Peter Elias
University of Warwick
Institute for Employment Research
There seems little reason why it (the RGs social classification) should not be replaced by a
more flexible, discerning and less stultifyingly simplistic apparatus, for the purpose of primary
research ... This would reflect a more sophisticated form of social theory, and the more
powerful and advanced data handling resources that are available in the computer age.
Furthermore, its abandonment might usefully contribute to our emancipation from a particular
perception of British society... that we continue to live in a society which is an unchanging
hierarchy of five grades ... (Szreter, 1984. p. 538)
The classification (RGs social classification) is therefore here to stay. Like any statistical
instrument its use requires great care as well as an understanding and experience of it
limitations. Because it is commonly used in describing a complex society that is continuously
evolving and changing it has likewise to evolve and change, and this feature imposes a restraint
on any comparisons over a long period of years that are made using it. (Leete and Fox, 1977, p.7)
Abstract
This paper illustrates some of the problems associated with the current definition of social
class. These problems have arisen because of the explicit aim of those who have been
responsible for maintaining the social classification over the past fifty years or more,
that continuity should be preserved. While this is a laudable aim in its own right, it
fails to achieve the necessary balance required by the social statistician, to elaborate
statistical instruments such that they reflect the changes that are taking place within our
society. The pace of change is shown to have quickened over the last fifteen years. Related
changes in the skill structure of the employed population are not adequately reflected in the
social classification. The definition of Social Class, in terms of the underlying occupational
classification, does not take advantage of the redefinition of skill as operationalised in the
Standard Occupational Classification. The paper concludes that, if the social classification
is to be preserved, it must be updated. Information on the distribution of earnings within
occupations would seem to be the most appropriate mechanism for effecting this revision.
There are, however, new approaches to data coding being developed which will help to provide
continuity and facilitate changes in the classification. Occupational coding of the
1991 Census of Population was performed at a more detailed level than is published, yielding
a degree of ‘backward’ compatibility between the new occupational coding scheme and the previous
scheme. Via computerised text coding programmes, other data sets now embody a much more detailed
coding scheme which facilitates both forward and backward compatibility. Consideration should
be given to the introduction of these methods in the next Census of Population.
Introduction
Social Class based upon Occupation is the name now given to what was previously referred
to as the Registrar-General’s social classification of occupations, a classification which
facilitates the grouping of individuals or households according to the occupations people hold,
thereby creating five or six broad groups for statistical investigation of the conditions within
which people live, their health, housing and hygiene, their social and economic experience and
their attitudes, values and preferences. The class categories are currently defined (*1*) with
reference both to the occupations people hold and their status in employment(*2*).
Using information from the Censuses of Population for 1981 and 1991, this paper examines the
conceptual basis and definition of social class based upon occupation. By comparing and
contrasting social class categories with the Standard Occupational Classification, some
tentative conclusions are drawn about the current definition of social class categories in
terms of their fitness for purpose.
The following sections of this paper explore issues relevant to the definition of social class
categories. In particular, consideration is given to:
| · | the conceptual basis of social class based upon occupation; |
| · | the concept of skill and recent changes in the occupational and skill structure of the population; |
| · | the consequences of not revising social class in line with revisions made for the introduction of the Standard Occupational Classification; |
| · | the current allocation of occupational groups to social class categories. |
The conceptual basis of social class
Szreter (1984) and Leete and Fox (1977) trace the origins of the present social classification to
an 1887 address to the Royal Statistical Society given by the Assistant Registrar-General,
N A Humphreys, in which he called for research on mortality to be undertaken by grouping the
population into broad groups reflecting their social standing. By 1910 the medical statistician,
Dr THC Stevenson, had introduced the social classification of occupations based upon three
perceived grades of society: the upper, middle and lower classes. Eight classes were introduced
initially; five were formed from the three basic classes by defining two intermediate groups
between the three main groups, three additional categories were added covering textiles, mining
and agricultural workers who could not be allocated easily within the fivefold structure of the
main classification.
By 1920 the three ‘anomalous’ categories which lay outside the five main groups had been merged
within these groups according to the level of skill of the constituent occupations within the
categories. From this point on it is apparent that ‘skill’, variously defined as the complexity
of job-related tasks or the education or training required for entry into an occupation, became
more than just a name to cover categories within the middle range of the social classification,
but was being used as an important organising concept for the maintenance of the social
classification.
Various refinements to the classification have been made over the past seventy years, for example,
the introduction of employment status information (eg whether or not a person has supervisory
status) to assist in the allocation to a social class category within a particular
occupation group, and the splitting of the large and heterogeneous social class III (skilled
occupations) into non-manual and manual categories to facilitate the recombination of the
social classes into two broad groups, manual and non-manual occupations. Basically however,
the structure of the classification has remained unchanged.
Social class is a derived classification. Respondents to censuses and surveys are not asked
directly to indicate their social class status. Instead, social classification is achieved by
mapping from the occupation unit group, the most detailed level of coding of occupational
information, and status in employment to the social class category. For this reason, revision
of the underlying occupational classification usually precipitates a major revision to this
mapping.
