148). Because the sample size varies greatly across nations, we use G2 standardized by the sample size, rather than unstandardized G2, as our goodness-of-fit measure.

l5 We are grateful to Robert M. Hauser for suggesting this backward elimination method of evaluating cross-national differences in the estimates of effect matrix pa- rameters. Tests of cross-national variation in other effect matrices are: SO2 (G2 = 53.49; df = 9; P < .001), M01 (G2 = 27.92; df = 9; P < .001), M02U (G2 = 76.41; df = 9; P < .001), M02S (G2 = 41.59; df = 9; P < .001), M02F (G2 = 33.28; df = 9; P < .001), and M02P (G2 = 33.96; df = 9; P < ,001). It is possible to test whether all the effect matrices are uniformly higher or lower in one nation than in another, instead of testing the cross-national differences separately for different effect matrices (see, e.g., Yamaguchi 1987; Hout 1988, 1989; Wong 1990, 1992). Indeed, Xie (1992, pp. 387-89) shows how the log-multiplicative layer effect model may be applied to "topological" kinds of models. As can be seen from the parameter estimates in table 4, however, when any two nations are compared, estimates of some effect matrices are higher while others are lower in one country. It was our objective to disaggregate the pattern of association, 'and what we are concerned with is the cross- national variations in different effects, not uniform difference. We are grateful to Raymond Wong and Yu Xie for sharing their GLIM macro programs with us.

l6 All the statements concerning cross-national differences are based upon statistical testing in the 0 by Q by N three-way table context using backward elimination. For example, when we test the difference in the SO1 parameter between England and France, we eliminate the cross-national difference in this parameter betweekhe two nations from the model that contains all interaction parameters involving cross- national variations.

strong." The advantages, as we hypothesized, come from two sources: having better access to high qualifications (Sol) and avoiding low quali- fications (S02). These two mechanisms are compatible and operate in many nations, but the relative importance of these two mechanisms in producing service-class advantages varies across nations.

Note that the advantages of the service class are not particularly re- duced in the socialist nations. Strong SO1 and SO2 parameters in Hun- gary and Poland suggest that reducing the privileges enjoyed by the service class proved difficult even among socialist nations. In fact, ser- vice-class advantages appear to be more pronounced in these countries than in some other nations. l8

The skilled and nonskilled working classes face disadvantages in the attainment of education. Manual exclusion (MO1)-disadvantaged access to high qualifications for men of farm and manual working class origins-is present in England, France, Japan, Nothern Ireland, and Scotland. Unskilled inclusion (M02U)-the propensity for sons of the unskilled working class to have low or minimum qualifications-is present in all nations, whereas the same propensity for sons of the skilled working class (M02S) is found in all nations except Japan, Poland, and Scotland. These results suggest that the extent of working-class handi- caps depends upon their skill and nation. The nonskilled working class occupies a disadvantaged position in all 10 nations, while the skilled working class is more handicapped in some nations than others.

Our prediction that social policies introduced in socialist countries after World War I1 would reduce the disadvantages of the manual classes is not systematically supported by the analysis. However, in Hungary and particularly Poland the combination of the effects indicating manual dis- advantage is smaller than in many nations. Besides Ireland, Sweden, and West Germany, Poland and Hungary are the only countries for which the

l7 The only exception seems to be Northern Ireland; neither SO1 nor SO2 effects are significant at the .05 level of significance, although both coefficients have the predicted sign. The lack of a significant SO2 effect implies that the relative chances of ending in the lowest level of Q category are the same for sons of service class (I +II) and of routine nonmanual class (111) origins. What seems to be distinctive in Northern Ireland and also in Ireland is that sons of manual workers (skilled and nonskilled workers and farmers) and of the petty bourgeoisie have a relatively strong propensity to be allocated into the lowest Q level. Therefore, the difference in attainment of the lowest level of Q lies between sons of manual workers and the petty bourgeoisie on one hand and sons of white-collar workers on the other (see also Raftery and Hout (1993, table 1)for the relatively undifferentiated chances of primary education attainment between the service class and clerical and sales workers in Ireland).

l8 The SO1 parameter in Hungary is significantly higher than that in England, North- ern Ireland, Poland, and Scotland, and the SO2 parameter in Poland is significantly higher than that in all countries except Hungary and Scotland.

manual exclusion (M01) effect is not significant. However, in Ireland, Sweden, and West Germany, the lack of the M01 effect is counteracted by high working-class inclusion effects (M02U, M02S). In Poland M02S is not significant and M02U is significantly smaller than in any other country except Scotland. In Hungary the effects are at an intermediate level. Thus, taking all these effects together, Poland appears as the coun- try with the lowest level of working-class disadvantage. A social policy designed to enhance educational opportunities may have helped the working classes in Poland. For Hungary the evidence is less striking. For neither Poland nor Hungary do our results show any evidence of "positive discrimination," whereby sons of the manual classes would have better access to high qualifications and avoid low qualifications.

Sons of the farming class face disadvantage in educational achievement in all nations, and sons of the petty bourgeoisie are also handicapped in some nations. M02F parameters are consistently high and significant in our 10 nations. This handicap is generally stronger than that experienced by the nonskilled working class. In England, however, the value of M02F is the smallest among the nations and is smaller than that of M02U, probably because the farm sector includes a relatively high pro- portion of prosperous farmers. These results lead us to suspect that, while reducing working-class disadvantage is not easy, cultural and institu- tional barriers along with a lack of incentives to educational attainment may be even greater when it comes to the farm sector. We are not able to find evidence of the effect of social policies designed to enhance educa- tional opportunities for the farming sector in Sweden (Jonsson 1987, 1988, 1993). The propensity of the Swedish farming class to have low or minimum qualification (M02F) is actually a bit higher than that in most other countries. l9

Qualifications and Class Positions

Tables 5 and 6 present the estimates of our QD effect matrix parameters before and after controlling for class origins and the goodness-of-fit statis- tics for our 10 nations. In the qualification-destination two-way table (table 5), our model of QD effect matrices generally fits the data sets of our 10 nations. The standardized G2's and associated P-values show reasonable fits to all nations (see rows 6 and 7 of table 5). The fit is least satisfactory in the case of Ireland. But even here the model misclassifies

l9 Sweden probably has a high M02F value because the Swedish survey was con- ducted in 1974 and our sample did not include young men of age below 30. Improved schooling in the farm sector would perhaps be found if we had access to a more recent survey that included young people more likely to have benefited from the policy.

BY DESTINATION STATISTICS

IN THE QUALIFICATION

  England         Northern       West
  and Wales France Hungary Ireland Taoan Ireland Poland Scotland Sweden Germany
Goodness of fit:                    
G2 ...............................                    
df ................................                    
Pa ...............................                    
% misclassified ..............                    
% reduction in GZ ........                    
Gz(S, 1,483)' .................                    
P* ...............................                    
Effect matrix parameter:                    
HQ1 ............................                    

NOTE.-SES are given in parentheses. "For G2. Compared to the independence model. 'G2standardized to the smallest N (1,483); see n. 14 for a fuller explanation

For G2(S,1,483).

*P< .as. ** P < .01.

* *

sza~3s0s'nH

30momm

YZZS-?10 '""=4 '

,' g

V)

* *

!i

-%?:a

d-\~m-mm C)

d :,,,,,,3225-2

8

* B

.

5szggg "i: -2

?1?-?9-?

w -1 -d-2=mdc .

^+

.

* * a

3

aG%5-%G

,mQI10oID 3 91-?1'?1

I ---I -,,Nb '0.3

3

P

i

C)

***

%5-ks&s . . B

mmo.006 . .

