@Bergman2001

(2001) - Manfred Max Bergman, Dominique Joye

Journal: Cambridge studies in Social research
Link::
DOI::
Links::
Tags:: #paper #SocialClass #CAMSIS #ISCO-88 #Wright #Goldthorpe
Cite Key:: [@Bergman2001]

Abstract

The purpose of this article is to inform researchers in the social and political sciences about the main social stratification scales in use today. Six stratification schemas are described in this text: the Cambridge Social Interaction and Stratification Scale (CAMSIS), Swiss Socio-Professional Categories (CSP-CH), John H. Goldthorpe’s class schema, the International Standard Classification of Occupations (ISCO-88), Donald J. Treiman’s prestige scale, and Erik Olin Wright’s class structure. Their theoretical backgrounds and assumptions are discussed, as are their structural and methodological aspects. General problems of contemporary stratification research are covered, and suggestions for future research directions within this field are proposed.

Notes

“Where they diverge, however, is in the explanation of how these titles relate to stratification. For example, occupational titles can have stratifying functions due to (a) the socio-economic relations which individuals share with each other on the basis of their occupations, (b) class interests based on the differential relations of occupations to authority and capital, (c) scarce, yet desirable, resources in the form of skills and knowledge that go along with these occupations and that can be transformed into advantage and power, and (d) differential social status or prestige that represent the symbolic value of occupations and correspond to variations in advantage and power.” (Bergman and Joye, 2001, p. 5)

“nature of the work performed by a worker has been used widely as a factor that can be grouped in meaningful ways so as to reflect social stratification within a society” (Bergman and Joye, 2001, p. 6)

“ISCO-88 classifies work according to, first, tasks and duties related to an occupation and, second, relevant skills that are necessary for fulfilling the formal and practical requirements of a particular occupation (International Labour Office, 1990; Elias, 1997a; 1997b). The most recent version emphasizes four skill levels, encompassing both formal education and informal training along with work experience as important classification criteria.” (Bergman and Joye, 2001, p. 7)

“Within a hierarchical framework, major groups (1-digit code) are subdivided into 28 sub-major groups (2-digit), which are subdivided yet again into 116 minor groups (3digit) and 390 unit groups (4-digit)” (Bergman and Joye, 2001, p. 8)

“For example, a nuclear physicist belongs to the unit group 2111 (physicists and astronomers), which are part of the minor group 211 (physicists, chemists and related professions), who belong to the sub-major group 21 (physical, mathematical and engineering science professionals), who are part of group 2 (professionals).” (Bergman and Joye, 2001, p. 8) A very Biological approach. Essentially classification based on species, genus etc

“Elias (1997b: 13) summarizes problems relating to the validity and reliability of the ISCO-88 classification schema as follows: • The extent and quality of the occupational data to be coded. The data to be coded may be too brief for application of a relevant occupational code, uninformative or may be ambiguous in its interpretation; • Instruments for the application of coding rules may be poorly formulated, leading to differences in their interpretation by different coders; • Poor coder training procedures may lead to errors in the application of coding rules; • Human error, which may be a result of fatigue and boredom – coding occupational information is usually a difficult and unrewarding task; • The classification itself may be poorly constructed, or may rely upon distinctions which cannot be readily operationalised in a particular context.” (Bergman and Joye, 2001, p. 9)

“The creators of this classification scheme implicitly assume that, first, an occupation can be reduced to a specific set of isolated tasks and duties, and that skills can be reduced to formal and informal education and on-the-job training schemes; second, that the tasks, duties, and skills of each occupation have been captured sufficiently; third, that tasks, duties, and skills neither interact nor can each of these compensate for another in the successful performance of a job; and, fourth, that a set of tasks, duties, and skills relating to an occupation are invariant across time and cross-national contexts.” (Bergman and Joye, 2001, p. 10)

“Central to Goldthorpe’s class schema are employment relations – cast in a functionalist perspective – in industrial societies, i.e. societies, which, according to Goldthorpe and his colleagues, operate on the basis of technical and economic rationality.” (Bergman and Joye, 2001, p. 11)