The two latest revisions of the allocation of occupations to social class categories took place
in 1960 and 1980. No major revision took place in 1990 following introduction of the Standard
Occupational Classification (SOC).
The concept of skill and the Standard Occupational Classification
Social class categories are defined with reference to the skill content of the constituent
occupational unit groups, modified somewhat by information about status in employment.
Exactly how this process has been achieved during the various revisions to the social
classification has never been made explicit. In the process of developing the Standard
Occupational Classification, skill categories were established for the process of redefining
and aggregating occupation unit groups. Later in this paper, a cross-check is made between
the skill categories of the SOC and social class. Useful though this may appear, it raises
the issue of whether or not ‘skill’ is a meaningful concept for organising any classification
based upon occupational information.
The concept of ‘skill’ and the meaning of ‘skill acquisition’ have been the subject of intense
debate, particularly among sociologists, stimulated by Braverman’s (1974) thesis on the
relationships between technological and organisational change and deskilling. The debate
has widened to cover the social processes underlying the conceptualisation of skilled versus
unskilled work, particularly the roles of gender and power relationships both within and beyond
the workplace. Interestingly, among other academic groups who address labour market issues,
‘skill’ is often taken as a relatively fixed and determinable attribute of a particular job.
Skill is used in this context as a generic term which embodies notions of the technical
complexity of a job and the required knowledge and experience of the worker. Skill
acquisition and deskilling refer to the movement of individuals into and out of such jobs,
often associated with work-related training or as a result of loss of employment, through
redundancy for example.
A recent and detailed study of skill within the context of British labour markets drew upon a
variety of academic approaches to these concepts, yielding some useful and interesting insights
into both the meaning of skill and the impact of technological and organisational changes on
trends in skill development. Based upon the various studies conducted within the ESRC Social
Change and Economic Life Initiative (SCEL), Penn, Rose and Rubery (1994) summarise evidence
which indicates the double direction of skill change. Within most jobs covered by these
investigations, recent increases in skill and responsibility were reported by employees and
employers. On the other hand, many new jobs provided little opportunity for skill acquisition.
Women appeared to be faring worse in this process than men, particularly given the growth in
part-time jobs in the service sector of the economy (Gallie, 1994; Elias, 1994; Horrell et al.
1994). Men who became unemployed also appeared to have suffered significantly in terms of their
subsequent skill utilisation or skill development (Elias, 1994). Perceptions of skill, as derived
from participants (employees and employers) in this major study, indicated that it was not just
technical competence and the overall complexity of the job which were generally regarded as
indicative of skill, but personal attributes such as ‘ability’ were often stressed as a defining
characteristic of a skilled job (Burchell et al., 1994).
These studies indicate that the distribution of skill, like income, has widened in Britain over
the last decade or so. More importantly, they show that skill is a meaningful and measurable
concept. While people find some difficulty in disentangling the skill requirements of the job
in which they work from their own abilities (Rose, 1994), it appears to be the case that both
employers and employees can identify and agree upon the types of jobs which are skilled from
those which are less skilled and that skill is predominantly an acquired characteristic related
to a particular occupation and associated with a fairly lengthy period of training or studentship.
Monitoring changes in the skill structure of the population requires, therefore, that certain
‘skill’ groupings should be defined in a consistent fashion across a significant period of time.
The ‘skill groupings’ used in this section derive from the Standard Occupational Classification.
The conceptual basis of this classification rests upon the definition of skill levels and
skill specialisations. ‘Skill level’ refers primarily to the level of formal
qualifications required for a person to be recognised as fully competent for entry into the
occupation concerned. Where formal qualification requirements are not stated or have not yet
been recognised, reference is made to the duration of training and/or work experience normally
required for occupational competence. ‘Skill specialisations’ refer to the materials handled,
tools and equipment used, fields of knowledge required, etc.
The nine major groups of the SOC are built around four skill levels. These were not formally
specified in the definition of the SOC (OPCS, 1990), because of the possible scope for confusion
with National Vocational Qualification (NVQ) levels which were in the process of development at
the same time. Table 1 relates the four skill levels to the minor and sub-major group structure
of the SOC(*3*).