. . .s

~~~Z":2,,,,

*** E

3

%C~S%G

03md-003 .

?-?tN.?-? o$

3----tbh~3C ~

3

* I I

$

.

-B

2

ms&G%G

:,,,,,,

3

N100.\Dbr

"322":

1

g

a

I

***

* -*

4 a

-*

10,oooos

rOm10mr-h

5

LI

-

-

2

m

G

,01033

+'-GG-G

m*10-0-0

;

C)

* .-LI

c

M s

d-5-%Gm? . .b 9

,,,3h\D

N3-N33 "i 'V1 8

. .

l'j3j I' z'=z='r-N '

,a,s

%

-C .

* * ;:2 *g

k5-105-& 9 9, -g3

m03m0.m

Oql?-?-$2 Ee

---mz","m r

"1

1

65 44

. . ....... ,gg 53

. . .......

. . . .......

. .

'i'iiii g3

. . . ....... 8;

. . . ....... .&2 .-s

. . ....... .

. . .......

. . . .......

. . " 42%

. . .

..iii'i'i $5

. . . .......

. . ....... % s2>,,

...........

2 ; i i i ''' 83 ,2&,-""

&-C'-OO

" . : :sv,a$ I{~2j~""

'' Y-m-hlNNN 4.cLIE~

'goao~~o~

g.gg6ag~~

3 3 3 zzsa--*

hZZdddd4

&?

only 1.2% of the cases and the reduction in C2 over the independence model is over 98%. Therefore, our QD effect matrices appear to be an adequate representation of the pattern of QD association in our 10 na- tions. As we found in the OQ transition, the strength of the QD effects varies across nations. The cross-national variations in all our QD effect matrices proved to be ~ignificant.~'

In order to obtain parameter estimates for QD effect matrices after controlling for class origin, we introduced OQ effect matrices and OD effect matrices (discussed in the next section) as controls in the three-way table. This model includes all three pairs of associations (QD, OQ, and OD) represented by effect mat rice^.^' The fit of this model for each soci- ety, which is shown in table 6, is good according to the standardized C2's (rows 6 and 7), the percentage of cases misclassified (row 4), and the percentage reduction in G2 from the baseline model of independence (row 5). The percentage reductions in the estimates of QD effect matrix parameters after controlling for class origin are shown in the bottom of table 6. They indicate the extent of difference between the gross and the net association of qualification and destination. The main feature emerg- ing from these numbers is that the reductions are generally small; qualifi- cations appear to play the role of allocating individuals to different class positions, regardless of their class origins.22

The only exception is LQ2F, the propensity for people with low quali- fications to move into the farming sector.23 Controlling for class origin reduces this propensity. The result suggests that sons born in the farming class are likely to have low qualifications and at the same time to stay in the farming sector; consequently, part of the association between low

20 Tests of cross-national variation in our QD effect matrices using backward elimina- tion are: HQ1 (G2 = 31.32; df = 9; P < .001), HQ2 (G2 = 21.99; df = 9; P < .01), LQ1 (G2 = 422.41; df = 9; P < .001), LQ2U (G2 = 150.20; df = 9; P < .001), LQ2S (G2 = 168.40; df = 9; P < .001), LQ2F (G2 = 128.42; df = 9; P < .001), and LQ2P (G2 = 112.54;df = 9; P < .001).

This is a model of no three-way interaction among qualification, destination, and origin. We believe that the assumption of no three-way interaction holds, because we can achieve a satisfactory fit to the data for each country without introducing a three-way interaction term. However, this assumption may not hold in the case of the U.S. data set (Hout 1988).

22 This finding is in broad agreement with some earlier studies (e.g., Treiman and Terell 1975; Meyer, Tuma, and Zagorski 1979) that examined the correlation between educa- tion and the son's occupational status after the father's status was controlled. However, the extent of reduction in the effect of eduction on class destination appears to be smaller than those reported in other studies that used extensive background variables (e.g., Bowles and Gintis 1976; Featherman and Hauser 1978; Bielby 1981; Ishida 1993).

23 The HQ2 effect in France became nonsignificant after controlling for class origin. However, the reduction was only 13%, and the effect of HQ2 was barely significant in the two-way table.

qualifications and farming-class destination is spurious. In England and Scotland, the effect of LQ2F actually became nonsignificant after control- ling for class origin. Farming-class destinations in these countries are likely to include a relatively high proportion of large, prosperous farmers, and, when we control for the flow from farm origin, the farming-class destination is not considered to be an undesirable position that merely absorbs people with no or minimum qualifications.

Hungary, the only country with a nonprivate collective agriculture, also stands out. In Hungary LQ2F is significantly larger than in all other countries studied, but, at the same time, controlling for social origin produces a smaller decline in LQ2F effects than elsewhere. In the Hun- garian institutional context workers with low education are thus more definitively allocated to the agricultural sector, and yet this nexus is less bound to the intergenerational reproduction of the agricultural classes. The strong association between low education and farm destination does not seem to be determined to the same extent by the orientation toward farming nurtured in agricultural families.

The detailed parameter estimates of the net QD effects disclose the mech- anism of transition from qualification to destination. Highly qualified indi- viduals possess better access to service-class positions in all nations. The HQ1 effect is consistently strong and significant across all countries. At- tainment of high qualifications also confers on individuals a better chance of avoiding manual work in some nations. When the HQ2 effect is not sig- nificant, these countries tend to have a relatively strong HQ1 effect. We expected that the state socialist countries would show a high value for HQ2 effect because of the state's wish to exploit to the maximum the human investment it has made. HQ2 parameters in Hungary and Poland are sig- nificant and higher than in some nations but by no means all.24

The comparison of HQ1 and HQ2 effects within each country suggests that the inclusion mechanism (HQ1) is stronger than the exclusion mecha- nism (HQ2). Therefore, the major source of advantage for the highly- qualified in the labor market appears to come from their privileged access to service-class positions. The only exception is in Japan, where the HQ1 effect is weaker than the HQ2 effect. In fact, the Japanese HQ1 effect is well below the same effect in most nations.'' This figure is most likely to be accounted for by a widespread organizational practice in Japanese

24 The HQ2 parameter in Hungary is significantly higher than that in England, North- ern Ireland, and West Germany, and the HQ2 parameter in Poland is significantly higher than that in England and Northern Ireland.

25 Pairwise comparisons of the HQ1 effect between Japan and each individual nation using backward elimination showed that the Japanese parameter is significantly smaller than that in all nations except Ireland and Scotland.

business: young university graduates are initially assigned to routine cleri-

cal grades and only later in their career move into administrative or

managerial positions (Cole and Tominaga 1976; Ishida 1993). The ten-

dency to include the highly qualified in the service class is thus less

marked in Japan than in other nations. Complementarily, the relatively

high HQ2 effect for Japan indicates that in this country holders of high

education are strongly protected against obtaining manual occupations

(Ishida, Goldthorpe, and Erikson 1991).

There are different classes into which people with low or minimum qualifications are likely to be allocated. First, individuals with low or minimum qualifications are likely to be excluded from entry into service- class positions (LQ1). Second, individuals with low or minimum qualifi- cations have a high propensity to move into unskilled working-class posi- tions (LQ2U). The LQ2U effect is generally substantial except in Poland; a close association between low qualification and unskilled positions consti- tutes an important aspect of the transition from qualification to destination.

Third, the poorly qualified have a high propensity to move into skilled working-class positions in some nations. The chances of inclusion of the poorly-qualified in the skilled working class, however, appear to be al- ways smaller than their chances of inclusion in the unskilled. This finding probably reflects not only the fact that some skilled jobs require more than minimum qualifications but also that the unskilled working class forms a distinct category of relatively undesirable positions which the poorly qualified are least able to avoid.

Fourth, with the exception of England and Scotland, low qualifications tend to be associated with farming; we observed consistently high values of LQ2F. Finally, the petty bourgeois destination also appears to be accessible to those with low qualifications in some nations, although the extent of inclusion is generally weaker than in the working classes and the farming sector. Like the LQ2F effect, the LQ2P effect is likely to be affected by the composition of the petty bourgeois destination. For exam- ple, in Japan, we expect a relatively large segment of this class to be proletarianized and composed of small, usually familial, business enter- prises (Patrick and Rohlen 1987; Ishida et al. 1991), and this segment probably absorbs people with minimum qualifications.