“Goldthorpe divides social stratification schemas into models that focus on either class structure or social hierarchy. Class structure refers to conceptualisations relating to the social positions of actors as identified by their relations within the labour market. In contrast, social hierarchy refers to an approach that, according to” (Bergman and Joye, 2001, p. 11)

“Goldthorpe, is interested in a single hierarchical dimension, e.g. prestige, status, economic resources, etc. He places his class schema into the former category and objects to the latter because it suggests a “vertical,” i.e. ascending/descending,” (Bergman and Joye, 2001, p. 12)

“oldthorpe’s class schema rests on a tripartite thematic division: employers, who purchase labour from employees and, thus, have authority over them; self-employed workers without employees, i.e. those who neither buy nor sell labour; and employees, who sell their labour to employers and, thus, are under their authority.” (Bergman and Joye, 2001, p. 12)

“Considerations which entered into the construction of Goldthorpe’s class schema, include: • Transformation of property into corporate forms. • Bureaucratisation of labour and organizations. • Authority, specialized knowledge, and expertise. • Sectorial divisions of occupations, especially with regard to agriculture vs. other sectors. • Job rewards and job-entry requirements. • The nature of the labour contract and the conditions of employment.” (Bergman and Joye, 2001, p. 12)

“When Runciman (1990) asked “How many classes are there in contemporary British society?,” he received this response: “As many as it proves empirically useful to distinguish for the analytical purposes in hand” (Erikson & Goldthorpe, 1992: 46).” (Bergman and Joye, 2001, p. 14)

“e take the view that concepts – like all other ideas – should be judged by their consequences, not by their antecedents. Thus, we have little interest in arguments about class that are of merely doctrinal value. (Erikson & Goldthorpe, 1992: 35)” (Bergman and Joye, 2001, p. 15)

“first, one of the goals in the social sciences is to explain social phenomena, including the antecedents, form and function, as well as the consequences, of social stratification and mobility. This goal seems difficult to attain by a recitation of statistical coefficients in the absence of explanatory tools in the form of an empirically grounded social theory” (Bergman and Joye, 2001, p. 15)

“Second, an emphasis on the results of a class schema during its construction may seduce its creators into adjusting the class categories post hoc in order to improve their fit to a desired set of empirical results.” (Bergman and Joye, 2001, p. 15)

“The class schema that we have developed ... is in its inspiration rather eclectic. We have drawn on ideas, whatever their source, that appeared to us helpful in forming class categories capable of displaying the salient features of mobility among the populations of modern industrial societies and within the limits set by the data available to us. (1992: 46)” (Bergman and Joye, 2001, p. 16)

“The various adjustments to the Goldthorpe class schema, combined with unclear procedural descriptions, raise concern about post hoc data fitting and the reliability of the current version” (Bergman and Joye, 2001, p. 16)

“His aversion to hierarchies is even more confusing, considering that his class schema was constructed in part under considerations of hierarchies (e.g. authority and working conditions). More generally, it seems difficult to conceive of a class schema, which is independent of hierarchy, which distinguishes qualitatively the classes from each other, but which relies on hierarchies as fundamental building blocks” (Bergman and Joye, 2001, p. 16)

“Different occupations may indeed occupy similar positions in a hierarchy and, concurrently, be subjected to different technical or economic realities. However, this does not invalidate hierarchies. Instead, a counter-argument could be made, which also does not invalidate the class schema: occupations subsumed in one class in Goldthorpe’s class schema may be located on completely different positions in hierarchical positions of prestige, status, etc.” (Bergman and Joye, 2001, p. 17)

“Even if the most detailed 11-class schema is considered, one wonders how homogeneous the groupings across these dimensions really are (cf. Goldthorpe, 1997; Prandy, 1998” (Bergman and Joye, 2001, p. 17)

“It is not quite clear why farm workers should be as predominant in his tripartite subdivision (see figure 2),” (Bergman and Joye, 2001, p. 17)