Table 1: Skill Levels, Major and Sub-major Groups of the Standard Occupational Classification
| Skill Level | Major or sub-major Groups | Constituent minor groups |
| Level 4 | Managers and administrators (exc. managers/proprietors in in agriculture and services) | 10, 11, 12, 13, 14, 15, 19 |
| | Professional occupations | 20, 21, 22, 23, 24, 25, 26, 27, 29 |
| Level 3 | Managers/proprietors in agriculture and services | 16, 17 |
| | Associate professional and technical occupations | 30, 31, 32, 33, 34, 35, 36, 37, 38, 39 |
| | Craft and related occupations | 50, 51, 52, 53, 54, 55, 56, 57, 58, 59 |
| | Buyers/brokers and sales representatives | 70, 71 |
| Level 2 | Clerical and secretarial occupations | 40, 41, 42, 43, 44, 45, 46, 49 |
| | Personal and protective | 60, 61, 62, 63, 64, 65, 66, 67, 69 |
| | Service occupations | |
| | Sales occupations (exc. buyers/brokers and sales representatives) | 72, 73, 79 |
| | Plant and machine operatives | 80, 81, 82, 83, 84, 85, 86, 87, 88, 89 |
| | Other occupations in agriculture, forestry and fishing | 90 |
| Level 1 | Other elementary occupations | 91, 92, 93, 94, 95, 99 |
At the highest level (level 4) are professional occupations, middle and senior managerial
occupations and specialist managerial jobs. These are occupations which require a degree or
equivalent or a corresponding amount of work experience. At level 3 are jobs requiring high-level
vocational education and training, but not equivalent to a degree. Level 2 encompasses a wide
range of occupations, all of which require a good standard of general education. Most
occupations classified at this level will also provide further work-based training to new
entrants. Level 1 occupations do not usually require formal educational qualifications but
may have an associated short period of formal experience-related training.
Trends in the skill structure of the population
Changes in the skill structure of the population arise as a consequence of organisational and
technological changes in the production of goods and services, coupled with changes in the
education and training of persons entering employment or in paid work. Distinguishing such
structural trends from changes in the social standing of particular occupations is difficult,
given that the social classification is often used for both purposes. Routh (1980) provides
some indication of these trends, using data sources which, it is claimed(*4), are compiled
separately from the social classification. His study of occupational change from 1911-71,
reproduced at Table 2 below, reveals the slow but steady nature of skill changes throughout
this period. The ‘Professional’ group, covering social class I and parts of II and III
increased very slowly between 1911 and 1931, then by about 0.2 percentage points per year
from 1951 to 1971.
Table 2: Occupational changes 1911-71
| % distribution of workforce | 1911 | 1921 | 1931 | 1951 | 1961 | 1971 |
| Professional | 4.1 | 4.5 | 4.6 | 6.6 | 9.0 | 11.1 |
| Employers, managers | 10.1 | 10.5 | 10.4 | 10.5 | 10.1 | 12.4 |
| Clerical | 4.8 | 6.7 | 7.0 | 10.7 | 12.7 | 13.9 |
| Foremen and manual | 81.0 | 78.3 | 78.1 | 72.2 | 68.1 | 62.6 |
Source: Routh, 1980, p.5.
The category ‘employers and managers’ (Social Class II plus certain occupations allocated to
Social Class I and Social Class III in Routh’s study) remained remarkably steady to 1961,
before growing at 0.2 per cent per annum from 1961 to 1971. The number of persons employed in
clerical occupations has shown a steady growth throughout the sixty year period covered by this
analysis.
For information on the changes taking place between 1971 and 1981, the various studies undertaken
by the Institute for Employment Research show that the strongest growth in occupational employment
was observed in health and welfare professions, estimated to have grown at just over 3 per cent
per annum, with the category ‘managers and administrators’ growing at 2.3 per cent per annum
(Lindley et al,. 1987).
For the period 1981-91 it is possible to reconstruct Census of Population data for 1981 on to
the Standard Occupational Classification. While this makes comparison with earlier years rather
difficult, it does provide a reasonably accurate representation of skill changes over the last
decade. Figure 1 shows the result of this comparison. Immediately it becomes apparent that
the group ‘managers and administrators’ has been expanding at a rate unsurpassed in any preceding
decade this century. Within this major group it is the sub-major group ‘corporate managers and
administrators’ which has shown most growth, expanding at a rate averaging 4-9 per cent per
annum. No other category has shown such remarkable growth, although some professional and
associate professional categories show growth rates averaging more than 3 per cent per annum
over the decade. Equally remarkable though is the extent of the decline in ‘craft and related’
occupations, mainly as a result of the fall in employment in the engineering sector.
It appears, therefore, as if the pace of occupational change has quickened during the 1980s,
with important repercussions in terms of the changing skill structure of the population.
Further insight into the nature of these changes requires a more detailed examination of
changes in the 1981-91 period.
Figure 1: Changes in Occupational Structure, 1981-91, United Kingdom: SOC Major Group
Source: Lindley and Wilson (1994)
Detailed changes in skill structure, 1981-91
Although some general indication of the broad changes in skill structure can be gained from the
preceding analysis, many users have a significant interest in the detail associated with
transformations. This has, with previous intercensal comparison, proved difficult due to
the discontinuities introduced with the revision to the classification of occupations. For
the 1981 and 1991 Censuses, special provision was made in an attempt to preserve some continuity
over this decade. First, the occupations were coded in the 1991 Census using a more detailed
version of the SOC which facilitates regrouping of occupations to the 1980 Classification of
Occupations (C080), the scheme which was used for coding occupations in the 1981 Census.