Poland shows values for the LQ effects that differ considerably from those of the other nations. In Poland exclusion of those with low qualifi- cations from the service class (LQ1) is much more marked than in other nati~ns.'~On the other hand, the various inclusion effects (LQ2) are

26 The LQ1 effect in Poland is significantly greater than the same effect in other countries, according to the pairwise comparison between Poland and each individual nation using backward elimination.

generally smaller in Poland than in the other nations, except for LQ2F.27 What is then common to those with low qualifications in Poland is not so much inclusion into the class of workers, farmers, or the petty bour- geoisie but exclusion from the service class; failure to obtain qualifications appears to be a serious and consistent disadvantage in access to profes- sional, administrative, and managerial positions. Poland also differs from all other nations in that the LQ2P effect is negative. This means that persons with low qualifications are not only disadvantaged in entry into the service class but also have a lower propensity to move into petty bourgeois positions. It also means a correspondingly higher propensity for people with middle or higher levels of qualifications to take up self- employment. These peculiar Polish features may reflect the state's control over the allocation of various class positions. The state probably acts as a gatekeeper excluding the poorly qualified, especially from governmental and industrial bureaucracies. It may also exercise some control over ac- cess to self-employment in the sense that entries into the petty bourgeoisie are often facilitated by some forms of qualificati~n.~~

Role of Qualifications in Class Reproduction and Mobility

In this final section, we address the issue of whether patterns of class reproduction and mobility are mediated by means of qualifications. In order to accomplish this task, we first need to represent with a model the association between origin and destination, as we have done in the cases of the OQ and QD associations. Erikson and Goldthorpe (1987a, 1987b, 1992a, 1992b; Ishida et al. 1991) have already analyzed the pat- terns of class reproduction and mobility among the nations examined in this article and proposed a model of core social fluidity. The model of core social fluidity is characterized by a number of different effects op- erating in the origin by destination class mobility tables. They are the inheritance (IN), hierarchy (HI), sector (SE),negative affinity (AFl), and positive affinity (AF2) effects. We believe, following Erikson and Gold- thorpe, that the patterns of association between origin and destination are captured by these effect matrices because we define class mobility within the context of class structure rather than status hierarchy (Gold- thorpe 1987). 29

*'The LQZU effect is significantly smaller than that of all other nations, and the

LQZS effect is smaller than that of all nations except France.
28 This point was suggested to us by Polish sociologists (Kazimierz Skomczyiiski and
Krzysztof Zagorski) at the CASMIN Conference in Reisenburg, Germany, March 1988.

29 We are fully aware of alternative models that emphasize the "hierarchical" or
"vertical" dimension of mobility processes (e.g., Hauser 1984; Hout 1989; Hout and
Hauser 1992), and we firmly believe in the utility of such models in addressing the

We directly build upon Erikson and Goldthorpe's specification of effect matrices but are forced to modify their core model for two reasons.30 First, we employed a sixfold version of the class schema, rather than their sevenfold version, by collapsing farmers (IVc) and farm workers (VIIb). The farm workers category was too small for the analysis of the three-way table involving qualifications. Collapsing these categories caused the sector and the negative affinity effects in the core model to disappear because both effects involved the farm workers category.

Second, and more important, our analysis attempts to capture the effect of qualifications on patterns of class reproduction and mobility, but some of the effects specified by Erikson and Goldthorpe confound mobility processes that involve the use of qualifications with those that do not. Therefore, we modified the inheritance effect and the positive affinity effect in their core model. For the inheritance effect, each class has its own class reproduction parameter (DIG)-unlike the model pro- posed by Erikson and Goldthorpe, which assigned the same class inheri- tance parameter to the service class and the petty bourgeoisie and to the routine nonmanual, skilled, and nonskilled class. Our decision is based on the understanding that class reproductions result from different social processes, which may or may not involve qualifications (Robinson and Kelly 1979; Robinson 1984).

The positive affinity effect proposed by Erikson and Goldthorpe (AF2) is divided into two components: one (AF2A) related to mobility between class I +I1 and class 111,and the other (AF2B) related to mobility between the service and the petty bourgeois class, between the petty bourgeois and the farming class, between the skilled and the nonskilled working class, and from farming-class origins to nonskilled working-class destina- tions. We hypothesize that a high propensity for mobility between class If11 and class I11 (AFZA) is likely to be associated with the use of qualifications. Sons of class I11 families have a high propensity for up- ward mobility into the service class, probably because they attain high qualifications. On the other hand, some highly qualified sons of service- class origin may move into class I11initially while awaiting their promo- tion into class I+11.

research question about the dependence of the son's socioeconomic status on the father's (Duncan 1979; Goodman 1979)and the mediating role of education. However, given that we are working with the six-category version of the class scheme, not the full version, and that we are primarily concerned with the effective use of qualifica- tions by different classes in the process of class reproduction and mobility, Erikson and Goldthorpe's (1987a, 19876, 1992a, 19926) representation of the association be- tween origin and destination appears to be more suitable for our purpose.

30 The data sets used in Erikson and Goldthorpe (1987a, 1987b, 1992a, 1992b) are identical to ours, but our analysis covers an age range (30-64) different from theirs (20-64). See the data section for details.

We hypothesize that affinities among the petty bourgeoisie, the farming class, and the working classes (AF2B) are independent of qualifications. The affinity between the petty bourgeoisie and the farming class, we suggest, arises out of a relatively easy transfer of capital between the farming sector and the small-business sector (Yamaguchi 1983; Erikson and Goldthorpe 1987a), which is often found in countries such as Ireland and Northern Ireland (Hout 1989).~' The affinity between the petty bour- geoisie and the service class is likely in part to be caused by the blurred boundary between the two classes. Large proprietors who employ more than 25 employees are in the service class, whereas small proprietors are in the petty bourgeoisie. Similarly, some petty bourgeois proprietors may become service-class salaried employees if the firm is incorporated. In addition, part of positive affinity AF2B also arises out of similarity in the manual nature of the work, that is, the "blue collar block," and the relatively easy flow from the farming sector to unskilled work. All these social processes are hypothesized to take place independently of the at- tainment of qualifications.

Because we do not modify the hierarchy effects (HI1 and H12) pro- posed by Erikson and Goldthorpe, we introduce five distinct effect matri- ces (inheritance, hierarchy 1, hierarchy 2, positive affinity A, and positive affinity B), which represent the core pattern of class reproduction and mobility among industrial nations. They take the form shown in appen- dix C. Erikson and Goldthorpe (1987b, 1992b; Ishida et al. 1991) have also identified some national variants of the core model. In order to produce acceptable fits, some effects matrices were modified or added for specific countries. Following Erikson and Goldthorpe, we have also introduced these variants, which are shown in appendix C.32

Given the model of social fluidity that represents the patterns of associ- ation between origin and destination, we estimate the impact of qualifi- cations on this model of social fluidity. We first obtain the estimates of the effect parameters in the collapsed origin by destination two-way ta- ble. We then compute the estimates of the same effect parameters after we introduce OQ and QD effect matrices in the three-way table. There- fore, following Yamaguchi's (1983) innovative paper, we propose to iden-

31 In fact, the positive affinity effects (AFZB) are strong in Ireland and Northern

Ireland. 32 We did not, however, introduce an extra positive affinity matrix (which operates on the pair of cells indicating mobility between class IVab and IVc, in addition to the positive affinity effect already operating on these cells) for Northern Ireland as sug- gested by Erikson and Goldthorpe (19876, p. 155). This is because we had partitioned Erikson and Goldthorpe's version of positive affinity into two components and allowed them to have separate effects. We were able to produce an acceptable fit for Northern Ireland without introducing an additional positive affinity.

tify (1) mobility processes that are likely to go through the qualification

channel (i.e., the changes in the effect parameters) and (2) mobility pro-

