Australian Journal of Political Science

ISSN: 1036-1146 (Print) 1363-030X (Online) Journal homepage: www.tandfonline.com/journals/cajp20

Class, capital, and identity in Australian society

Jill Sheppard & Nicholas Biddle

To cite this article: Jill Sheppard & Nicholas Biddle (2017) Class, capital, and identity
in Australian society, Australian Journal of Political Science, 52:4, 500-516, DOI:
10.1080/10361146.2017.1364342

To link to this article: https://doi.org/10.1080/10361146.2017.1364342

Published online: 14 Aug 2017.

Submit your article to this journal

Article views: 17197

View related articles

View Crossmark data

Citing articles: 14 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=cajp20
AUSTRALIAN JOURNAL OF POLITICAL SCIENCE, 2017
VOL. 52, NO. 4, 500–516
https://doi.org/10.1080/10361146.2017.1364342

ARTICLE

Class, capital, and identity in Australian society

Jill Shepparda and Nicholas Biddleb
a
School of Politics and International Relations, The Australian National University, Acton, ACT, Australia; bANU
Centre for Social Research and Methods, The Australian National University, Acton, ACT, Australia

ABSTRACT ARTICLE HISTORY

Despite a comparatively ‘flat’ social structure and lack of obvious Accepted 16 July 2017
class-based cleavages, Australian society is stratified by objective,
multidimensional measures of social class. Using data from a July KEYWORDS
Social class; measurement;
2015 survey of a random sample of Australian citizens, latent class capital; Australia
analysis identifies six class types in Australian society, based on
the distributions of cultural, social, and economic capital among
respondents. The resulting classes are categorised as ‘precariat’,
‘ageing workers’, ‘new workers’, ‘mobile middle’, ‘emerging
affluent’, and ‘established affluent’. The precariat is characterised
by high numbers of retired pensioners, the ageing worker class
the highest mean age, and the new worker class by its low rate of
unemployment. The established middle class accounts for one
quarter of the adult population, while the emergent affluent class
has the youngest mean age, and the established affluent is the
most advantaged. We also show Australians are acutely aware of
their class identity.

Introduction and overview

Australian society is not commonly characterised by obvious class cleavages. Rather, ‘a
perceived social equality among people [is] said to structure social interaction, and equal-
ity of opportunity [is] argued to be a major feature of society’ (Western 1991, 14). Further,
while Australian politics were once characterised by a manual/non-manual labour clea-
vage, Kemp’s (1978) findings on post-war voting behaviour has justified the subsequent
widespread view that Australians’ political allegiances cut across conventional class
lines. The (relative) absence of visible class signifiers in Australia is a source of national
pride (Kemp 1978, 287). Certainly, Australian society has enjoyed comparatively high
rates of socio-economic equality, but there is increasing evidence of stratification within
the Australian population (Whiteford 2014). Given the context of increased socio-econ-
omic inequality, we examine the existence of social class in Australian society.
Academic interest in social class in Australia is reasonably dormant, with research
tending to focus on the distribution of income or wealth (Leigh 2013; Wilkins 2014), or
intergenerational mobility (Mendolia and Siminski 2015). By contrast, the recent inter-
national literature has explored social structures that entrench income and wealth dispar-
ities (for example, Piketty 2014; Chetty et al. 2014; Dorling 2015). Generally, the role of

CONTACT Jill Sheppard jill.sheppard@anu.edu.au

© 2017 Australian Political Studies Association
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 501