Second, a ½% sub-sample of the 1981 Census of Population for England and Wales was recoded
to the SOC. It is this latter resource which provides the detailed view of occupational
changes presented in this section.
Table 2 shows the ‘top-40’ SOC unit groups, so chosen in terms of the absolute percentage
change in employment in England and Wales between 1981 and 1991(*5*). In line with the evidence
shown at the level of SOC major groups, most of the SOC unit groups in the ‘Managers and
Administrators’ major group are areas of occupational growth, particularly in banking,
finance and computer systems. Interestingly, the unit group category for senior managers
demonstrates a decline, a result which must be viewed cautiously given the fact that this
group did not exist in the 1980 Classification of Occupations and could, therefore, be subject
to a higher degree of specification error than is the case with other, more clearly defined,
managerial functions.
Growth in the professional occupations is associated with the increase in employment in
management consultancy occupations, software engineering and the higher education
professions. In associate professional and technical occupations, underwriting and
investment analysis stand out as major occupational growth areas. In clerical occupations,
only ‘legal secretaries’ gain a place in the top-40 occupations defined in terms of the
percentage change in the 1981-91 period. In the ‘craft and related’ group of occupations,
computer engineers rank thirteenth overall. Considerable growth is also in evidence in
caring occupations in the ‘Personal and Protective Services’ major group. At the low-skilled
end of the spectrum, only shelf-fillers gain a place in this ‘top-40’.
Considering the implications of this evidence for changes in skill structure between 1981 and
1991, it is fairly clear that the net change in skill structure is towards the skilled end of
the spectrum. Evidence of a broadening of the distribution of employment by skill, as was
inferred from the SCEL studies, is not immediately apparent. In part, this may be due to the
fact that mobility around the skill distribution may have increased, with more movement between
different skill levels. While the net effect of changes in skill structure is to shift the
general level of skill in the employed population upward, offsetting movement may well arise
for particular groups, or for workers with specific employment histories (i.e. those experiencing
unemployment, breaks for family formation).
The growth in managerial occupations is remarkable. Confirmation of these trends was sought
from independent sources (Elias, 1995; Elias and Gregory 1994) from which it was concluded that
the growth exhibited in the period 1981-91 has been accelerating in the post-1991 period.
The direction and extent of this growth runs counter to popular debate about the ‘delayering’ of
management within organisations. Obviously, both growth and delayering could occur
simultaneously, as large organisations ‘downsize’ their workforces, with displaced middle
managers joining smaller growing businesses.
Managerial occupations: how real are the changes?
The strong growth in the numbers of persons employed in managerial occupations evidenced in the
preceding section raises questions about the definition of an occupation. This is particularly
relevant to the issue of social classification, in that any tendency towards occupational
descriptions which lead to an increase in the usage of managerial codes will tend to ‘inflate’
social class II at the expense of other categories. In other words, if people are more prone
to describe their work in managerial terms, without there being any ‘real’ or underlying change
in their work routines, employment status and earnings, the ensuing growth in social class II
will be artefactual. This was an issue explored in Elias, Wilson and Green (1995), from which
this section is drawn.
Evidence on the growth of managerial duties as distinct from managerial occupations is obtained
from the Labour Force Survey. According to the 1991 Labour Force Survey, approximately
three-quarters of the 4.1 million persons employed in the ‘Managers and Administrators’
major group of the SOC stated that they had managerial responsibilities. Of the quarter
who stated that they did not have managerial duties, most were managers/proprietors of small
businesses in agriculture or services. This group tended to regard themselves as self-employed.
Interestingly, approximately 0.8 million persons who were classified to occupations other than
managers and administrators stated that they had managerial responsibilities. Foremost among
these occupations were senior nursing staff, teachers and sales representatives.
Looking back to the 1981 Labour Force Survey, a similar question was used to determine whether
or not a person had managerial duties. This analysis confirms that the main growth over the
decade has been in those occupations which have a managerial title. However, the number of
persons outwith the managerial category who state that they have managerial responsibilities
has increased by 60 per cent. In summary, therefore, it appears to be the case that the growth
in managerial jobs is associated with a corresponding increase in the number of people reporting
that they have managerial duties. The managerial category would, in fact, be broader than it
currently is and would have appeared to have grown more rapidly if certain occupations in the
associate professional category were included on the basis of the managerial duties involved
in such jobs.
Do social class categories reflect changes in skill structure?
Given that the social class categories are, ostensibly, based upon the concept of skill, the
question posed in this section is whether or not the remarkable changes evidenced via the SOC
are reflected in the structure of social class. Figure 2 shows the distribution of the employed
population (males and females separately) by the six main social class categories. For females,
corresponding information for 1971 cannot be obtained, due to the fact that married women were
allocated to the social class category of their husband.
Figure 2: The Social Class structure of the economically active population. Great Britain: males, 1971-91; females 1981-91
Significant change is reflected in Class II, managerial and technical occupations, together with
a small decline in class III N and each of the manual class categories(*6*). At face value,
therefore, social class appears to reflect the changes discussed in the preceding section.