cesses that are independent of education and likely to involve more "spe-

cific" resources, possession of which affects chances of a particular pat-

tern of mobility (i.e., no change in the effect parameter^).^^

Tables 7 and 8 present the estimates of all the OD effect matrix param- eters and the goodness-of-fit statistics before and after controlling for qualification for our 10 nations.34 The second part of table 8 shows the percentage change (or reduction) in the estimates after controlling for qualification. Four findings stand out in tables 7 and 8. First, qualifica- tions play different roles in class reproduction. They are crucial for the reproduction of the service class in most countries. In Hungary, Sweden, and West Germany, the tendency for class reproduction among the ser- vice class is no longer significant when we control for qualifications. In England, Poland, and Scotland, the extent of class inheritance is substan- tially reduced, by more than 40%. These results suggest that the service class uses qualifications to reproduce its privileged position. The tendency for the sons of the service class to stay in the same class is in large part explained by their superior education.

However, in Ireland and Northern Ireland, this role of qualifications in the reproduction of the service class is much less pronounced. While the extent of service-class reproduction is strong, as indicated in the high values of the estimates, the percentage reductions after controlling for qualifications are small-10% in Ireland and 22% in Northern Ireland. The service class in these two countries does not appear to use qualifica- tions to pass on its privileged position from one generation to the next.35 In Japan, the class inheritance parameter for the service class was not significant in the origin by destination two-way table. Ishida et al. (1991) have already reported this weak tendency to class reproduction among

33 Because our qualification variable includes vocational and professional qualifica- tions, not simply the levels of education, education may not be labeled as "general- ized" resources, as done by Yamaguchi (1983). Yamaguchi, however, recognizes that the distinction between generalized and specific resources is an analytical one and that some qualifications should be treated as "specialized" resources (p. 720). For a similar technique, see Miiller et al. (1989).

34 The goodness-of-fit statistics for the three-way table after controlling for qualifica-

tion are already shown in table 6. 35 The reason for this observation is not clear, but it may be related to the composition of the service class. The proportion of the self-employed (self-employed professionals and large proprietors) in the service class may be greater in Ireland and Northern Ireland than in other societies. It is then not surprising that the inheritance of the service class is relatively independent of qualifications in these nations (for the compo- sition of the service class in Ireland and Northern Ireland, see Hout [1989, pp. 42-47]).

the Japanese service class. This weak tendency results in part from a special negative affinity effect that acccommodates a low tendency for the sons of the service class to be downwardly mobile into the manual working classes.

The reproductions of the petty bourgeoisie and the farming class are not affected by qualifications. Among our 10 nations, the reductions in the ex- tent of class reproduction for class IVab and class IVc +VIIb after control- ling for qualifications are small, the highest being only 11%in Japan. The tendency for the petty bourgeoisie and the farming class to inherit their fathers' positions is independent of qualifications and, therefore, is more likely to be the result of direct transmission of fixed capital, whether in the form of land or private business. Their propensity for immobility may also result from the transmission of cultural properties, such as attachment to land and the culture of independence, and the development of these cul- tural traits is not always compatible with better qualifications that open up opportunities other than small business and farming.

The reproductions of the routine nonmanual class and the manual working classes are not much affected by qualifications in any nation. The extent of reproduction of the routine nonmanual class is limited, being significant only in Hungary, Ireland, and Northern Ireland. Class I11is often considered to be a transient class (Stewart, Prandy, and Black- burn 1980), and the important feature of this class with regard to its reproduction and mobility is captured in our AF2A effect, that is, the exchange between class I11 and class I +11. The tendency for sons of the skilled working class to stay in their class appears to be largely indepen- dent of qualifications in all nations. The lack of effect of qualifications may be related to our educational scheme, which does not group together all the relevant vocational qualifications as a separate category; basic vocational training is included in the lowest category, while some techni- cal qualifications are in the middle category.36 The reproduction of the nonskilled working class is generally not affected by qualifications either. In Hungary the tendency for sons of the nonskilled to move out of their class of origin (because the parameter estimate is negative) is in large part explained by qualifications. It turns out that these sons have a rela- tively high tendency to move into the service class, as we find in Poland

36 It should be noted that in France apprenticeship is included in the middle category, but nonetheless the transmission of the skilled working class is independent of qualifi- cations. In France class reproduction and mobility are in general not much dependent upon qualifications, a finding consistent with Konig and Miiller's (1986) conclusion that organizational mobility, rather than mobility through the attainment of qualifica- tions, is more important in France than in West Germany (cf. Robinson and Garnier 1985).

TABLE 7
England         Northern       West
and Wales France Hungary Ireland Japan Ireland Poland Scotland Sweden Germany
Goodness of fit:                  
Gz ............................... 30.11                  
df ................................ 15                  
pa ................................ 012                  
% misclassified .............. 1.2                  
% reduction in GZ ........ 98.1                  
.................GZ(S, 1, 483)' 18.14                  
pd ................................ 255                  
Effect matrix parameter:                  
....................DIG(I+ 11) 1.229**                  
(.137)                  
DIG(II1) ........................ 118                  
(.162)                  
DIG(1Vab) .................... 1.115**                  
(. 104)                  
.............DIG(1Vc+VIIb) 3.528**                  
(.161)                  

TABLE 8 ESTIMATES IN THE ORIGIN BY QUALIFICATIONTHREE-WAYTABLEAND THE

OF OD EFFECT MATRIX PARAMETERS BY DESTINATION PERCENTAGEREDUCTIONIN THE ESTIMATESOF OD EFFECT MATRIX AFTER CONTROLLINGFOR QUALIFICATIONS
  England and Wales France Hungary Ireland Japan Northern Ireland Poland Scotland Sweden West Germany
Effect matrix parameter: ....................DIG(I+ 11)                    

DIG(1Vc+VIIb) .............

DIG(V+VI) ..................

(see below), and this upward mobility is in part facilitated by the attain- ment of qualifications.37

Second, the effect of hierarchy 2 (HI2) is in part explained by the attainment of qualifications. In Ireland, Poland, and Scotland the HI2 effect becomes nonsignificant after controlling for qualifications. For the rest of the nations-except for Northern Ireland, where HI2 was not significant before controlling for qualifications-the reduction ranges from 23% to 39%. The HI2 effect indicates a low propensity for sons of service-class families to be downwardly mobile into the unskilled working class and a corresponding low propensity for sons of the unskilled work- ing class and the farming class to be upwardly mobile into the service class. The reduction in its effect after controlling for qualifications sug- gests that barriers to upward movement into the service class from the farming and unskilled classes can be in part ascribed to the lower levels of qualifications obtained by their sons. Barriers to downward movement from the service class to the unskilled class can also be explained in part by the superior qualifications obtained by sons of the service class.

In contrast, the effect of hierarchy 1 (HI1) is not much affected by qualifications in most countries.38 The effect of HI1 implies barriers to mobility between the service class and the intermediate classes (111, IVab, and V+VI) and between the intermediate classes and the rest of the manual classes (VIIa, VIIb, and IVc). Our findings suggest that these barriers do not necessarily arise out of differential access to qualifica- tions.39

Third, qualifications play different roles in two types of positive affin- ity. For positive affinity A (AF2A), which relates to a high propensity of mobility between class I+I1 and class 111, the exchange between these two classes is likely to be facilitated by the attainment of qualifications.

37 When we modified the positive affinity matrix (AFZA) in Hungary to include a pair of cells indicating mobility between the service class and the skilled and nonskilled working class (similar to a pair in Poland), the tendency for sons of the nonskilled to move out of their class origin disappeared. In other words, the mobility of these sons of nonskilled origin was captured by the modified AFZA matrix, which allows these sons to have a higher tendency to move up into the ranks of the service class.