“he fewer classes to classify occupational titles, the less likely we find homogeneity within the classes, the more confusing is the meaning of the class schema, and the less convincing is a class structure from both a theoretical and empirical point of view.” (Bergman and Joye, 2001, p. 17)

“The 5-class version in particular is both parsimonious and has high face validity.” (Bergman and Joye, 2001, p. 18)

“Erik Olin Wright’s model of social stratification can be described as a materialist and neo-Marxist conceptualisation of class structure with occasional Weberian leanings.” (Bergman and Joye, 2001, p. 18)

“According to Wright (1985; 1997; 1998a; 1998b), Marxist writers of late have adopted at least four strategies to deal with the middle class, which impinge on one of the central tenets of Marxist ideology – a polarization of antagonistic class relations between the bourgeoisie and the proletariat: (1) the middle class as an ideological illusion; (2) as a segment of another class (e.g. the “new petty bourgeoisie” or “new working class”); (3) as a new class, distinct from the bourgeoisie, proletariat, or petty bourgeoisie; or (4) the middle class as belonging to more than one class, simultaneously. As we shall see, Wright’s mapping of class structure clearly belongs to (4)” (Bergman and Joye, 2001, p. 18)

“ater texts, however, he changed sides and not only agreed with writers such as Roemer (1982) on exploitation as the key feature in the relationship between the classes, but modified and expanded Roemer’s ideas to develop the latest version of his own stratification mode” (Bergman and Joye, 2001, p. 19)

“he rejection of domination as the defining feature was based on two insights: first, he conceded that domination does not automatically include exploitation (e.g. parents often dominate their children without necessarily exploiting them); second and more importantly, he understood that neo-Marxist models based on domination of one class over another in conjunction with, for instance, gender or ethnicity, become fractured, multifaceted, context-bound, and entangled in complex authority and power relations beyond materialist and realist perspectives (cf. Dahrendorf, 1959).” (Bergman and Joye, 2001, p. 19)

“first, the material welfare of one class has to depend o” (Bergman and Joye, 2001, p. 19)

“20 the exploitation of another class.” (Bergman and Joye, 2001, p. 20)

“econd, the material welfare of one class must depend on the efforts of another class, i.e. the rich appropriate surplus value from the labour of the poor.” (Bergman and Joye, 2001, p. 20)

“These two conditions – economic oppression and acquisition of surplus value – constitute materialist exploitation. Accordingly, Wright presents the following definition: classes are “positions within the social relations of production derived from these relations of exploitation” (1998: 13)” (Bergman and Joye, 2001, p. 20)

“Assets that define the bourgeoisie and the proletariat are ownership of means of production and of labour, respectively” (Bergman and Joye, 2001, p. 20)

“right suggested two different assets that are prevalent in modern capitalistic societies, especially due to the division of labour: bureaucratically controlled organizational assets and skills.” (Bergman and Joye, 2001, p. 20) middle class assets

“However, organizational assets bestowed upon individuals on the basis of their position as managers or supervisors in an organization or institution are different in nature from assets based on means of production and labour, since the latter two can be owned while it is difficult to conceive of ownership of organizational assets.” (Bergman and Joye, 2001, p. 20)

“he organization o” (Bergman and Joye, 2001, p. 20)

“21 the means of production and, thus, the differential relationship to authority creates surplus value beyond its expenditure in terms of labour or means of production.” (Bergman and Joye, 2001, p. 21)

“Skills, although these can be owned individually, also produce surplus beyond the expenditure of acquiring the skills, especially if unions, associations, professional credentials, or bureaucratisation protect such skills via, for instance, institutional accreditation and certification. Organizational assets can be used to extract surplus labour, as can skills, as long as a skill differential is protected and maintained between the experts and the non-experts, and as long as the value of the skills outweighs the cost of acquiring these skills” (Bergman and Joye, 2001, p. 21)

“According to Wright’s most recent work, the following are the conditions responsible for the class structure in modern complex societies: • In line with Marxist thinking, owners and wage labourers form two distinct meta-classes, where the owners of the means of production exploit the wage labourers by appropriating the surplus value produced by wage labourers. • In modern capitalist societies, assets are not limited to the ownership of production and labour, but include skills and organizational assets, which produce amongst the wage labourers a differential ordering of social structure according to the latter two assets.” (Bergman and Joye, 2001, p. 21)