Yet the changes do not appear to reflect the sharp changes in occupational structure which
have become apparent in the post 1981 period. For males it appears as if the 1971-81 changes
are simply replicated in the 1981-91 period. This, it will be argued, is a consequence both
of the application of the maximum continuity rule and a failure to update the allocation of
occupations to social class categories.
The maximum continuity rule (see OPCS, 1991, p.10) was developed as a method for dealing with
the consequences of creating new occupational groups in the SOC which could straddle social
class categories. Each of the new occupation unit groups was allocated to the same social
class category as that to which they would have been allocated in the Classification of
Occupations, 1980. Two interesting examples showing the operation of this rule can be
drawn from nursing occupations and cleaning occupations(*7*).
For the former group, and in line with the conceptual basis of the SOC, the relatively
diverse group of nursing occupations was split into three groups: Nurses were defined as
occupations associated with patient care and for which a qualification at level 3 was a
prerequisite. This led to the definition of the SOC unit group 340 within the major group
of associate professional and technical occupations. ‘Assistant nurses and nursing auxiliaries’
were defined as occupations associated with patient care for which a qualification at level 2
only was required. This category (SOC 640) was allocated within the caring occupations
classified in major group 6 of the SOC ‘Personal and protective service occupations’. A
third and relatively minor category was created for dental nurses.
Using the recoded ½% sub-sample of the 1981 Census of Population in conjunction with the
10% sample of the 1981 Census of Population and the component coding scheme used for the
1991 10% sample, it is possible to analyse changes in employment by either C080 or SOC.
Figures 3, 4 and 5 show the results of this exercise for these nursing occupational categories.
For the broad and heterogeneous C080 category 'Nurse administrators and nurses' Figure 3
shows that their total employment in England and Wales grew by 13.8 per cent between 1981
and 1991 . The rates of growth for males and females were 17.2 and 13.5 per cent respectively,
with males constituting approximately 10 per cent of the occupational category. However,
analysis of these changes by the new SOC categories (Figures 4 and 5 below) shows that by far
the major part of this growth was focused in the category 'Assistant nurses and nursing
auxiliaries'. Total employment change in this occupational category is 34.5 per cent over
the decade, compared with only 4.0 per cent for the category 'nurses'. In other words, the
above average rate of growth of the broader C080 nursing category arises because of the
remarkable rate of growth of assistant nurses and nursing auxiliaries.
In the C080, 'nurse administrators and nurses' were allocated to Social Class II 'Managerial
and technical occupations'. Using the maximum-continuity rule, the derived SOC categories
'nurses', 'assistant nurses and nursing auxiliaries' and 'dental nurses' are all allocated to
the same social classifications. Given that the major changes that have taken place have arisen
in the assistant/auxiliary nurse category, it is obvious that allocation of this new group
elsewhere in the social classifications would indicate change in social structure. The
present relationship between the SOC and the social classification does not permit this.
The second example concerns cleaning occupations. For this group of occupations, comprising
the C080 categories 'other domestic and school helpers' and 'cleaners, window cleaners, chimney
sweeps and road sweepers', a number of reclassifications were introduced. In particular, job
titles such as 'domestic' found in the former category were equated with 'cleaner' found in the
latter category. Together, these large components of the two C080 categories were merged to form
the SOC category 'cleaners'. Figures 6, 7 and 8 illustrate the percentage change in employment
for associated C080 and SOC categories between 1981 and 1991.
Figure 3: Changes in Nursing Occupations as defined in C080 (Nurse administrators and nurse)
1981-1991 England and Wales
Figure 4: Changes in Nursing Occupations as defined in the SOC (Nurses):1981-1991, England and Wales
Figure 5: Changes in Nursing Occupations as defined in the SOC (Assistant nurses and nursing auxiliaries):
1981-1991, England and Wales
Figure 6: Changes in Cleaning Occupations, 1981-1991 England and Wales: C080
Figure 7: Changes in Cleaning Occupations, 1981-1991 England and Wales: C080
Figure 8: Changes in Cleaning Occupations, 1981-1991 England and Wales: SOC
In total, employment is seen to decline in both C080 categories. In the case of 'cleaners,
window cleaners, chimney sweeps and road sweepers' the overall decline is the result of a
strong decline in female employment being partially offset by growth in male employment.
For the SOC category the distinction between male and female occupational change becomes
even sharper. The number of females working in cleaning occupations declines by 25 per
cent, contrasted with growth of 14.3 per cent for males.
In the C080 'other domestic and school helpers' were previously allocated to Social Class
IV 'Partly skilled occupations' whereas the category 'cleaners, window cleaners, chimney
sweeps and road sweepers' were allocated to Social Class V 'Unskilled occupations'. This
presented a problem for the mapping from SOC to the social classifications (see pp15-16,
Standard Occupational Classification Volume 3), resulting in a decision to allocate the
whole of the new SOC occupational category 'cleaners' to Social Class V. This gives rise
to some discontinuity between the 1981 and 1991 interpretations of the social classifications.
