The classical tradition of class analysis
The classical tradition of class analysis in Britain incorporated theoretical underpinnings derived from both Marx and Weber. It was part of a wider attempt to create an alternative theoretical paradigm to structural-functionalism which had dominated during the 1950s and 1960s. However, class was not clearly defined within this discourse. Two separate approaches ran in tandem and occasionally coalesced. The neo-Marxist emphasis on class as based upon property relations was paralleled by other concepts of class that were rooted within the division of labour of capitalist societies. The most popular of the latter were the concepts of the âmiddleâ and âworking classâ, which referred to groupings internal to the broader category of non-propertied labour. These binary concepts dominated much non-economic sociology. Indeed, a great deal of routine sociology during the 1960s and 1970s simply incorporated class as its predominant concept. Indeed, Stinchcombe â a leading American sociologist at the time â famously quipped that âsociology has only one independent variable, classâ (1968). In the field of educational sociology, a great deal of research examined the question of why working-class children failed to achieve similar results at school when compared to their middle-class counterparts (see Bernstein, 1975; Douglas, 1964; Floud, Halsey, & Martin, 1957; Willis, 1977). Similarly, political sociology had a preoccupation with why significant numbers of working-class voters failed to support left-of-centre social democratic and socialist parties (see MacKenzie & Silver, 1968; Nordlinger, 1967; Parkin, 1967).
A series of interconnected changes problematised the hegemony of these traditional class approaches within sociological discourse. The first involved manifest lacunae within the dominant paradigm itself. These included the absence of research on other aspects of structured inequality such as gender, ethnicity, age and disability. A great deal of empirical research at that time in economic sociology focussed exclusively upon samples of male factory workers (see Goldthorpe, Lockwood, Bechhofer, & Platt, 1968a, 1968b, 1969 as well as Blackburn & Mann, 1979). Since the late 1970s, there has been an efflorescence of research in these other spheres of inequality at the expense of class analyses.
A second element in the demise of specifically Marxist forms of class analysis involved the increasingly turgid and arcane (not to mention scholastic) nature of much of the literature in the area (see, e.g. Cutler, Hindess, Hirst, & Hussain, 1977; Jessop, 1988). Much of this was engaged in narrow debates about the political economy of modern capitalist societies which seemed increasingly distanced from developments outside the narrow confines of academia.
A third feature was the collapse of the simple notions of the âworking classâ and âmiddle classâ amongst sociologists. My own publications at that time had a part to play in this. My empirical research on skilled manual workers showed how they were a distinct stratum within the wider matrix of social stratification (see Penn, 1985b, 1990). This was part of a wider investigation into the disjunction between sociological and historical discourses concerning the chronology and trajectory of the British working class. The empirical results of this research threw into doubt the long cherished belief that there had ever been a homogenous working class with common interests or experiences in Britain.
The final element in this shift of emphasis away from classical notions of class was changes in the world that lay outside academia. These included the Falklands War: this puzzled many stratification experts like Howard Newby who stated at the 1983 Social Stratification conference in Cambridge that ânobody realized that the working class was nationalisticâ! I remember being perplexed by this odd claim which revealed a lack of knowledge about popular consciousness both historically and at the time it was made. This lack of connection with the world around them amongst economic sociologists also included the wider impact of the implosion of the Labour Movement symbolised by the 1984/1985 Miners Strike in Britain (see Penn, 1985a). Indeed, the evident disconnection between sociology and events was a major factor in the creation of the Social Change and Economic Life research initiative by the Economic and Social Research Council in the mid-1980s (see Gallie, 1994). The 1980s also witnessed successive Conservative Party electoral victories in Britain and led, ultimately, to the defeat of the Left within the Labour Party and the creation of Blairite âNew Labourâ. The sociology of class was left floundering as the strong relationship between class and voting in Britain (which actually only dated from 1945) disappeared before their eyes (see Robertson, 1984). This rendered the discussion of working-class âdeviantâ support for the Conservative Party (see Parkin, 1967; Taylor, 1978) irrelevant as working-class conservatism became the new norm and was encapsulated in notions such as âBasildon Manâ! (see Evans, 1999). These developments threw into doubt many taken-for-granted assumptions amongst sociologists and led many to cease to investigate class differences at all (see Lee & Turner, 1996). Overall, these elements outlined above combined to make class analysis both unfashionable and, in many sociologistsâ eyes, largely irrelevant to their main substantive concerns.
More recently, there has been an infusion of American-style approaches to the study of social inequality (see, e.g. Lambert, Connelly, Blackburn, & Gayle, 2012). By this I mean an application of rigorous statistical modelling and an emphasis on dimensions of stratification. This paradigm is seen clearly in the publication policies of the most prestigious US journals in sociology such as the American Sociological Review, the American Journal of Sociology and Social Forces. In the USA, such approaches have long represented the dominant paradigm (see Blau, Duncan, & Tyree, 1967; Coleman et al., 1966; Featherman & Hauser, 1978). This new approach has involved considerable debate about how best to measure individual-level inequality and which occupational scheme is the best for assessing social stratification (see Connelly, Gayle, & Lambert, 2016; Lambert & Bihagen, 2014). Many data sets such as the British Household Panel Study and the European Social Survey now allow the use of a wide range of such measures based primarily not only upon occupation, but also on educational attainment and patterns of social interaction such as friendship and marriage.
One problem with this ânew waveâ revisionism is that it is difficult to express results in a language that is intelligible to more than a few specialists. As a consequence, most sociologists outside the specialist field of social stratification still rely on common sense categories to interpret their results and/or they use the myriad of possible measures/schemes in an arbitrary and âad hocâ fashion. This is exemplified in the popular handbook by Shaw et al. (2007) which provides long lists of different conceptual schemes for measuring structured inequalities without any useful guidance as to which might be the most appropriate in any particular context.
Clearly, there is a myriad of different ways of measuring inequality available currently to empirical sociologists. All rely, in the main, on data collected about occupations.1 All are rooted in the way such occupational data were originally categorised by the UK Census authorities around the time of the First World War (see Penn, 1985b). The basis for the first systematic subdivision of the population based upon occupational position was a specific demographic and public health interest in the problem of infant mortality rates (see Registrar General, 1911; Stevenson, 1928). The categories used were as follows:
(1) | Upper |
(2) | Intermediate |
(3) | Skilled |
(4) | Intermediate skill |
(5) | Unskilled |
(6) | Textile workers |
(7) | Coal miners |
(8) | Agricultural labourers |
The main heuristic behind this seminal classification was to be able to demonstrate that there was a clear, graded increase in rates of infant mortality from class 1 to class 5 and thereafter random variation. The purpose was to galvanise political action in Parliament and Whitehall to tackle preventable causes of infant mortality. All subsequent official and almost all sociological categorisations have been based upon this initial model. New occupations are forced into these procrustean templates. This is problematic as it is by no means self-evident that the axial principles that underpinned the initial classification still hold.
Such occupational classifications all assume a hierarchy to these underlying measures; terms such as âgradientâ (see Marmot, 2004, 2015) or âladderâ (see Ipsos MediaCT, 2009) crop up regularly. Interestingly all ...