social class in Australian social (and political) life has received little attention in the past 20
years, mirroring international trends (for example, Clark, Lipset, and Rempel 1993; Waal,
Achterberg, and Houtman 2007). In this context, we study the distribution of both self-
assessed and objectively measured social class in Australia, and the extent to which Aus-
tralians accurately identify their objective class membership.
This study adopts Savage et al.’s (2013) framework and methodology to identify the
existence of latent classes and class membership in Australia, based on a telephone-
based probability sample survey conducted in July 2015 (Sheppard and Biddle 2015).
Drawing on Bourdieu’s (1984) multidimensional concept of social class, Savage et al.
(2013) use latent class analysis of British survey respondents’ economic, cultural, and
social capital to identify class membership. Savage et al. identify seven coherent, latent
classes within British society, including what they term the ‘precariat’, who tend to be
the most disadvantaged, an entrenched ‘elite’ class, and five classes in between character-
ised by varying combinations of capital possession. We extend Savage et al., through the
inclusion of a measure of self-assessed social class, to analyse the relationships between
Australians’ self-perceptions and their more objective social position.
The paper commences with a discussion of the measurement of social class in the pol-
itical science and sociological literatures, and the identification of social classes in Austra-
lia. The following section details the original survey data collected and methods employed
in the study, focusing on the translation of the Savage et al. (2013) methodology to Aus-
tralia. Then follows the analysis, wherein we outline the major finding of our study: the
existence of six latent classes in Australian society. Based on the socio-demographic
characteristics of members of each class, we assign class names: the precariat, the
ageing worker class, the new worker class, the established middle class, the emergent afflu-
ent class, and the established affluent class. We provide comprehensive portraits of each
class, based on empirical analysis of the original survey data. Finally, we discuss the impli-
cations of these findings for our understanding of Australian society.

Identifying and studying social class

Broadly, ‘social class’ can describe any categorisation of society by hierarchical strata,
according to individuals’ social, economic, or demographic characteristics. Commonly,
class is understood as having three hierarchical categories: working, middle, and upper.
When asked to describe their class status, the vast majority of the population in liberal
democracies identify as either working or middle class (see, for example, Jackman and
Jackman 1973; Cameron and McAllister 2016). Academically, the study of social class
has centred on the role of individuals’ occupation in determining their status, while
recent approaches have expanded on occupation to also consider individuals’ social and
cultural habits, as well as non-occupational economic resources. The academic literature
on social class can accordingly be described as comprising three waves: a first wave based
on Marxist theory; a second wave emphasising the central role of occupation; and a third,
and current, wave that incorporates the additional dimensions of social and cultural
capital.
First wave scholars of social capital Wright (2005) focus, per Marx, on the roles of
labour, capital, and political power in class formation and inter-class dynamics (Weber
1946; see, for example, Robinson and Kelley 1979). Savage et al. (2013) describe this
502 J. SHEPPARD AND N. BIDDLE

approach as a ‘moral’ measure of an individual’s standing within a community (akin to

Weber’s measure of status). Research from this wave is more recently critiqued as
making a priori assumptions about the pre-eminence of capital-rich classes and the lack
of inter-class mobility (Robinson and Kelley 1979; Goldthorpe and Marshall 1992).
Instead, the first wave of social class research is largely theoretical. By including empirical
evidence of social mobility and the weakening of capital-labour relations, ‘consideration
should also be given to theories holding that class relations are in fact of diminishing
importance for life-chances and social action’ (Goldthorpe and Marshall 1992, 382).
To that end, the second wave emerged in the 1970s to overtake the first wave’s deduc-
tive, relatively unempirical model of class. This wave is marked by the creation and wide-
spread adoption of what is often termed the ‘Goldthorpe schema’ (alternatively the
‘Erikson-Goldthorpe-Portocarero’ or ‘EGP schema’) of measuring class (Erikson and
Goldthorpe 1992; see also Erikson, Goldthorpe, and Portocarero 1979). In contrast to
the earlier Marxist and Weberian conceptualisations of class, this approach to measure-
ment claims not to assume hierarchical or static relations between classes, instead empha-
sising individuals’ occupational status (for example, employee or employer, service or
manual workers) as a heuristic for what is in reality a multidimensional, nuanced class
structure. This schema has influenced how Australian social class is measured and ana-
lysed, including in designing indices of occupational prestige and determining educational
inequalities (Baxter et al. 1991; Marks 1999; McMillan, Beavis, and Jones 2009).
A related stream of research introduces ‘new classes’ into the EGP schema. ‘New class’
researchers argue that the high-grade (e.g. business owners and chief executive officers,
etc.) and low-grade (e.g. middle managers and highly skilled technicians, etc.) professional
classes within the EGP do not adequately account for employer–employee relations
(Güveli, Need, and Graaf 2007a). The proposed ‘technocrat’ and ‘social and cultural
specialists’ classes (at both the high-grade and low-grade professional levels) take into
account the inherent antagonism between management and worker. Holding positions
of comparative prestige, technocrats are more likely to work in supervised, capital-sup-
porting roles, while social and cultural specialists have skills that are not easily supervised,
and tend to work in public sectors (Güveli, Need, and de Graaf 2007b). Accordingly, the
two types of professionals possess different levels of social and cultural capital, and display
different political and social behaviours (Oesch 2008).
Expanding upon the ‘new class’ theories, Savage et al. (2013, 4–5) argue that, although
the Goldthorpe schema has a stronger basis in sociological theory than its antecedents, it
does not capture differences in cultural and social practices, and income and wealth, that
are not necessarily distributed along occupational lines. Rather, this critique contends that
externally valid study of social class – to the extent that researchers understand class as
something consisting of symbolism, ritual and cultural practice, beyond just occupational
status – requires broader measurement than just the Goldthorpe schema. To this end, in
2011 Savage et al. (2013) conducted the Great British Class Survey (GBCS) in the United
Kingdom to take into account social, cultural, and financial cleavages orthogonal – or
potentially orthogonal – to the Goldthorpe schema.
Per the chronological narrative of social class research, this – the third and current wave
– draws on Bourdieu’s (1984) conceptualisation of class as comprising three forms of
capital: social, cultural, and economic. Arguably the most important distinction
between this and the previous wave of research is the aforementioned orthogonality of
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 503

