Issue 9	July 1995

Social Research Update is published quarterly by the Department of Sociology, University of Surrey, Guildford GU2 7XH, England. Subscriptions for the hardcopy version are free to researchers with addresses in the UK. Apply by email to sru@soc.surrey.ac.uk.

Official Social Classifications in the UK

David Rose

David Rose is Associate Director and Professor in the ESRC Research Centre on Micro-social Change, University of Essex. He is also Director of the Essex Institute for the Social Sciences. Currently he is convening the ESRC Review of OPCS Social Classifications. He is joint author of Property, Paternalism and Power (1978), Social Class in Modern Britain (1988) and Changing Households (1994) and edited Social Stratification and Economic Change (1988).

The practice of officially classifying the British population according to occupation and industry began in 1851. The occupational element was gradually increased from 1881, and in 1887 the idea was first mooted by the Assistant Registrar General that, for mortality analyses, the population might be divided into broad groups based on social standing. However, it was only in 1911 that government recognised that the concepts of occupation and industry were distinct and so included separate questions on each in that year's Census. Although at this time it was still not thought possible to classify occupations without reference to industry, we can take 1911 as a watershed year in the history of official UK social classifications based on occupations. The Registrar General's Annual Report for 1911 (not published until 1913) included, without any fanfare, a summary of occupations designed to represent 'social grades'. These were later referred to as 'social classes' and were used for the analysis of mortality data and the 1911 Census tables on Fertility of Marriage. Thus were inaugurated the Registrar General's Social Classes (RGSC), re-named in 1990, Social Class based on Occupation.

The history of social class schemes

The person responsible for devising the social class scheme was T H C Stevenson, a medical statistician in the General Register Office. His 1913 classification mixed occupational and industrial groups. While Stevenson conceived society as divided into three basic social classes (the upper, middle and working classes), in fact he produced an eight-fold classification by introducing intermediate classes between the upper and middle classes and between the middle and working classes; and adding three industrial groups for those working in mining, textiles and agriculture.

In 1921 there was a major revision of the class scheme. This was made possible by the introduction of the first proper classification of occupations. The three industrial social classes were re-allocated between the other classes and the new five class scheme was used for the analysis of infant and occupational mortality and fertility. Not only did this revision give stronger emphasis to 'skill', but there is persuasive evidence that the revision was constructed in the light of knowledge of mortality rates. Thereby it produced the mortality gradients so long cherished by those who use RGSC for this purpose.

In 1928, Stevenson gave the first indication of the conceptual basis of his scheme. In a paper on 'The Vital Statistics of Wealth and Poverty' read to the Royal Statistical Society, Stevenson noted that 'culture' was more important than material factors in explaining the lower mortality of the 'wealthier classes'. And 'culture' (which for Stevenson included knowledge of health and hygiene issues) was more easily equated to occupation than to income and wealth. Hence Stevenson inferred social position from occupation as an indicator of 'culture'. He admitted that the assignment of occupations to classes was largely an empirical matter and thus dependent on individual judgement. The validation of the scheme could be seen in the results it produced, for example the mortality gradients associated with the scheme which showed a uniform increase as one moved down the social scale.

Thus, the Registrar-General's class scheme rested on the assumption that society is a graded hierarchy of occupations. The five basic social classes recognised by the Office of Population Censuses and Surveys (OPCS) were described, from 1921 to 1971, as an ordinal classification of occupations according to their reputed 'standing within the community'. In 1980, this was changed so that social class was equated directly with occupational skill. Unfortunately, OPCS did not explain the principles behind this reconceptualization. Since only about 7 per cent of cases were awarded different class codes when the same data were coded according to both 1970 and 1980 procedures and then cross-classified, it seems that, in practice, the changes in mechanism for allocating occupations to classes achieved little beyond further obfuscation of the already somewhat unclear class categories themselves. The 1990 Standard Occupational Classification (SOC) is much more explicitly related to skill in the sense of competence required to perform in an occupation. Again the introduction of SOC affected the allocation of occupations to classes, although its full effect was deliberately minimised (see Elias, 1995).

However, these are by no means the only changes made to the official social class scheme since 1921. At every subsequent Census changes have been implemented not only in the method of classifying occupations but also in the allocation of particular occupations to social classes. These make time-series comparisons extremely difficult. In 1931, for example, male clerks were demoted from class II to class III; in 1960 airline pilots were promoted from class III to class II; and in 1961 postmen and telephone operators were demoted from class II to class IV. These piecemeal changes have important cumulative effects, yet have to be made because of social and employment change.

