Notes on: Savage, M. Devine, F., Cunningham, N., Taylor, M.,Li. Y., Hjellbrekke, J., Le Roux, B ., Friedman, S., & Miles, A.  (2013) 'A New Model of Social Class? Findings from the BBC's Great British Class Survey Experiment'. Sociology 47 (2): 219--50.

Dave Harris

Escalating social inequalities have led to a new interest in the analysis of social class especially '"cultural class analysis"'derived from Bourdieu and some feminist work (220).  In this new model, economic cultural and social capital is combined and applied to the UK.  The BBC survey is the largest survey of social class ever conducted 161,400 respondents to a web survey, for combined with 'a parallel national representatives survey'.  Overall, the established middle class, or service class, and the traditional working class persist, but there are five other classes as well, some of which show increased social polarisation.

Originally, class was studied using the Registrar General's schema, based on '"standing within the community"' (221), and then "skill".  Six classes resulted with professionals at the top and unskilled manual workers at the bottom.  However, this was not adequately informed by social theory, and this led to the development of the Goldthorpe scale [technically  the 'Erikson-Goldthorpe -PortoCaero'model] which was able to successfully challenge a 'rival Marxist framework of Erik Olin Wright'.  Here there were seven classes defined according to employment position, distinguishing between employees and employers, and between those on labour contracts and those in a more service relationship with their employers like professionals and managers.  This model led to the National Statistics Socio - Economic Classification scheme as an official definition.

However, that model can itself be criticised, despite its successes, especially in analyzing social mobility:

  • It focuses on 'criterion validity'(222), measuring 'features of the employment relations which are held to define class relationships'.  This misses out the wider cultural and social activities and identities which do not link closely to class position, as in patterns of cultural consumption.
  • It works pragmatically as a way of managing 'standard nationally representative surveys with a moderate sample size' -its preferred form of analysis, log-linear modeling, requires 'reasonable cell sizes', and this was apparently the reason why the study did not separate out an elite within the service class.  Larger sample sizes do show differences within the classes.
  • Classes were still based on measures of income and wealth, and this has been criticised by those who want to examine 'moves between income groups, rather than occupational classes' as a form of social mobility.  This will become important if 'income variation within occupations is growing'.
  • Feminist critique has pointed to important symbolic and cultural aspects of class and the way it works on 'forms of stigmatisation and marking of personhood and value'

Recent studies of social mobility has challenged the validity of the scheme, again for taking salaried middle class as too homogeneous a group, and for over stressing the divide between manual and non manual occupations.  In particular, there seemed to be substantial 'cross national differences with respect to qualification levels, job autonomy, career prospects…  organisation of production, etc.' (223), providing different 'occupational realities' in different countries, say between skilled worker and supervisor ['two formally identical categories']

Bourdieu has been influential in thinking of the new scheme, based on combinations of economic capital, cultural capital, and social capital.  Different stocks of each of these can be combined to produce a more complex model of social class, with multiple dimensions.  The social and cultural are particularly important when considering 'social reproduction and cultural distinction'.  We can also reexamine the formation of classes, the way in which class boundaries are generated.  Bourdieu has led to a number of subsequent sociological studies [including Bennett et al].  Until recently, however, there have been few questions on cultural and social capital in surveys, hence the importance of the BBC work, which collects a number of data from web surveys since 2001, although this is the first sociological one.

The BBC survey asked questions about cultural capital -- leisure interests, musical tastes, use of the media, food preferences-- which were rather similar to Bennett et al., 'The most sophisticated study of cultural capital ever conducted in Britain'(224).  Questions on social capital attempt to 'measure the range of people's social ties', whether they knew anyone in  37 different occupations.  The questions on economic capital asked about household income, savings, and the value of housing.  Additional information was gathered on 'household composition, education, social mobility and political attitudes'.  The survey launched in 2011 and responses were 'enthusiastic' until 161,400 surveys had been submitted.  However, there was a 'strong selection bias'towards the well-educated, and this led to the separate representative face to face survey using identical questions, this time with 1026 respondents, operated as a quota surveyed using a market research firm.  This helps calibrate the over representation of the professions in the original BBC survey.  Attempts to weight this survey by referring back to Bennett et al was not possible, because respondents were 'highly unrepresentative of their peers in these classes'(225) , with far more cultural and social capital, and 'the very act of participating in the [BBC survey] was a 'performative" way of claiming cultural stakes'.  However, this problem can be dealt with by combining it with the other survey.