However, the fact that male and female employment changes are in such marked contrast between
1981 and 1991 does suggest that the new SOC category may not be as homogeneous as was intended.
If the decline in female employment is a result of the continuing decline in domestic services,
whereas the growth in male employment is indicative of the expansion of office cleaning services,
it is arguable that such changes are indicative of the workings of quite different social and
economic forces and should be recognised as such in the social classifications.
In the current allocation of occupational groups to social class categories out of date?
Social class categories are determined by reference to Standard Occupational Classification
unit group codes and information concerning status in employment (self employed, manager,
foreman/supervisor or employee/apprentice/trainee . This section examines and queries the
allocation of particular SOC unit groups to social class categories.
The relationship between social class categories and the SOC is of interest for a number of
reasons. First, if occupation remains as one of the basic building-blocks of more aggregated
social classifications, it is relevant to examine the current allocation to determine how
effectively the criteria underlying the allocation of occupation groups to these broader social
categories have been operationalised. The current allocation of SOC unit groups to social class
categories was effectively ‘rolled-over’ from the 1980 Classification of Occupations, which in
turn brought forward the tradition of social class banding of occupations developed in the
interwar period. There is good reason, therefore, to suspect that some of these allocations
may now appear inappropriate. Secondly, the SOC itself has an hierarchical structure which
competes with social class as a broad categorisation of occupational structure. It is relevant,
therefore, to address the issue of how much additional information the social class
categories yield over and above the major group structure of the SOC.
For 1991 we can investigate the social class structure of each major and sub-major group of the
Standard Occupational Classification. Given that the major and sub-major groups of the SOC are
formed from groupings of occupations which are deemed similar in terms of the education,
qualifications, training and work experience required for competent performance of their
associated tasks and that social class categories are supposedly hierarchical with respect
to ‘skill’, this exercise should reveal a reasonably close correspondence between social class
categories and sub-major groups of the SOC.
Looking first at two sub-major groups within SOC Major Group 1 (managers and administrators),
the information presented in Figures 9a and 9b shows that this is generally the case. Nearly
all of employment classified to sub-major group 1a (corporate managers and administrators) is
categorised to social class II (managerial and technical occupations). In sub-major group 1b
(managers/proprietors in agriculture and services), about 10 per cent of employed persons are
classified to social class III N (skilled occupations-non manual). This arises because self
employed persons in SOC unit groups 172, 174, 176 and 178 (proprietors of hairdressing,
restaurants, other catering, entertainment and sports establishments and the proprietors of
butchers and fishmongers establishments) are classified to III N, whereas all other proprietors
(eg of garages, hotels, bed and breakfast, pubs, clubs and other retail establishments) are
classified to social class II.
Figure 9a: The Social Class structure of SOC sub-major group ‘Corporate ‘Managers and Administrators’
Figure 9b: The Social Class structure of SOC sub-major group ‘Managers/proprietors in Agriculture and Service’
SOC Major Group 2 is comprised of four sub-major groups: 2a (science and engineering
professionals), 2b (health professionals), 2c (teaching professionals) and 2d (other
professional occupations). For sub-major groups 2a and 2b, there is a complete
correspondence between the SOC sub-major group and the corresponding social class
category (category I - professional occupations). For teaching professionals, these
occupations split between social class categories I and II. This arises because university
academic staff and education officers, school inspectors are classed as ‘professional’ whereas
HE and FE academic staff and all school teachers are classified to social class II (managerial
and technical). In sub-major group 2d (other professional occupations), librarians, archivists
and curators are classified to social class II, as are social workers.
Figure 10a: The Social Class structure of SOC sub-major group ‘Science and Engineering Professionals’
Figure 10b: The Social Class structure of SOC sub-major group ‘Health Professionals’
Figure 10c: The Social Class structure of SOC sub-major group ‘Teaching Professionals’
Figure 10d: The Social Class structure of SOC sub-major group ‘Other Professional Occupations’
For Major Group 3 (associate professional and technical occupations) it would be expected that the
majority of occupations within this major group should be classified to social class II (managerial
and technical occupations). This is apparent from Figures 11a to 11c, but there are some
interesting exceptions. Draughtspersons are classified to social class category III N, as
are camera, sound and video equipment operators, professional athletes and driving instructors.
Figure 11a: The Social Class structure of SOC sub-major group ‘Science and Engineering Associate Professionals’
Figure 11b: The Social Class structure of SOC sub-major group ‘Health Associate Professionals’
Figure 11c: The Social Class structure of SOC sub-major group ‘Other Associate Professional Occupations’
In Major Group 4 (clerical and secretarial occupations) it can be seen from Figures 12a and 12b
that most employed persons are allocated to social class category III N (non-manual skilled
occupations), except for clerical jobs in warehouse/stores record keeping. Employees in this
occupational area are classified to social class category IV (partly skilled occupations) and
foremen to social class category III M (skilled manual occupations). Telephone operators are
classed to category IV (partly skilled occupations, whereas other office communication system
operators and receptionist/telephonists are social class III N.