Allocating people to social classes

While individual occupations have been reallocated to different classes, the overall shape of the model has changed very little during the past sixty years. The classes are now described as follows:

I

Professional etc occupations

II

Managerial and Technical occupations

III

Skilled occupations

(N)

non-manual

(M)

manual

IV

Partly-skilled occupations

V

Unskilled occupations

It should already be apparent that social class is a derived classification achieved by mapping occupation and employment status to class categories. But where do the raw data come from? The principal source of reliable national data comes from the decennial Population Census. Census data on occupation, employer and employment status are collected for the whole working population and 10 per cent of these data are then coded and used in analyses. Between Censuses, government social surveys also collect data required for social classifications, as do many academic studies. In addition valuable data for health and medical researchers are derived from death registration records which include the occupations of deceased persons.

In practice individuals are assigned to social classes by a threefold process. First, they are allocated an occupational group, defined according to the kind of work done and the nature of the operation performed. Each occupational category is then assigned as a whole to one or other social class and no account is taken of differences between individuals in the same occupation group, e.g. differences of education or level of remuneration. Finally, persons of particular employment status within occupational groups are removed to social classes different from that allocated the occupation as a whole. Most notably, individuals of foreman status whose basic social class is IV or V are reallocated to social class III, and persons of managerial status are (with certain minor exceptions) placed in social class II (see OPCS, 1980:vi and xi; OPCS, 1991:12).

Ultimately, however, occupations are placed in social classes on the basis of judgements made by the Registrar-General's staff and various other experts whom they consult, and not in accordance with any coherent body of social theory. This is why the classification has rightly been described as an intuitive or a priori scale. This does not mean, however, that no theory lay at the basis of Stevenson's scheme. His notion of 'culture' as embodied by occupation did depend upon a definite view of social processes. We can see this when we examine why the scheme was deemed necessary.

Leete and Fox (1977) note that the social class scheme was designed to analyse infant mortality, although Boston (1984) reports Stevenson's primary interest as being in the study of fertility, and especially the then vexed question as to whether the upper and middle classes were reproducing themselves. Clearly, the two issues are related, but, as Szreter (1985) shows, they must also be seen as central to one of the earliest debates in UK social science - that between the eugenicists and the environmentalists.

Although the first published application of the class schema in 1913 was to the interpretation of infant mortality statistics, the real inspiration for its construction came from the nineteenth-century debate about differential fertility, between hereditarian eugenicists on the one hand and environmentalists on the other. Stevenson, an environmentalist and advocate of interventionist public health measures, developed the social class scheme in order to test and disprove eugenicist theories, Thus, while there is no explicit sociological theory to support the scheme, it is nevertheless an assertion of the importance of social factors in accounting for observed trends in fertility and mortality.

For many sociologists, SEG is seen as a better measure than social class for social scientific purposes. Its seventeen groups can be collapsed to produce a scheme not dissimilar from that proposed by Goldthorpe and his associates for their studies of social mobility and class structure (see Erikson and Goldthorpe, 1992). The reason for this similarity is apparent: SEG is a measure of employment status (not 'skill' or 'social standing') and thus speaks theory without knowing it. The fact that the SEG scheme was originally proposed to government (in 1950) by David Glass, Goldthorpe's intellectual predecessor in the mobility field, goes far in accounting for its social scientific potential.

Thus we have two official ways of classifying people in terms of their employment, based on different principles and conceptions and without any straightforward mapping between them. Given this fact, and given also that the two schemes have remained largely unaltered for decades while society itself has changed rapidly, OPCS recently requested the Economic and Social Research Council to undertake a review of RGSC and SEG. Not only was it thought that two occupationally-based social classifications might be one too many, but increasingly government departments and academic researchers have bemoaned the fact that the classifications can ignore the 40 per cent of the population who are not in paid employment.

The ESRC Review is in two phases. Phase 1, now completed, examined whether there was a continuing need for official social classifications; whether the current ones needed revision; and, if so, how that revision might be achieved. Phase 2, now underway, is charged with the task of producing, validating and testing a single, revised social classification which could be used from the time of the next Census (see Rose, 1995).