The national survey was used first to develop 'robust summary measures' to develop the new model of class.  Thus for social capital, the 37 occupations mentioned above were coded by a particular Cambridge Social Interaction and Stratification Scale, which is 'widely validated', number of contacts was measured and then the 'means status scores of the occupations' of contacts, ranging from 85 to 5.  The range for each respondent could also be noted.  The table shows that the average respondent knows socially someone in 13 out of 34 valued occupations, the mean status score of contacts is 41, and the mean range is 62 [I have rounded up these figures].  In the end, only some measures proved robust enough - range was omitted because it varied with a number of contacts any way.  The mean status score of occupations was used, together with the mean number of social contacts reported.

For cultural capital, there is still much debate about how to define and measure it.  Bourdieu's categories have been questioned in recent research, and elite culture in the UK seems to be more liberal and tolerant, while there are more middle and upper class omnivores.  Instead, an 'inductive analysis of cultural taste' was pursued without assuming a division between high and low.  Specifically, multiple correspondence analysis was pursued on 27 cultural variables in the national sample [producing the usual Bourdieu type maps, 227, 228].  The first axis differentiated the culturally disengaged from the engaged, the second one distinguished those interested in highbrow culture, 'engagement with classical music, or attending stately homes, or museums, art galleries, jazz, theatre and French Restaurants' (227), as opposed to popular culture, 'associated with sport, using information technology, and popular and contemporary music' (226).  The latter can also be seen as '"emerging" cultural capital'.  The issue of engagement or disengagement was then compared to age and  class ['of respondent and of main earner when the respondent was a child'].  The first axis was 'clearly aligned with social class', with 'the routine classes' more disengaged, and with an age split reflecting the split between highbrow and emerging cultural capital.  As with Bennett et al, the implication is that there are 'two different modes of cultural capital'(227). 

Highbrow culture scores provided a maximum of 30 if people engaged 'often' in all these activities. Popular culture is further defined as 'engagement with video games, social networks sites, the Internet, playing sport, watching sport, spending time with friends, going to the gym, going to gigs and preferences for rap and rock'.  These activities were scored, with a maximum of 32.

Economic capital was based on 'household income, household savings and house price', combined to produce an overall variable of "assets", and the distribution is revealed in a table on 229.  These are measures for households not individuals, and they may reflect significant differences in individuals earning as, but the team preferred to household measure as 'more likely to tap the economic resources available to individuals', although there may well be a suppressed issue of gender inequality (228).

How might these measures be used to decide where to place the main class boundaries?  This is an inductive approach working from the results to see which classes emerge, 'latent class analysis', based on 'the most parsimonious way to group people to classes'(229).  The idea of the latent variable assumes that some parameters 'differ across unobserved sub groups', and it is these that provide the categories.  [Apparently] factor analysis assumes continuous latent variables.  [There are other technical reasons on 229].

Six variables were standardised - 'mean status scores of contacts, total number of contacts, highbrow cultural capital, emerging cultural capital, income and assets', and a latent class analysis carried out.  The national survey was used to correct the the BBC survey by combining the national values with a weighted version of the BBC values [rather a strange weight of 1/161,400 -- implies total dominance of the national survey?]. The effect was the ability 'to allocate classes' to all the respondents from the BBC survey derived from nationally representative data'[with a further statistical test]. The result was a seven class model.  Tables 5 and 6 show the results [below].

Savage tables

Further socio demographic details of the latent classes are also available on such matters as gender, ethnic minority, graduates, those with jobs and professions and management, and those from professional or senior management families, while another table shows that certain occupations were over represented -- eg in the elite especially chief executive officers, IT and telecommunications directors, various other directors and then judges and managers [It is actually quite a large table providing a good deal of detail about the main occupations in each class, 231-33]. This is one way to show that these groups are actually 'sociological meaningful', so that the latent class analysis 'is producing relevant findings' (233).  It is worth comparing them to the earlier classification schemes based on Goldthorpe.