“if wage labourers are divided across two axes (i.e. low, medium, and high skills; low, medium, and high organizational assets), a mapping of a class structure emerges that includes twelve classes: three owner classes and nine wage labourer classes, separated by ownership in the first instance, and sub-dividing the wage labourers across two dimensions – skill and relationship to authority. The following reproduced table (e.g. Wright, 1997; 1998a; 1998b) illustrates Wright’s class structure (Wright II):” (Bergman and Joye, 2001, p. 22)

“Table 3: Wright II Class Structure.” (Bergman and Joye, 2001, p. 22)

“Only class 1, class 3, and class 12 remain true to classical Marxism” (Bergman and Joye, 2001, p. 22)

“Wright employed ideas regarding “contradictory class locations” (Carchedi, 1977) to describe the mutual exploitation of some classes, thus offering the most pronounced departure from classical Marxism.” (Bergman and Joye, 2001, p. 22)

“From this perspective, it could be argued that a society which manages to produce wage labourers who are either marginally skilled or have a minimum of organizational assets will profit from exploiting each other, while, concurrently, being exploited by the owners of production” (Bergman and Joye, 2001, p. 23)

“Although problematic due to its neglect of property not associated with production (e.g. real estate, securities, stocks, bonds, etc.), property is measured from survey questions relating to employment status, and, if respondents are self-employed, focussing on how many workers they employ. Educational attainment or occupational codes are often used to determine respondents’ expertise. Authority is the most complicated and contested dimension. This dimension has been constructed from answers to questions relating to self-classification about management tasks at respondents’ places of work, participation in workplace policy decisions, or the ability to impose sanctions on subordinate workers.” (Bergman and Joye, 2001, p. 23)

“n other words, little convincing arguments have been put forth to justify the chosen criteria that ostensibly demarcate classes; instead, if continua are suggested by the theory, why not work with continua empirically (cf. Halaby & Weakliem, 1993)” (Bergman and Joye, 2001, p. 24)

“Thus, ownership of the means of production cannot be separated from skills or management assets because many owners are also skilled managers. Furthermore, successful managers and experts nowadays acquire ownership of means of production, for example through various stock options, profit sharing schemes, or simply through purchasing a part of the organization. Hence, the dividing line between owners and wage labourers may be more blurred than suggested by Wright.” (Bergman and Joye, 2001, p. 25)

“As with Goldthorpe’s class schema, users tend not to use the full range of classes as proposed by the author” (Bergman and Joye, 2001, p. 25)

“One wonders if such theoretical and ideological baggage may introduce unnecessary complications while, neglecting other stratification phenomena. In particular, the scheme depends entirely on the labour theory of value because, otherwise, exploitation has little meaning.” (Bergman and Joye, 2001, p. 25)

“ccording to Donald J. Treiman (1977: 1), the answer to the question “‘What sort of work do you do?’ provides the single best clue to the sort of person one is” in the sense that it not only positions individuals within a social structure based on their occupation, but, moreover, allows inference of attitudes, experiences, and life-style from prestige ratings of their occupations. Treiman’s work inscribes itself in the Chicago school of stratification, that is, a structural-functional perspective that borrowed ideas from especially Talcott Parsons and Emile Durkheim.” (Bergman and Joye, 2001, p. 26)

“sing the four nested levels of the International Standard Classification of Occupations (ISCO), Treiman’s occupational prestige scores for each occupation within an ISCO level are averaged to produce a score for occupational groups as summarized by ISCO. The most recent versions have been constructed in collaboration with Harry Ganzeboom and are markedly more sophisticated with regard to their methodological rigor (Ganzeboom & Treiman, 1992; 1996).” (Bergman and Joye, 2001, p. 26)