38 In Northern Ireland the effect of HI1 became nonsignificant after controlling for qualifications. It should be noted, however, that the reduction was only 21%, and the effect of HI1 was barely significant in the two-way table. In all other instances where an effect became nonsignificant after controlling for qualifications, the average reduction was 59%.

39 It should be noted that the use of a more refined measure of the qualification variable (such as one that distinguishes between general compulsory education and basic vocational education or apprenticeship) would probably reduce the effect of HI1 after the control, because barriers to mobility between the skilled and the unskilled working class are part of the HI1 effect.

In Poland, however, AF2A relates to a high propensity of mobility be- tween the service class and the two manual working classes (see app. C for the special AF2A matrix in Poland). The reduction in its effect after controlling for qualifications suggests that the attempts of the state to recruit men of manual working-class origin into public bureaucracies and correspondingly to demote men of service-class origin were in part accomplished through differential access to qualifications.

In contrast, positive affinity B (AF2B) is not much affected by qualifi- cations except for West Germany. AF2B relates to instances of a high propensity of mobility involving the petty bourgeoisie, the farming class, and the two working classes.40 Our results suggest that the intergenera- tional fluidities among these classes as specified in the AF2B matrix take place in large part independently of the attainment of qualifications. Social processes underlying these affinities are more likely to be the results of a transfer of capital, an increase or decrease in the scale of business, or a similarity in the manual nature of work. In West Germany, however, the effect of AF2B becomes nonsignificant after controlling for qualifica- tions. This is probably because of the special AF2B matrix for West Germany-the pairs of cells indicating exchange between class I11 and class IVab are added (see app. C). AF2B in West Germany thus consists of an area of white-collar mobility and an area of blue-collar mobility, and mobility within the white-collar region is likely to be facilitated by the attainment of qualifications.

Fourth, the effect of negative affinity (AF1) in Japan, the special na- tion-specific matrix, is in part explained by the attainment of qualifica- tions. The negative affinity indicates a low propensity for sons of service- class origin to be downwardly mobile into the manual working classes. Our results suggest that qualifications are in part used by these sons of the service class to avoid this downward mobility. However, negative affinity (AF1) in West Germany, which shows a low propensity for sons of the nonskilled working class to move into the white-collar positions of classes I +11, 111,and IVab, is not much affected by qualifications. These barriers to upward mobility for the nonskilled working class appear not to be the result of inferior qualification^.^'

In summary, when we look across societies, a cross-national similarity

40 It should be remembered that in Hungary AF2B does not include the pair of cells indicating mobility between I+II and IVab and that in Ireland AF2B includes the pair of cells indicating mobility between IVc and If11 (see app. C and Erikson and Goldthorpe [1987b, pp. 153-551 for justification).

41 The importance of qualifications in accounting for the AF1 effect may be underesti- mated because our qualification variable does not put apprenticeships in a separate category in Germany.

can be found in the effect of qualifications upon class reproduction and mobility. The reproduction of the petty bourgeoisie and the farming class shows a striking similarity across nations. Among our 10 nations, the reproduction of these classes is not mediated by qualifications. A cross- national similarity also appears to dominate in the influence of qualifica- tions upon the hierarchy effects and the positive affinity effects. These findings lead us to conclude that class reproduction and mobility involve different social processes which are differentially affected by the attain- ment of qualifications and that the effect of qualifications on class repro- duction and mobility is remarkably uniform across all nations.

SUMMARY AND CONCLUSIONS

This article addressed three related themes about the relationship among class origin, class destination, and education. For each of the bivariate associations between these variables, we derived a set of effect matrices that were constructed to represent distinct social processes underlying these associations. Using effect matrices allowed us to identify similarities in the overall pattern of associations and at the same time to investigate variations in the relative importance of different social processes.

We empirically investigated the first theme, equality of opportunity for education, by evaluating the association between class origin and educational attainment. The development of educational systems took place in all our industrial nations in the 20th century, although to varying degrees (Benavot and Riddle 1988; Meyer, Ramirez, and Soysal 1992). Nonetheless, opportunities for education, that is, relative access to educa- tion by class origin, proved to be far from equal in these nations. Not only is the attainment of education affected by class origin in all nations, but also the association between class origin and education falls into a similar pattern across nations. Cross-national differences, however, are found in the extent or strength of these patterns of association.

Our examination of the relationship between class origin and education did not provide unequivocal support for the prediction that the state socialist nations would show a looser connection between class origin and education because of social policies designed to enhance the educational opportunities of the manual classes. The two socialist nations included in our study differ in this respect. While in Hungary the sons of the skilled working class may have benefited from such policies, in Poland the sons of both the skilled and unskilled working classes appear to be the least educationally disadvantaged among all nations studied. For both nations, however, our results did not provide any evidence of "posi- tive discrimination," whereby the sons of the manual classes occupy an advantaged position in educational attainment.

The second theme discussed in this article concerned the allocation of class position by education. Our analysis suggests that, with the exception of the farm sector, education plays the role of allocating individuals to different class positions independently of class origin. Only for children of farmers do we find results suggesting that they are likely to have low qualifications and also to take up agricultural employment later in life. The pattern of allocation to class positions given a specific level of educa- tion is highly similar in our 10 nations. The strength or extent of associa- tion, however, varies across nations. Cross-national variation in the ex- tent of various effects appears to reflect differences in socioeconomic history and in institutional arrangements regarding the use of educational credentials in the labor market across nations4* However, cross-national variation does not obviously correspond to different "societal types." The two state socialist nations, Hungary and Poland, show as much variation between them as there is among other nations in general. How- ever, the peculiar features of Hungary and Poland that distinguish them from each other are likely to be explained by a common causal factor- the political intervention of the state. In Poland the state appears to effectively channel access to the service class and the petty bourgeoisie, while in Hungary the state tends to allocate those with low qualifications into the cooperative farm sector. It is this causal interpretation that the state socialist nations may share in common, rather than the actual pat- tern or strength of the association between education and class positions.

Furthermore, Hungary and Poland appear to deviate more from other nations than capitalist nations do. It is precisely because of the critical role of the state, which can shape the process of allocating qualified individuals to various class positions, that the variation in such processes can be more marked among state socialist nations than among capitalist nations. Unlike the relationship between class origin and education, where the target of social policy was relatively clear-to enhance the educational opportunities of the sons of manual-class origins-policies regarding allocating qualified people into different class positions were more diverse.

Our third theme centered around the role of education in class repro- duction and mobility. Our findings lead us to conclude that class repro- duction and mobility involve different social processes, which are in turn differentially affected by educational attainment. A single measure ex-

42 The findings of our comparison of 10 nations are in broad agreement with the results reported in studies that focused on the historical development and institutional pattern of the linkage between educational credentids and labor market outcomes (Maurice, Sellier, and Silvestre 1982; Konig and Miiller 1986; Allmendinger 1989; Ishida 1993; Miiller 1994).