Class one: elite.  Members here are 'the most advantaged privileged group in the UK' with high levels of all forms of capital [including a mean household income of 89,000 pounds and an average house price of 325,000.  These figures are 'almost double that of the next highest class'. Savings are also 'exceptionally high'.  They have the highest number of social contacts, and only the second highest mean status score [unsurprisingly -- if they have a wide circle of friends this must include those of lower classes? ].  They score highest on highbrow cultural capital and even have 'moderately high scores on emerging cultural capital'.  They also have other social advantages, 'the lowest proportion of ethnic minorities, the highest proportion of graduates, and over half come from families where the main earner was in senior management or the professions'.  They have restricted upward mobility into their ranks.  They are occupationally narrow, with over representations [as above], and over representation of elite university graduates, a disproportionate location in the southeast of England close to London.  They seem to be a definite elite within a class, promising to unite class and elite analysis.

Class two: Established middle class.  Good scores on economic capital, a larger class than the elite with 1/4 population, probably understood as the bulk of the old service class, with a high proportion of managers and professions.  There are more social contacts here than in any other classes' and these tend to be high status.  They are highly culturally engaged both for highbrow culture and emerging cultural capital, showing they are cultural omnivores.  Their possession of the three kinds of capital make them secure although not as wealthy as the elite.  They have a high proportion of graduates and a majority work in the professions or management as did their families. However, 'they are more open than the elite' especially in terms of access by ethnic minorities.  They are not particularly over represented by particular occupations although professionals in public service are 'modestly high'.  They live away from London and outside the southeast, in the provinces, and this 'is a sizeable bulwark of "Middle England"'(236).

Class three: Technical middle class.  This is another distinctive and original group.  It only takes up 6% of the population, but it is relatively prosperous, especially with household savings and expensive houses.  It is therefore the 'second most prosperous class in terms of economic capital'.  However, socially and culturally 'it is much more restricted' (237), with a low number of social contacts, even though these tend to be high status - 'it presumably socializes nearly exclusively with other professional experts', so it is not the poor or disadvantaged who have the most restricted social networks, but this group!  Culturally it has relatively low scores for both kinds of cultural capital and 'therefore appears to be relatively culturally disengaged': it is both relatively socially isolated and culturally apathetic.  There is a lower proportion of graduates and professionals and managements, but certain occupations are over represented, for example aircraft pilots.  So members 'have achieved good economic rewards often without distinctive credentials'.  59% of members are women.  Researchers and scientists and technicians are over represented, with graduates from established and prestigious universities, slightly more likely to be in science and technology.  Locations are primarily in the southeast where the jobs are likely to be found, but these are suburban dwellers.  Overall, this category seems to support Savage's earlier argument that there is 'a distinctive technical group somewhat at odds with the larger section of the middle classes who are more oriented towards the arts and humanities'.

Class four: New affluent workers.  This class scores highly on emerging cultural capital but low on highbrow versions, but is not culturally disengaged.  Economic capital values are moderate, so people here are 'economically secure without being very well off'.  There are high scores on social contacts although they have moderate status.  Thus there are moderate scores on all three capitals with a particular interest in emerging cultural capital.  Social origins are mostly non middle class, and few have been to university.  57% are men.  Where there are graduates, new universities are overrepresented.  There is a high proportion of young people.  Various white collar and blue collar jobs appear, largely in the private sector and 'in customer facing occupations' (239).  They are 15% of the population.  They tend to be over represented in 'old manufacturing centres of the UK'.  This group has not particularly benefited from educational expansion, nor from major inherited resources, but have achieved relative security and are relatively engaged.  The term affluent worker is used to indicate 'the need to understand this is 'an unusually fluid grouping'(240).