the three forms of capital. It is possible for individuals to possess one form of capital, but
not another (particularly in the case of social and cultural capital) (Bourdieu 1984;
Bennett, Emmison, and Frow 1999; Prieur and Savage 2013). Further, the three forms
are not necessarily evenly distributed within or between individuals, and the resulting var-
iance allows researchers to identify more fine-grained, nuanced class groups than the more
blunt Goldthorpe measurement (Savage et al. 2013). However, the second and third waves
are not entirely without common ancestry and shared ideas. The recent focus on the dis-
tribution of capital across income groups has precedents from second wave theories such
as Wright’s (1985) ‘contradictory class locations’ and even Savage et al.’s (1995) second
wave era applications of Bourdieu.
Cultural capital has formed a less cohesive field of study, with varying measurements
and interpretations across academic disciplines (Lamont and Lareau 1988; Throsby
1999; Kingston 2001). Studies of cultural capital – in Bourdieau’s concept – in Britain
and Australia have found that cultural tastes form in ordered and meaningful ways,1
but that larger gaps exist between those who do and do not engage in cultural consump-
tion, than between those who engage in ‘highbrow’ and ‘lower brow’ (for example, listen-
ing to rock music or watching horror films) cultural activities (Bennett, Emmison, and
Frow 1999; Bennett et al. 2009). The study of economic capital, particularly the distri-
bution of income and wealth within and between societies, is profuse both internationally
and in Australia.
Published criticisms of the GBCS are few in number but comprehensive in nature. Mills
(2015, 2014) questions the theoretical basis of the study, the coverage of the initial prob-
ability-based survey sample, and the external validity of the study’s ‘elite’ class, which com-
prises only 6% of the GBCS sample. Dismissing the GBCS as a ‘data dredging exercise’,
Mills (2014, 437) nonetheless makes cogent points about the limitations of the GBCS’s sec-
ondary non-probability based sample. Elsewhere, his critique is preoccupied with what he
deems the investigators’ overambitious inference of GBCS data to create a ‘class map’ of
British society. To this extent, Mills’ caution is entirely justifiable.
However, these points do not negate the validity of the GBCS as a sample-based snap-
shot of latent social class structure in British society, with all the constraints that sample-
based data entail. Silva (2015) outlines a similar critique of the GBCS, urging researchers’
and other audiences’ understanding of the study as a ‘public relations exercise’. Conver-
sely, we do not personally agree with Silva’s (2015, 386) contention ‘that the academic
content needs to be distinguished from the media audience … but they are too closely
linked in the case of the GBCS’. Rather, academic researchers should be encouraged to
seek out non-academic audiences and influence public debate.
Mirroring the international literature on social class, the few Australian studies of social
class have reached little agreement on either the existence of class in Australian society or
its role in broader social behaviours. The variations in findings (for example, Alford 1973;
Jones 1974; Kemp 1978) can largely be explained by different conceptualisations and
measurement of social class, spanning Marxist and Weberian accounts to later, Gold-
thorpe-influenced approaches (Kemp 1978; Graetz and McAllister 1994). A large
number of Australian researchers have used occupation and occupational prestige scales
based on Australian Government classifications (Broom, Jones, and Zubrzycki 1965;
McMillan, Beavis, and Jones 2009). Such measurement has identified both a decline in
the rigidity of class structure, as social and occupational mobility has increased, and
504 J. SHEPPARD AND N. BIDDLE