pressing the mediating role of education in the association between class

origin and class destination is an often-used summary, but it does not

capture the differential roles of education in the process of class reproduc-

tion and mobility. We need to distinguish different social processes gener-

ating class reproduction and mobility and to assess the effect of education

on each social process separately.

In concluding this article, we would like to discuss implications of our analysis for the study of comparative macrosociology. An influential hypothesis in this area is that of Featherman, Jones, and Hauser (1975; hereafter FJH), who emphasized a basic similarity in the pattern of rela- tive mobility chances or social fluidity across industrial nations (see also Grusky and Hauser 1984; Hauser and Grusky 1988; Ganzeboom et al. 1989; Hout and Hauser 1992; S#rensen 1992; Erikson and Goldthorpe 1992a, 1992b; Jones 1992). Our results suggest that their hypothesis can be extended to cover also the pattern of differential access to educational opportunities and the pattern of class allocation by education. Similarity among nations is found not only in the pattern of interaction between origin and destination but also in the way in which both origin and destination are related to education in industrial societies. The emphasis of the Featherman et al. hypothesis on relative rather than absolute chances of mobility seems to give it the potential to include also associa- tions involving education. Given the wide variety in the institutional forms of the systems of education and qualification in various countries, this conclusion is not at all self-evident.

The basic similarity in the pattern of association between origin and education and between education and destination, however, does not preclude some variations in the extent of association. Our attempt to account for such variation by a macrosociological variable proved to be not so successful. Hungary and Poland, whose political regimes separate them from the capitalist societies, did not cluster into a pattern that we could name the typical state socialist pattern. We are, therefore, inclined to share the skepticism of Erikson and Goldthorpe (1987b, pp. 161-2) about the utility of macrosociological variables, at least of different politi- cal regimes, for explaining the differences in the extent of association involving education. However, we would like to add that if similarities between these two nations are to be sought, they should be found in the causal interpretation of the associations involving education-political intervention-rather than the actual pattern and extent of such associa- tions. In other words, in order to better understand variations and simi- larities across nations, we must consider cross-national differences in the differential resources and incentives of various classes and the institu- tional context within which these associations take place (see Miiller and Karle 1993)-specifically, in the cases of Hungary and Poland, the role of the state in influencing allocation processes.

Finally, it is important to emphasize that a cross-national similarity appears to emerge again in the way in which education mediates the association between origin and destination. The role of education in class reproduction and mobility shows an impressive uniformity across indus- trial nations. Here again there exists a potential for extending the FJH hypothesis. A cross-national similarity in the outcomes, that is, the pat- tern of social fluidity, is likely to be achieved by a cross-nationally uni- form mechanism, that is, the differentially effective use of educational resources by different class origins.

APPENDIX A

TABLE A1

Orlglnal 10-Category Version

I Higher-grade professionals, administrators, and officials; managers in large in- dustrial establishments; large proprietors I1 Lower-grade professionals, administrators, and officials; higher-grade techni-

+.

00 cians; managers in small industrial establishments; supervisors of nonmanual N employees

I11 Routine nonmanual employees in administration and commerce; sales personnel;

other rank-and-file service workers IVa Small proprietors, artisans, etc., with employees IVb Small proprietors, artisans, etc., without employees IVc Farmers and small holders; other self-employed workers in primary production VIIb Agricultural and other workers (including family workers) in primary production V Lower-grade technicians, supervisors of manual workers VI Skilled manual workers VIIa Semiskilled and unskilled manual workers (not in agriculture)

I + I1

I11

IVah IVc + Wb V+VI VIIa

Six-Category Version

Service or professional-managerial

Routine nonmanual

Petty bourgeoisie

Farming class

Skilled workers

Unskilled workers

APPENDIX B

TABLE B1

DESCRIPTION VARIABLES
OF QUALIFICATION FOR 10 NATIONS

Qualification Level
by Country Description

England and Wales (ENG)
and Scotland (SCO):

Low: la ... . . . . .... . . . . . . ..... Inadequately completed general elementary education: left school before school leaving age. lb . . . . . . . .. . .. . . ......... General elementary education: met compulsory educational requirement; completed nonselective school with an 0-level

certificate; completed selective school within the limits of the term of compulsory education. lc ........... ............ General elementary education and basic vocational qualification: apprenticeship up until a City and Guilds Exam or general education of an upgraded minimum with at most an apprenticeship. Medium: 2a ........ ............... Intermediate vocational qualification: school leavers from selective schools with an apprenticeship or a City and Guilds Exam up until an Intermediate or Final certificate.

+.

00
2h .. . . . . . . . . . .. . . ..... . . . Intermediate general qualification: selective secondary education without A-level certificate. G, 3a . . . . .. . . ..... . ...... . . . Maturity level: secondary school certificate, A-level or equivalent, with or without vocational qualification. High: 3b ....................... Higher education, lower-level tertiary certificate: higher technical school diplomas, City and Guilds Full Technological Certificates, Higher National Certificates/Diplomas (HNC, HND, Level-C qualifications). 3c .................. . . . . . Higher education, upper-level tertiary certificate: university diplomas and professional qualifications (Level-B qualifica- tions). France (FRA):

Low: la ................... .... Inadequately completed general elementary education: no formal certificate. lb ...................... . General elementary education: C.E.P. (Certificat d'6tudes primaires); Certificat d'6tudes complementaires; diplbme d'6-

tudes primaires prkparatoires; certificat d'ktudes primaires d'adultes.
Medium:

lc . . ........ . . . . .... . .... General elementary education and basic vocational qualification: C.A. P. (Certificat d'aptitude professionnel); B.E.P. (Brevet d'Ctudes professionelle); Brevet de compagnon; B.P. (Brevet de maitrise); 1'" partie de B.E.C. (Brevet d'en- seignement commercial); B.E.I. (Brevet d'enseignement industriel); B.E.H. (Brevet d'enseignement hbtelier); B.E.S. (Brevet d'enseignement social).

2a .. .. ... . ... . . . . . .. . . . . . Intermediate vocational qualification: zemepartie: B.E.C., B.E.I., B. E.H., B. E.S.; B.T. (Brevet de technicien); Bac

technique lkepartie; Diplbme d'616ve brevet6 d'une,6cole nationale professionelle ou d'un lyc6e technique d'Ctat. 2h ....................... Intermediate general qualification: B.E.P.C.; Brevet Elementaire; Brevet Superieur; Bac's 1'" partie. 3a .. . .. . . . .. . . . . . . . ... . . . Maturity level: all Bac's 2'me partie; university propaedeutic; D.U.E.L.

TABLE B1 (Continued)

Qualification Level
by Country Description

High:

3b ....................... Higher education, lower-level tertiary certificate: Capacit6 en droit; B.T.S. (Brevet de technicien superieur); D.U.T. (dipl6me universitaire de technologie); dipl6mes d'6coles d'enseignement param6dical; dipl6me dVtudes superieures techniques.

3c ....................... Higher education, upper-level tertiary certificate: license, maftrises, doctorat d'ktat, doctorat de 3"e cycle. Hungary (HUN): Low: la ....................... Inadequately completed general elementary education: did not complete the sixth (in old system) or the eighth (in new

system) grade of the folk school. General elementary education: completion of primary school (six or eight grades) without any further education. General elementary education and basic vocational qualification: primary school and attendance of a vocational or voca-

tional-technical school with at least one year of schooling.
Medium:

P 2a ....................... Intermediate vocational qualification: graduated from vocationally oriented intermediate school (Technika, technical in- termediate schools) without ~atura,r with less than four of Gymnasium; Burgher student with sLbsequent vo- cational training or apprenticeship.