“Despite subsequent failures to replicate similar levels of goodness-of-fit when regressing education level and income on subjective prestige rating (e.g. Featherman & Stevens, 1982), SEI and its derivative, the ISEI (Ganzeboom & Treiman, 1992; 1996) have enjoyed great popularity in the social and political sciences” (Bergman and Joye, 2001, p. 27)

“Treiman’s Theory of Occupational Prestige rests on six implicit and explicit propositions:” (Bergman and Joye, 2001, p. 28)

“All complex modern societies are organized into similar societal functions in order to maximize efficiency (e.g. production of goods including foodstuffs, transport, surplus management, differentiated education, etc.) (cf. Durkheim, 1933; Aberle et al., 1950; Blau, 1964).” (Bergman and Joye, 2001, p. 28)

“Efficient logistic organization of these functions is satisfied through a similar division of labour across all complex modern societies (cf. Aberle et al., 1950).” (Bergman and Joye, 2001, p. 28)

“Division of labour creates a social stratification due to the differential in control over scarce, yet desirable, resources” (Bergman and Joye, 2001, p. 28)

“ifferentials in control over these resources produce differentials in power (i.e. the ability to achieve whatever ends are desired; cf. Weber, 1947; Treiman also quotes Shils, 1968: 110-111).” (Bergman and Joye, 2001, p. 28)

“Differentials in power (i.e. control over scarce and desired resources) result in differentials in privileges for the members of societies. Two processes are in operation to turn resources (e.g. skills, authority, economic resources, and power) into privilege: first, some resources such as skills are rare and valued, so the value of these skills is increased on the labour market and the price for some work relative to another increases; second, differential control over these scarce resources is bureaucratically enforced through, for example, professiona” (Bergman and Joye, 2001, p. 28)

“29 licensing, certification, and accreditation.” (Bergman and Joye, 2001, p. 29)

“Power and privilege are highly valued in all societies. Thus, powerful and privileged occupations are associated with high prestige ratings and prestige ratings for all occupations do not vary within or between modern complex societies.” (Bergman and Joye, 2001, p. 29)

“In sum, the subjectively attributed prestige of a specific occupation is (a) linked to the privilege and power which individuals enjoy based on their occupational titles, (b) invariant across social and cultural groupings, and (c) similar across all complex modern societies” (Bergman and Joye, 2001, p. 29)

“To defend these six claims, Treiman invests considerable effort in reanalysing data, which ostensibly reject the Cultural Hypothesis, i.e. that prestige hierarchies reflect cultural values and thus vary across different value systems and cultures” (Bergman and Joye, 2001, p. 29)

“Treiman suggests too many ad hoc explanations for disconfirming evidence. Exceptions to his model fall into at least four categories: (1) if occupational prestige ratings of one society differ significantly from others, the society may be declared not modern or complex” (Bergman and Joye, 2001, p. 30)

“(2) alternatively, divergence may be attributed to reduced data quality or measurement imprecision; (3) if prestige ratings of some occupations diverge across societies or social groups within society, they are declared “exceptions that prove the rule” (e.g. teachers, clerks, soldiers, police officers); (4) some divergences, e.g. the greater inter-societal agreement on prestige of whitecollar occupations, compared to blue-collar work (i.e. the inter-societal variation in the social organization of manual work), are either ignored or insufficiently discussed.” (Bergman and Joye, 2001, p. 30)

“Social stratification may not be captured by one single universal prestige scale. Prestige may in fact be a by-product of the social position, which is bestowed differentially upon occupations” (Bergman and Joye, 2001, p. 30)

“Treiman goes to great lengths to argue for the universality of occupational prestige because he believes that if prestige ratings vary across societies, then societies can no longer be compared because of the impossibility to attribute variations between societies to either variations in measurement or variations in societal structures.” (Bergman and Joye, 2001, p. 31)

“The causal chain embedded in Treiman’s theoretical proposition is complex and often difficult to falsify. While he argues that division of labour begets differential access to resources, differential access to resources begets differentials in power, differentials in power beget differentials in privilege, and differentials in privilege beget differentials in prestige, one could imagine the causal stream between the couplets to reverse or to interact with a third element. For instance, privilege may have an effect on power but little effect on prestige;” (Bergman and Joye, 2001, p. 31)