Class five: Traditional working class.  This is one of the 'clearly less advantaged classes' with moderately poor scores on economic capital, even though 'members predominantly own their homes'.  They have restricted social contacts with only average data scores, only moderate scores on highbrow cultural capital, and low scores on emerging cultural capital.  All these measures of capital are therefore low, but they are not totally deprived.  There are few graduates, and an over representation of traditional working class occupations, including 'menial white collar occupations'.  This group is particularly under represented in the BBC survey.  It is predominantly female.  Many graduates tend to have come from those unis  specializing in mature or part time students '(such as Birkbeck and the Open University)'.  They are located in the old industrial areas, especially in Scotland, Wales and Northern Ireland, and they can be seen 'as a residue of earlier historical periods'.

Class six: Emergent service workers have a modest household income but are more likely to rent.  There is a significant number of social contacts with moderate status scores, more emergent cultural capital and high levels of engagement in things like 'youthful musical sporting and Internet activities', but with low highbrow cultural capital.  It is marginal in the sense that it has low economic capital but higher social and cultural capital.  It is relatively young, with a high proportion of ethnic minorities.  Members are not graduates or from middle class families, but they are more culturally engaged unlike the traditional working class.  They work in lower levels of the service sector, and appear to be the groups who 'are "making their way" in a range of relatively insecure occupations' (241).  Any graduates are overrepresented from universities specializing in arts and humanities.  It is an urban group living in cheaper areas of the larger cities.

Class seven: Precariat.  These are economically the poorest, they have a small social range with a low mean status, and low scores for both highbrow and emerging cultural capital.  This is the most clearly deprived class, even though they are relatively large, consisting of 15% of the population.  They are to be found in old industrial areas, especially Stoke on Trent!  They are unlikely to have attended university.  They are overrepresented 'amongst the unemployed, van drivers, cleaners, carpenters, care workers, cashiers and postal workers'.  This is a group with high amounts of insecurity on all measures.

So this is a new model of class, different from Goldthorpe, based on inductive rather than deductive approaches, and shedding more light on cultural and social dimensions and their role in class division.  The plan is to produce more qualitative research in the future.  All the measures are still controversial, of course.  Economic capital might obscure divisions within households.  Social capital measures are based on validated scores, although the division into highbrow and emergent groupings departs rather from Bourdieu ['more nuanced', 243 is their claim]: whether emerging cultural capital will go on to produce a robust form requires further investigation.

Wider implications include a less than clear fit between occupational profiles and seven classes, which suggests 'the need for caution', and investigations on 'whether specific occupations are clearly differentiated from each other'(244).  Overall, with the exception of CEOs, 'we have found no clear affiliation between specific occupations and our latent classes'(245), implying that class is less a matter of occupation than the result of a multi dimensional process, spreading occupations between different classes.  There does seem to be a distinctive elite, and this is the first time that that group has been particularly identified in the class analysis: this has critical implications for the old notion of the service class, and does demonstrate more clearly 'the accentuation of social advantage at the top of British society'.  At the opposite end there is now a precariat, 15% of the population, again this was not identified before in class analysis.

Only two of the seven classes conform to the old models of middle and working class.  The boundaries between them might be becoming blurred, as in much recent discussion.  This scheme sees the established middle class as corresponding to the service class in the old analysis.  It is still the largest single class, with 25% of the population, so there is a secure group in British society.  The traditional working class might be seen as the survivors of the old class, but they now only compromise 14% of the population and are relatively old, so they might be 'fading from contemporary importance'.  Nevertheless, overall, this is still only 39% of the national population, with the majority falling into other classes outside the conventional analyses.  The distinction between white and blue collar jobs is now 'of little value in unraveling these patterns'.

Despite obtaining moderate levels of economic capital, there is no strong association with conventional highbrow culture, unlike Bourdieu.  Instead, there have been new ways of acquiring economic and social capital.  It is also the case that economic capital does not always lead to extensive social networks.  Social and cultural engagement does not decline evenly with social class position.

The new affluent workers and the emergence service workers are perhaps the most interesting groups.  They can be seen as 'the children of the "traditional working class"'(246), arising from 'deindustrialisation, mass unemployment, immigration and the restructuring of urban space'.  They are engaged with the emerging cultural capital and have extensive social networks, so they might represent a new social formation not just a disengaged group.

Overall, social inequality has polarised, and traditional middle and working class divisions have become fragmented producing 'more segmented forms'.  More detail will be apparent from future publications and research.