Figure 12a: The Social Class structure of SOC sub-major group ‘Clerical Occupations’
Figure 12b: The Social Class structure of SOC sub-major group ‘Secretarial Occupations’
Occupations falling within Major Group 5 (craft and related occupations) should be classified
to III M (skilled manual occupations). This is apparent for approximately 80 per cent of
employment in this occupation major group (see Figures 13a, 13b and 13c), but again some
exceptions are noteworthy. In the construction trades, roofers, tilers, glaziers and
steeplejacks are placed in social class IV. In the skilled engineering trades, lathe and
press setter-operators are placed in III M whereas boring, grinding and milling machine
operators are allocated to IV. In other skilled trades, weaving and knitting machine
operators are social class III M but sewing machine operators are social class IV.
Figure 13a: The Social Class structure of SOC sub-major group ‘Skilled Construction Trades’
Figure 13b: The Social Class structure of SOC sub-major group ‘Skilled Engineering Trades’
Figure 13c: The Social Class structure of SOC sub-major group ‘Other Skilled Trades’
Major Group 6 (personal and protective service occupations) divides into two distinct sub-major
groups. Protective service occupations (fire, police, prison service and customs/excise officers)
are allocated between social class categories III N and IV; police and fire officers are classed
as III N, prison service officers as IV. Customs and excise officers (the occupational category
which requires the least training) are allocated to social class II. In the personal service
group of occupations, nursing auxiliaries and dental nurses are allocated to social class II,
whereas ambulance staff, nursery nurses and playgroup leaders to III M, and educational
assistants to social class IV. Other apparent anomalies include the classification of
hairdressers to III M and beauticians to IV.
Figure 14a: The Social Class structure of SOC sub-major group ‘Protective Service Occupations’
Figure 14b: The Social Class structure of SOC sub-major group ‘Personal Service Occupations’
Major Group 8 (plant and machine operatives) includes a significant number of SOC unit groups
for which the various social class allocations are not obvious. For example, tannery production
operatives are classed as III N whereas textile processing operatives are classed as IV.
Similarly operatives of machines and plant in glass, ceramics, paper products and rubber
are allocated to III M, with operatives of machines and plant in chemicals, gas and
petroleum to social class IV.
Figure 15a: The Social Class structure of SOC sub-major group ‘Buyers, Brokers and Sales Representatives’
Figure 15b: The Social Class structure of SOC sub-major group ‘Other Sales Occupations’
Figure 16a: The Social Class structure of SOC sub-major group ‘Industrial Plant and Machine Operators, Assemblers’
Figure 16b: The Social Class structure of SOC sub-major group ‘Drivers and Mobile Machine Operators’
Major Group 9, which covers elementary occupations would be expected to have a close relationship with social class category V (unskilled occupations). This is clearly not the case for agriculture workers (see Figure 17a), but the position of this group within the SOC was recognised as problematic. For non-agricultural elementary occupations (see Figure 17b) a variety of allocations are shown for constituent unit groups: ‘mates’ to carpenters are allocated to III M whereas ‘mates’ to builders are allocated to IV. Paviors and kerb layers are allocated to IV but other road construction workers are allocated to V. Slingers are allocated to III M but dockers are allocated to V. Postal workers are allocated to social class IV, whereas mail couriers are allocated to V. Kitchen hands are social class V, but counter hands are social class IV.
Figure 17a: The Social Class structure of SOC sub-major group ‘Other Occupations in Agriculture, Forestry and Fishing’
Figure 17b: The Social Class structure of SOC sub-major group ‘Other Elementary Occupations’
Conclusion
This paper has examined the conceptual basis and the structure of the classification known as
Social Class based upon occupation. In comparison with the skill-based Standard Occupational
Classification a large number of anomalies are in evidence. In part, these stem from the
maximum continuity rule, a procedure which was adopted to preserve the continuity of social
class defined with respect to the 1980 Classification of Occupations. This rule had two
consequences. First, it meant that the new distinctions made in the SOC between groups of
occupations on the basis of skill could not be differentiated within the social classification,
a schema which purports to distinguish occupations on the basis of skill. Second, it served to
perpetuate many of the allocations of occupation unit groups to social class categories which
existed within the 1980 Classification of Occupations many of which now appear puzzling.