“• If it were indeed true that most modern societies must accommodate the same functions, it is not clear that this invariance would hold across all levels of the causal chain, from power through privilege to prestige” (Bergman and Joye, 2001, p. 32)

“Does subjectively assigned prestige reflect sufficiently access to power and privilege? If it does not, prestige is of little use to social stratification research; and if it does, why not measure privilege and power more directly?” (Bergman and Joye, 2001, p. 32)

“It is not clear which criteria Treiman used to select his list of occupations. Most frequently, the compiling of an occupational list is accomplished by both reproduction of occupational titles from previous research and ad hoc selection/deselection of individual occupations; either of these strategies is problematic with regard to constructing and validating a stratification model (cf. Coxon & Jones, 1974)” (Bergman and Joye, 2001, p. 32)

“Grouping the scores of Treiman’s Standard International Occupational Prestige Scale into the International Standard Classification of Occupations (ISCO) roster is only defensible if ISCO groupings do share characteristics that have as their basis a dimension similar to prestige, thus power and privilege” (Bergman and Joye, 2001, p. 32)

“Some of the studies of occupational prestige that were used to validate the occupational prestige scale were crude in their occupational coding, sampling of raters and occupations, and measurement.” (Bergman and Joye, 2001, p. 33)

“Occupations, which were not present in all societies, were either grouped into existing classifications or dropped from the calculations. This, of course, strengthens the universalistic claims of the prestige scale because it may cover up social and cultural variations.” (Bergman and Joye, 2001, p. 33)

“Although Treiman claims invariability of occupational prestige ratings across societies, as well as across social groups and context (e.g. gender, age, education, income, occupational group, raters’ values and attitudes), numerous studies produce counterevidence that shows variation across some of these groupings and dimensions (cf. Coxon & Jones, 1978; 1979a; 1979b).” (Bergman and Joye, 2001, p. 33)

“According to the CAMSIS approach, individuals are embedded in socially moderated networks of relationships within which they engage in social, cultural, political, and economic interactions, which are qualitatively and quantitatively different from interactions with persons who are more distant from these networks” (Bergman and Joye, 2001, p. 34)

“ltimately, relationship networks are constituted by, and reproduce, hierarchical inequalities” (Bergman and Joye, 2001, p. 34)

“This relational perspective of social, economic, and political structuring proposes a certain regularity and patterning of interactions, as well as an interactive negotiation of relations and their consequence” (Bergman and Joye, 2001, p. 34)

“social structure is not something given a priori, but continuously negotiated and reconstructed according to human interactions and the meanings, which they ascribe to these.” (Bergman and Joye, 2001, p. 34)

“The most general theoretical assumptions of CAMSIS are that resources are distributed systematically and unequally, according to socially regulated relationships within particular networks.” (Bergman and Joye, 2001, p. 35)

“Occupations are still the single most significant indicator of someone’s location in the overall structure of advantage and disadvantage (Blackburn & Prandy, 1997), as well as a major source of social identity, which can be localized to individual professions or professional groupings” (Bergman and Joye, 2001, p. 36)

“CAMSIS is gender-sensitive, i.e. different scales are calculated for men and women, since holding the same occupation may have different implications for the persons’ social position, depending on their gender. Beyond gender, other social groupings can also be accommodated in future versions, such as ethnicity, religion, education, language group, level of urbanization, or professional qualifications.” (Bergman and Joye, 2001, p. 36)

“Based on these theoretical considerations, computational procedures are applied that quantify the probabilistic relationships between social actors according to information about their occupations.” (Bergman and Joye, 2001, p. 37)

“Starting from this type of analysis of a frequency table, analytical techniques are limited to techniques that deal with χ2 decomposition” (Bergman and Joye, 2001, p. 37)