2b ....................... Intermediate general qualification: left Gymnasia without Matura or Burgher student without further schooling. 3a ....................... Maturity level: certificate of a Gymnasium (general or Technika). High: 3b ....................... Higher education, lower-level tertiary certificate: certificate presupposing the maturity level hut not leading to the high- est certificate attainable. 3c ....................... Higher education, upper-level tertiary certificate: university education. Republic of Ireland (IRE) and Northern Ireland (NIR): Low: la ....................... Inadequately completed general elementary education: left school earlier than a year before the end of the term of com- pulsory education. lb ....................... General elementary education: attended school until at least one year before the end of the regular term of compulsory education. lc ....................... General elementary education and basic vocational qualification: apprentice training or attendance of a technical school.

High: 3b ..................... 3c ....................... Japan (JAP): Low: la ....................... lb .......................

Medium: 2ab ......................

High: 3c ....................... Poland (POL): Low: la ....................... lb ...................... lc ....................... Medium: 2a ....................... 2b ....................... 3a ....................... High: 3b .......................

Intermediate vocational qualification: intermediate educational certificate and either graduation from apprentice train- ing or attendance at a technical school. Intermediate general qualification: left academic secondary school with certificate below the A-level. Maturity level: entry qualification for university attendance, that is, A-level and comparable certificate (for the Repub- lic of Ireland, includes Leaving Certificate).

Higher education, lower-level tertiary certificate: the highest diploma, either HNC or HND Higher education, upper-level tertiary certificate: university education.

Inadequately completed general elementary education: left school before the end of the regular compulsory education. General elementary education: completed old system primary education (kyusei jinjo shogakko; kyusei koto shoggako) and new system lower secondary education (shinsei chugakko).

Intermediate vocational and general qualification: completed old system lower secondary level school (kyusei chugakko, jitsugyo gakko, shihan gakko) and new system upper secondary school (shinsei kotogakko). Maturity level and lower-level tertiary education: completed old system upper secondary schools (kyusei kotogakko, ko- sen) and new system lower tertiary education and higher technical education (shinsei tandai, kosen).

Higher education, upper-level tertiary certificate: university degree

Inadequately completed general elementary education: elementary school without certificate.
General elementary education: elementary school with certificate, vocational school without certificate.
General elementary education and basic vocational qualification: vocational school with certificate.

Intermediate vocational qualification: upper school without certificate, vocational orientations.
Intermediate general qualification: upper school without certificate, general orientations.
Maturity level: upper school with certificate, technical college without diploma, university without diploma.

Higher education, lower-level tertiary certificate: technical college with certificate.

Higher education, upper-level tertiary certificate: university diploma.

TABLE B1 (Continued)

wede en<^^^):

Low: la ....................... lb .......................

Medium: 2 a .......................

3a ....................... High: 02 3b .......................

m

3c ....................... West Germany (WG):

Low: la ....................... lb ....................... lc .......................

Medium: 2 a .......................

2b ....................... 3a ....................... High: 3b .......................

Description

Inadequately completed general elementary education: less than six years of elementary school education. General elementary education: elementary school graduation (6-8 years) and elementary school graduation with subse- quent vocational training of one year or less. General elementary education and basic vocational qualification: elementary school graduation with more than one year of subsequent vocational training.

Intermediate vocational qualification: elementary school graduation with an intermediate certificate and subsequent vo- cational training of more than one year. Intermediate general quali6cation: Intermediate school graduation without vocational training or with vocational train- ing of one year or less. Maturity level: maturity exam but no further education or with vocational training of one year or less.

Higher education, lower-level tertiary certificate: maturity exam plus several years of vocational training or university attendance without diploma. Higher education, upper-level tertiary certificate: graduation from a university-level institution.

Inadequately completed general elementary education: no formal education certificate.

General elementary education: folk school certificate without apprenticeship.

General elementary education and basic vocational qualification: vocational school with an industrial, craft, or commer- cial apprenticeship.

Intermediate vocational qualification: school for technicians, professional school (Bermfsfachschule), or technical school

with an industrial or commercial orientation. Intermediate general qualification: Intermediate Maturity Certificate (Mittlere Reife). Maturity level: Abitur.

Higher education, lower-level tertiary certificate: degree from an engineering school, technical college, or teachers college. Higher education, upper-level tertiary certificate: degree from a university.

i?Jo~~.-The summary descriptions and category numbers, except those for Japan, come from Konig et al. (1988, pp. 61-68)

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

...

8;;;

B;;;

4;;;

Lr:::

4;;;

-...

h"'

5;;;

8;;;

3:::

0;;;

1:::

u"'

9;;;

...

LI.::

2;::

!a;;;

,2;; ;

4;;;

%:::

....

....

....

....

....

....

....

....

....

....

....

....

.,..

....

....

....

....

....

....

....

....

....

....

....

....

....

-....

0::::

...

F;;;;

....

3.. ..

Lr::::

s;;;;

....

x::::

G....

u:::: 0;;;;

*....

c::::

9;;;;

LI::::

3;;;; 3;;::

..

.-....

L S : : .'n,

% ;aj ;El

u.::>

.eH

0H

a+=$$

a)HHHH

z

REFERENCES

Allmendinger, Jutta. 1989. "Educational Systems and Labour Market Outcomes." European Sociological Review 5:231-50. Alwin, Duane, and Robert M. Hauser. 1975. "The Decomposition of Effects in Path Analysis." American Sociological Review 40:37-47. Bechhofer, Frank, and Brian Elliott. 1981. The Petite Bourgeoisie: Comparative Studies of the Uneasy Stratum. London: Macmillan. . 1985. "The Petite Bourgeoisie in Late Capitalism." Annual Review of Sociol- ogy 11:181-207.

Bell, Daniel. 1973. The Coming of Post-Industrial Society. New York: Basic.

Benavot, Aaron, and Phyllis Riddle. 1988. "National Estimates of the Expansion of Mass Education, 1870-1940." Sociology of Education 61:191-210. Bernstein, Basil. 1977. Class, Codes and Control, vol. 3. 2d ed. London: Routledge & Kegan Paul. Bielby, William T. 1981. "Models of Status Attainment." Research in Social Strati- fication and Mobility 1:3-26. Blau, Peter M., and Otis Dudley Duncan. 1967. The American Occupational Struc- ture. New York: Wiley.

Bourdieu, Pierre. 1973. "Cultural Reproduction and Social Reproduction." Pp. 7112 in Knowledge, Education and Cultural Change, edited by Richard Brown. London: Tavistock.

. 1974. "The School as a Conservative Force." Pp. 32-46 in Contemporary Research in the Sociology of Education, edited by John Eggleston. London: Methuen.

Bourdieu, Pierre, and Jean-Claude Passeron. 1977. Reproduction in Education, Soci- ety and Culture. Beverly Hills, Calif.: Sage. Bowles, Samuel, and Herbert Gintis. 1976. Schooling in Capitalist America. New York: Basic.

Cole, Robert E., and Ken'ichi Tominaga. 1976. "Japan's Changing Occupational Structure and Its Significance." Pp. 53-95 in Japanese Industrialization and Its Social Consequences, edited by Hugh Patrick. Berkeley: University of California Press.

Collins, Randall. 1979. The Credential Society. New York: Academic Press. Davis, Kingsley, and Wilbert Moore. 1945. "Some Principles of Stratification." American Sociological Review 10:242-49.

De Graaf, Paul M., and Harry B. G. Ganzeboom. 1993. "Family Background and Educational Attainment in the Netherlands for the 1891-1960 Birth Cohorts." Pp. 75-100 in Persistent Inequality, edited by Yossi Shavit and Hans-Peter Blossfeld. Boulder, Colo.: Westview.

Duncan, Otis Dudley. 1975. Introduction to Structural Equation Models. New York: Academic Press. . 1979. "How Destination Depends on Origin in the Occupational Mobility Table." American Journal of Sociology 84:793-803.

Erikson, Robert, and John Goldthorpe. 1987a. "Commonality and Variation in Social Fluidity in Industrial Nations: A Model for Evaluating the FJH Hypothesis." European Sociological Review 354-77. . 19876. "Commonality and Variation in Social Fluidity in Industrial Na- tions: The Model of Core Fluidity Applied." European Sociological Review 3:145

66. . 1992a. "The CASMIN Project and the American Dream." European Socio- logical Review 8:283-305.

. 1992b. The Constant Flux: A Study of Class Mobility in Industrial Societies. Oxford: Clarendon.

Erikson, Robert, John Goldthorpe, and Lucienne Portocarero. 1979. "Intergenera- tional Class Mobility in Three Western European Societies." British Journal of Sociology 30:415-41.

Featherman, David L., and Robert M. Hauser. 1978. Opportunity and Change. New York: Academic Press.

Featherman, David L., F. Lancaster Jones, and Robert M. Hauser. 1975. "Assump- tions of Social Mobility Research in the United States: The Cases of Occupational Status." Social Science Research 4:339-60.