The ESRC review: a summary

The following is a summary of the key findings of Phase 1 of the ESRC Review. First, there is complete unanimity across central and local government, academics and the private sector that OPCS social classifications are necessary. Central government departments need them because they provide convenient summaries of complex data relevant to the analysis of social variation and thus to policy formulation, targeting and evaluation, as well as needs assessment. The many area classifications and indices of deprivation used by government departments in determining resource allocation include elements of the OPCS classifications. The classifications are also essential for monitoring the health of the population. In the private sector OPCS classifications are a vital element in the creation of area classifications by companies in the market analysis field. Academic researchers need the classifications for scientific analyses, especially in health, medical, geographic and demographic research.

Use of the classifications across government departments is widespread and multiple. Apparently for conventional reasons some departments favour use of RGSC and others of SEG. As an indication of the importance of the classifications to users, they are featured by request in more than thirty 1991 Census tabulations. For both government departments and academic users, the long time-series provided by RGSC in particular is of great value both in the interpretation of social trends and in policy evaluation.

Not least among the advantages of OPCS producing a standard government classification is that this leaves government in control of definitions and hence of the information which must be collected in order to produce classifications. Since OPCS collects the building block information and processes it, it is only sensible that OPCS should also determine how data are classified and thereby provide an approved standard for use in all government departments.

The current classifications have problems especially in terms of their obscure conceptual basis and limited population coverage. Nevertheless, for both pragmatic and theoretical reasons, occupationally-based classifications will continue to remain vital tools for scientific and policy analyses for the foreseeable future. This is because, pragmatically, they are based on routinely and widely-collected data and, theoretically, because it remains the case that a person's employment situation is a key determinant of life chances.

Among the key issues to be faced in Phase 2 of the Review is likely to be that of level of measurement for social classifications. The current RGSC is ordinal; SEG is a nominal measure. While there is considerable support in government departments for continuing to have an ordinal measure, many academic experts disagree. Among academics there is support for a nominal measure from those who, like Goldthorpe, would argue that since the categories of employer, self-employed and employee are central to any sociological class scheme (and to SEG), it is impossible to have ordered categories. Arguments that nominal measures are inherently less adequate than ordinal measures and cannot sustain serious statistical analysis are dismissed as being outmoded.

On the other hand there are those who believe it is possible and desirable to create continuous measures of social stratification. Advocates of the Cambridge Scale take this view, for example. A continuous scale may be divided into classes, of course, but a scale is claimed to be a more flexible tool with which to work.

Both the Goldthorpe class scheme and the Cambridge Scale are, of course, potential alternatives to RGSC and SEG and as such they are widely used in academic (but not government) research. Nevertheless, the reasons these schemes were created in part reflect dissatisfaction with RGSC in particular on theoretical, conceptual and technical grounds. The Cambridge Scale (CS) goes further in its implied critique, since its authors reject the whole idea of class analysis. Based on the scaling of survey respondents' occupational friendship and marriage scores, the CS is regarded by its originators as a broad measure of social stratification and social inequality. Ultimately the scale measures the market outcomes of different jobs and the lifestyle associated with them. It is not an attempt to measure the social structure and the way this creates different market capacities in different sections of the population. Indeed, the theoretical position of the authors of the CS rejects class analysis on the grounds that it is a static approach to what are fundamentally problems relating to social dynamics. Nor is CS a status scale. It is a measure of lifestyle determined by social experience and, ultimately therefore, significant social processes. It is designed to unite key features of both the social and the economic; and it raises questions about any attempt to analyse social inequality in terms of categorical measures (see Prandy, 1990).

While RGSC in particular has many problems which need to be resolved, we should not forget that it has been a very successful empirical tool whose use has had profound influence on social policy for several decades. Especially when we consider that it was created in the context of a debate between eugenicists and environmentalists in a time before serious theoretical social science had emerged in Britain, it is not surprising if the RGSC is now considered inadequate by many academic researchers. And beyond the problems already discussed, the revision of government social classifications will have to be considered within the framework of current intellectual debates surrounding class analysis. Principal among these, of course, are the unit of analysis problem, the relationship between gender and class, the appropriate reference occupation for the retired, the unemployed and other significant economically inactive groups, and the even thornier problem of the very relevance of class analysis to contemporary society and social scientific explanation. However, whatever the problems, it is intended that Britain should have a new social classification more fitted to the needs of the twenty-first century.

Bibilography

Social Research Update is published by:

Edited by Nigel Gilbert.

July 1995 © University of Surrey

Permission is granted to reproduce this issue of Social Research Update provided that no charge is made other than for the cost of reproduction and this panel acknowledging copyright is included with all copies.