“One of the techniques suitable for this purpose is correspondence analysis, although the original scale was calculated from smallest space analysis. As part of the family of dimensional analyses, correspondence analysis (CA) is an exploratory statistical technique that analyses the structure within simple two-way (and multiway) tables from some measure of association between rows and columns. CA scores categories over a series of dimensions according to which category values relate more or less to each other within the structure of a particular dimension in the sense that the scores maximize the row/column correlations. When the scores in the first (most influential) dimension show an even ordering of all occupations, they tend to reflect an order to patterns of social interaction, which corresponds closely with an order of social stratification (e.g. Prandy, 2000).” (Bergman and Joye, 2001, p. 37)

“n other words, the first dimension in CA or RC-II Models may be dominated by patterns of interaction in just one or a very few “problem occupations,” i.e. occupations where there exists a high proportion of husbands and wives who have the same, or highly related, occupational titles (e.g. in agriculture). Either explicit modelling or exclusion of these “diagonal” cases in such occupations must take place in order to prevent these cases from influencing unduly the results.” (Bergman and Joye, 2001, p. 38)

“heoretically, it is most closely related to symbolic interactionism.” (Bergman and Joye, 2001, p. 38)

“• Because at the base of its scale are relationships of a specific population – in most cases these are national samples – the stratification scores for each occupation are sensitive to the idiosyncrasies of this population. In other words, unlike other schemas, CAMSIS scores calculated for one population cannot be simply transposed to another, but must be calculated specifically for this population” (Bergman and Joye, 2001, p. 39)

“However, it lags behind other schemes with regard to cross-national validation and application to studies in the social and political sciences.” (Bergman and Joye, 2001, p. 39)

“• As with most other scales, some tinkering goes into the construction of the scores. For instance, the exclusion or deliberate modelling of so-called “problem occupations” – selected at the discretion of the scale constructor – may introduce important variations with regard to both model fit and stratification scores.” (Bergman and Joye, 2001, p. 39)

“espite its parsimony, or probably because of it, CAMSIS has met strong resistance from those who are deeply immersed in the assumptions of and narratives around the concept of class: for many, stratification automatically implies distinct classes. While theoretically and empirically convincing, CAMSIS is less attractive to an established narrative in the social sciences. This is compounded by the fact that many social scientists would like to preserve the theoretical distinction between class and status and, thus, incorrectly classifying CAMSIS as a status-based schema. By its nature, however, CAMSIS denies the distinction between class and status.” (Bergman and Joye, 2001, p. 40)

“The Swiss Socio-Professional Categories (CSP-CH) schema has been developed, first, to improve on both the eclecticism and complexity inherent to the construction of other international stratification schemes (notably Erik Olin Wright and John H. Goldthorpe’s class schemas), second, to produce a stratification scheme within the limitations of the data available in Switzerland, and, third, to examine idiosyncratic characteristics of social stratification in Switzerland, which may be lost if applying a standardized schema to Swiss data” (Bergman and Joye, 2001, p. 40)

“he stratification schemes discussed in this text display a tremendous variety in terms of their theoretical underpinnings and methodological constructions. Despite this variation, it is surprising how strongly they correlate with each other and how similar they are with regard to predictive validity.” (Bergman and Joye, 2001, p. 43)

“CAMSIS and Treiman are considered measures on an (ordered-)interval scale, while ISCO and Goldthorpe’s schema are nominal. Wright II, Wright III, and CSP-CH are nominal as well, although embedded sub-dimensions within the class schema (for Wright II and Wright III: management assets and skills/credentialed assets; CSP-CH: highest achieved education level) make them something between an ordinal and a nominal scale” (Bergman and Joye, 2001, p. 43)

“However, their effectiveness in terms of the maximization of explained variance (i.e. the fetish around the R2) is often overrated and misused as evidence to discredit competitors” (Bergman and Joye, 2001, p. 43)

“For instance, highest achieved education level should not be used in the same model as ISCO-88 or the CSP-CH class schema since these schemas use education level as a component within their construction.” (Bergman and Joye, 2001, p. 43)

“Practically all stratification schemas are based on occupational titles of individual positions, which contradicts the original notion of class as a collective social phenomenon (cf. Blackburn, 1998).” (Bergman and Joye, 2001, p. 45)