Gambetta, Diego. 1987. Were They Pushed or Did They Jump? Cambridge: Cam- bridge University Press.

Ganzeboom, Harry, Ruud Luijkx, and Donald Treiman. 1989. "Intergenerational Class Mobility in Comparative Perspective." Research in Social Strati$cation and Mobility 8:3-84.

Goldthorpe, John H. 1982. "On the Service Class, Its Formation and Future." Pp. 162-85 in Social Class and the Division of Labour, edited by Anthony Giddens and Gavin Mackenzie. Cambridge: Cambridge University Press.

Goldthorpe, John H., with C. Llewellyn and C. Payne. 1987. Social Mobility and Class Strmcture in Modern Britain, 2d ed. Oxford: Clarendon.

Goodman, Leo A. 1979. LLMultiplicative Models for Analysis of Occupational Mobility Tables and Other Kinds of Cross-Classification Tables." American Journal of Soci- ology, 84~804-19.

Griffin, Larry J., and Arne L. Kalleberg. 1981. "Stratification and Meritocracy in the United States." British Journal of Sociology 32:l-38.

Grusky, David B., and Robert M. Hauser. 1984. "Comparative Social Mobility Re- vised: Models of Convergence and Divergence in 16 Countries." American Sociolog- ical Review 49:19-38.

Halsey, A. H., A. F. Heath, and J. M. Ridge. 1980. Origins and Destinations. Oxford: Clarendon. Hauser, Robert M. 1978. "A Structural Model for the Mobility Table." Social Forces 56:919-53. . 1984. "Vertical Class Mobility in England, France and Sweden." Acta Socio- logica 27237-110, 387-90.

Hauser, Robert M., and David B. Grusky. 1988. "Cross-National Variation in Occu- pational Distributions, Relative Mobility Chances, and Intergenerational Shifts in Occupational Distributions." American Sociological Review 53:723-41.

Heyns, Barbara, and Ireneusz Bialecki. 1993. "Educational Inequalities in Post-war Poland." Pp. 303-36 in Persistent Inequality, edited by Yossi Shavit and Hans- Peter Blossfeld. Boulder, Colo.: Westview.

Hout, Michael. 1983. Mobility Tables. Beverly Hills, Calif.: Sage. . 1988. "More Universalism, Less Structural Mobility: The American Occupa- tional Structure in the 1980s." American Journal of Sociology 93:1358-1400. . 1989. Following in Father's Footsteps. Cambridge, Mass.: Harvard Univer- sity Press. Hout, Michael, and Robert M. Hauser. 1992. "Symmetry and Hierarchy in Social Mobility." European Sociological Review 8:239-66. Ishida, Hiroshi. 1993. Social Mobility in Contemporary Japan. Stanford, Calif.: Stan- ford University Press. Ishida, Hiroshi, John Goldthorpe, and Robert Erikson. 1991. "Intergenerational Class Mobility in Postwar Japan." American Journal of Sociology 96:954

75. Jones, F. L. 1992. "Constant Social Fluidity: A Comment on Recent Criticisms." European Sociological Review 8:255-59.

Jonsson, Jan 0. 1987. "Educational Resources. " Pp. 153-97 in Welfare in Transition, edited by Robert Erikson and Rune Aberg. Oxford: Clarendon.

. 1988. Utbildning, Social Reproduktion och Social Skiktning (Education, so- cial reproduction, and social stratification). Stockholm: Almqvist & Wiksell. , 1993. "Persisting Inequalities in Sweden." Pp. 101-32 in Persistent Inequal- ity, edited by Yossi Shavit and Hans-Peter Blossfeld. Boulder, Colo.: Westview.

Kerckhoff, Alan C., Richard T. Campbell, and Jerry M. Trott. 1982. "Dimensions of Educational and Occupational Attainment in Great Britain." American Sociolog- ical Review 47:347-64.

Kerr, C., J. T. Dunlop, F. H. Harbinson, and C. A. Myers. 1960. Industrialism and Industrial Man. Cambridge, Mass.: Harvard University Press.

Konig, Wolfgang, Paul Liittinger, and Walter Miiller. 1988. "Comparative Analysis of the Development and Structure of Educational Systems." CASMIN Working Paper no. 12. CASMIN Project, Mannheim.

Konig, Wolfgang, and Walter Miiller. 1986. "Educational Systems and Labour Mar- kets as Determinants of Worklife Mobility in France and West Germany: A Com- parison of Men's Career Mobility, 1965-1970." European Sociological Review 2: 73-96.

Krymkowski, Daniel. 1991. "The Process of Status Attainment in Poland, the U.S. and West Germany." American Sociological Review 56:46-59. Levy, Marion J. 1966. Modernization and the Structure of Societies. Princeton, N.J.: Princeton University Press. Mare, Robert D. 1981. "Change and Stability in Educational Stratification." American Sociological Review 46:72-87. ~at~j6,

Petr. 1993. "Who Won and Who Lost in a Socialist Redistribution in Czecho- slovakia?" Pp. 251-71 in Persistent Inequality, edited by Yossi Shavit and Hans- Peter Blossfeld. Boulder, Colo.: Westview.

Maurice, Marc, Francois Sellier, and Jean Jacques Silvestre. 1982. Politique d'educa- tion et politique industrielle en France et Allemagne (Educational politics and indus- trial organization in France and Germany). Paris: Presses Universitaires de France.

Meyer, John, Francisco Ramirez, and Yasemin Nuhoglu Soysal. 1992. "World Expansion of Mass Education, 1870-1980." Sociology of Education 65:128-49.

Meyer, John, Nancy B. Tuma, and Krzysztof Zagorski. 1979. "Education and Occu- pational Mobility: A Comparison of Polish and American Men." American Journal of Sociology 84:978-86.

Muller, Walter. 1994. "Bildung und Soziale Plazierung in Deutschland, England und Frankreich" (Education and social placement in Germany, England, and France). Pp. 115-34 in Gesellschaft, Demokratie und Lebenschancen: Festschrift fur Ralf Dahrendo$, edited by Hansgert Peisert and Wolfgang Zapf. Stuttgart: Deutsche Verlangs-Anstalt.

Miiller, Walter, and Dietmar Haun. 1994. L'Bildungsungleichheit im sozialen Wan- del" (Social change and educational inequality). Kolner Zeitschrqt fz'ir Soziologie und Sozialpsychologie 46: 1-42.

Muller, Walter, and Wolfgang Karle. 1993. "Social Selection in Educational Systems in Europe." European Sociological Review 9:l-23. Miiller, Walter, Wolfgang Karle, Wolfgang Konig, and Paul Luttinger. 1989. "Class and Education in Industrial Nations." International Journal of Sociology 19:3-39. Parkin, Frank. 1971. Class Inequality and Political Order. New York: Praeger. . 1979. Marxism and Class Theory. New York: Columbia University Press.

Parsons, Talcott. 1951. The Social System. Glencoe, Ill.: Free Press.

Patrick, Hugh T., and Thomas P. Rohlen. 1987. "Small-Scale Family Enterprises." Pp. 331-84 in The Political Economy of Japan, vol. 1. Edited by Kozo Yamamura and Yasukichi Yasuba. Stanford, Calif.: Stanford University Press.

Raftery, Adrian, and Michael Hout. 1993. "Maximally Maintained Inequality: Expansion, Reform, and Opportunity in Irish Education, 1921-75." Sociology of Education 66:41-62.

Robinson, Robert V. 1984. "Reproduction of Class Relations in Industrial Capital- ism." American Sociological Review 49:182-96. Robinson, Robert V., and Maurice A. Gzrnier. 1985. "Class Reproduction among Men and Women in France." American Journal of Sociology 91:250-80. Robinson, Robert V., and Jonathan Kelly. 1979. "Class as Conceived by Marx and Dahrendorf." American Sociological Review 44:38-58. Shavit, Yossi, and Hans-Peter Blossfeld, eds. 1993. Persistent Inequality. Boulder, Colo.: Westview. Simkus, Albert, and Rudolf Andorka. 1982. "Inequalities in Educational Attainment in Hungary, 1923-1972." American Sociological Review 47:740-51.

Smith, Herbert L., and P. P. L. Cheung. 1986. "Trends in the Effects of Family Background on Educational Attainment in the Philippines." American Journal of Sociology 91:1387-1408.

Serensen, Jesper B. 1992. "Locating Class Cleavages in Inter-generational Mobility." European Sociological Review 8:267-81. Stewart, A., K. Prandy, and R. M. Blackburn. 1980. Social Stratification and Occu- pations. London: Macmillan.

Szczepanski, Jan. 1978. "Early Stages of Socialist Industrialization and Changes in Social Class Structure." Pp. 11-36 in Class Structure and Social Mobility in Poland, edited by Kazimier Slomczynski and Tadeusz Krauze. White Plains, N.Y.:

M. E. Sharpe.

SzelCnyi, Szonja, and Karen Aschaffenburg. 1993. "Inequalities in Educational Op- portunity in Hungary." Pp. 273-302 in Persistent Inequality, edited by Yossi Shavit and Hans-Peter Blossfeld. Boulder, Colo.: Westview.

Tominaga, Ken'ichi, ed. 1979. Nihon no Kaiso Kozo (The structure of stratification in Japan). Tokyo: Todai Shuppankai.

Treiman, Donald. 1970. "Industrialization and Social Stratification." Pp. 207-34 in Social Stratification: Research and Theory for the 1970s, edited by Edward 0. Laumann. Indianapolis: Bobbs-Merrill.

Treiman, Donald, and Kermit Terrell. 1975. "The Process of Status Attainment in the United States and Great Britain." American Journal of Sociology 81563-83.

Treiman, Donald, and Kam-Bor Yip. 1989. "Educational and Occupational Attain- ment in 21 Countries." Pp. 373-94 in Cross-National Research in Sociology, edited by Melvin L. Kohn. Newbury Park, Calif.: Sage.

Wesolowski, Wlodzimierz, and Bogdan Mach. 1986. "Unfulfilled Systemic Functions of Social Mobility: 11, The Polish Case." International Sociology 1:173-87. Wong, Raymond Sin-Kwok. 1990. "Understanding Cross-National Variation in Oc- cupational Mobility." American Sociological Review 55:560-73. . 1992. "Vertical and Nonvertical Effects in Class Mobility: Cross-National Variations." American Sociological Review 57:396-410. Xie, Yu. 1992. LLLog-Multiplicative Layer Effect Models for Comparing Mobility Tables." American Sociological Review 57:380-95. Yamaguchi, Kazuo. 1983. "The Structure of Intergenerational Occupational Mobil- ity." American Journal of Sociology 88: 7 18-45. . 1987. "Models for Mobility Tables: Toward Parsimony and Substance." American Sociological Review 52:482-94.

 

Comments
  • Recommend Us