Revision of the social classification is now required if this classification is to remain as a
useful method of aggregating occupations with reference to skill levels. This will undoubtedly
be the case, even though the Standard Occupational Classification provides an alternative set of
broad categories which are essentially hierarchical with respect to skill. As Leete and Fox
(1977) predicted, social class categories will continue to be required by social statisticians,
because they provide a simple mechanism for condensing the complexities of occupational structure
to a differentiated set of social groups. But if social class is to be revised, this raises the
important issue of how it should be revised. Simply allocating occupations to class categories
on the basis of some subjective notions about the skill content of occupations within unit groups
may appear as illogical to social statisticians in 2015 as the current allocation of many
occupations to social class categories appears today. What is required is a more rigorous
and quantitative approach to the method through which this allocation is established. In this
respect, good use can be made of the fact that the Standard Occupational Classification has, as
its name implied, become the standard for all official surveys, including the New Earnings Survey,
Labour Force Survey, General Household Survey, Family Expenditure Survey and for other research
resources which continue to develop (notably the British Household Panel Study). It is proposed,
therefore, that a more detailed analysis of the distribution of earnings within SOC unit groups
should be undertaken. If the labour market rewards skills, it would seem appropriate to use
information on the distribution of earnings by occupation to assist in the identification of
skill levels and the allocation of SOC unit groups to broad skill categories. By this method,
social class based upon occupation would reflect significant income differences between the
social groupings, making it a more powerful tool for social, demographic and medical research.
All of which leaves Szreter’s (1984) point, that there is a need for greater flexibility in
the production of information categorised into social grades, and that modern data handling
techniques should facilitate this. At the time this remark was made, such a development was
not feasible. But recent advances in computer-assisted and computer-automated coding are now
moving rapidly to the point where semi-automated coding of textual descriptions of occupations
will be feasible. When this occurs, coding will relate to an index reference position for a
particular combination of words. If these data are stored as part of a census or survey
record, recombination of these codes to any future new version of the social classification
becomes trivial. The problem of retaining continuity yet maintaining a statistical instrument
as a modern tool of social analysis will have been solved.
References
Burchell, B., J. Elliott, J. Rubery and F. Wilkinson (1994). ‘Management and Employee Perceptions
of Skill’. Ch. 6 of Penn, Rose and Rubery.
Elias, P. (1994). ‘Occupational Change in a Working Life Perspective: Internal and External
Views’. Ch.3 of Penn, Rose and Rubery.
Elias, P. and M. Gregory (1994). The Changing Structure of Occupations and Earnings in Great
Britain, 1975-1990. Research Series No. 27. London : Employment Department.
Elias, P. (1995). ‘Managers and Administrators’. Ch. 1 of Occupational Studies, 1995.
P. Elias, R. Wilson and A. Green (eds.). Coventry: Institute for Employment Research,
University of Warwick (forthcoming).
Elias, P., R. Wilson and A. Green (1995). Review of the Economy and Employment 1994.
Occupational Studies: Part 1. Coventry: Institute for Employment Research, University
of Warwick.
Elias, P. and T. Staples (1995). Occupational Definition and Coding Conventions, 1981 and
1991 Censuses of Population. London: HMSO (forthcoming).
Gallie, D. (1994). ‘Patterns of Skill Change: Upskilling, Deskilling or Polarization?’ Ch 2
of Penn, Rose and Rubery.
Horrell, S. J. Rubery and B. Burchell (1994). ‘Gender and Skills’. Ch. 7 of Penn, Rose and
Rubery.
Lindley, R.M., R. Wilson, D. Bosworth and P. Elias (1987). Review of the Economy and
Employment 1987. Coventry: Institute for Employment Research, University of Warwick.
Leete, R. and J. Fox (1977). ‘Registrar-General’s social clauses: origins and users;
Population Trends, 8, 1-7.
Lindley, R.M. and R. Wilson. Review of the Economy and Employment: Occupational Assessment.
Coventry: Institute for Employment Research, University of Warwick.
Office of Population Censuses and Surveys (1990). Standard Occupational Classification:
Volume 1. London: Her Majesty’s Stationery Office.
Office of Population Censuses and Surveys (1991). Standard Occupational Classification:
Volume 3. London: Her Majesty’s Stationery Office.
Penn, R., M. Rose and J. Rubery (1994). Skill and Occupational Change. Oxford: Oxford
University Press.
Rose, M. (1994). ‘Job Satisfaction, Job Skills and Personal Skills’. Ch.9 of Penn, Rose
and Rubery.
Szreter, S.R.S. (1984). ‘The genesis of the Registrar-General: social classification of
occupations’. British Journal of Sociology, 35, 4, 522-46.
Routh, G. (1980). Occupation and Pay in Great Britain 1906-79. London: MacMillan.
Acknowledgements
This paper utilises source material from the Census of Population 1981 and 1991, gathered
together as part of the development work for the Standard Occupational Classification and
under the ESRC/ISC Census Development Programme. Tessa Staples at the Office of Population
Censuses and Surveys, Occupation Information Unit, has given generously of her time within
both of these work programmes. She also assisted with the production of comparable social
class distributions for 1981 and 1991.
Lynne Conaghan has done an excellent job in integrating the graphics and text in this document.
Thanks are also due to Jane Worton for re-formatting this paper for inclusion as an appendix to
the overall Report on Phase 1 of the ESRC Review of OPCS Social Classifications.