decoupling of class (to the extent that it exists) and party identification (for example,
Connell and Irving 1992, Ch 6). However, the role of class consciousness in structuring
Australians’ social identities remains comparatively low: Emmison and Western (1991,
291) note that ‘even “supporter of a sports club” has apparently more significance than
social class’ in determining how Australians view themselves as social beings.
Survey evidence has suggested that Australians enjoy comparatively high rates of both
intergenerational occupational mobility, and life-cycle career mobility (Graetz and McAll-
ister 1994). However, new data are beginning to suggest that Australia is less inter-gener-
ationally mobile than previously assumed. Specifically, Mendolia and Siminski (2015)
conclude that ‘Australia is not particularly mobile in an international context. It is less
mobile than the Scandinavian countries, as well as Germany, Canada and New Zealand,
but is more mobile than France, Italy, United States and United Kingdom.’ However,
few studies have examined the existence or mobility of social class status, beyond socio-
economic indicators. This study fills a substantial void in that regard.

Data and methods

The data used in this analysis come from a stratified random-digit dialled (landline and
mobile phone numbers) survey of 1200 Australians aged 18 and over. The survey was con-
ducted between 13 and 27 July 2015. The response rate was 21%.2 The resulting data have
been weighted in a two-stage weighting process. First, a weight is created to adjust for the
effects of telephony status (i.e. the number of landlines in a household and/or possession of
a mobile phone) on the likelihood that an individual will be randomly selected from the
population. Second, the final sample is weighted to reflect population distributions of age,
gender, and location. Unweighted, the mean age of respondents in the sample is 53.14
years. With weights applied, the mean age of respondents is 46.10 years. The dataset is
available for unrestricted download and reuse from the Australian Data Archive website.3
To replicate the Savage et al. (2013) study as closely as possible, social class is measured
here along three dimensions – economic, cultural, and social capital – with two measures
of each. The measures each have equal weight in the final model. Economic capital is
measured by respondents’ gross annual household income and their assets. Assets are
in turn measured by sum of the self-reported estimated value of property owned or mort-
gaged by the respondent, and their self-reported cash savings.
Per Bourdieu’s (1984) explicit delineation of cultural tastes, cultural capital is measured by
a count of ‘highbrow’ and ‘emerging’ cultural activities in which respondents have engaged
during the previous 12 month period. From a list of 15 possible activities, multiple corre-
spondence analysis (MCA) reveals two dimensions of activities, representing ‘highbrow’
and ‘emerging tastes’ respectively. MCA is used here to follow Savage et al. (2013), but is
only one of a range of possible methodological techniques for reducing multiple measures
into dimensional scales. Principal axis factor analysis and – treating the data as continuous
measures – principal component analysis also reveal two dimensions or factors within these
measures, with similar loadings.4 The MCA results are reported in the appendix.
Based on these results, the measure of highbrow cultural capital comprises a count of
respondents’ participation in a battery of cultural activities: gone to the opera, watched
ballet or dance, gone to the theatre, and listened to classical music. Participation in
each activity is a binary value (where 1 = yes). The ‘emerging’ cultural capital measures
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 505

counts whether respondents listened to rock and/or indie music, attended gigs, played
video games, watched sports, exercised and/or attended a gym, used Facebook and/or
Twitter, and listened to rap music during the previous 12 months. Five cultural activities
measured in the survey – listening to jazz, socialising at home, going to museums and gal-
leries, doing arts and crafts, and listening to hip-hop/rap – are excluded from both the
highbrow and emerging cultural capital scales, as they did not load onto either dimension
with any strength. Data on these activities are discarded from the analyses.
The third dimension of social class in this study, social capital, consists of the number of
a range of occupations among respondents’ social networks, and the mean prestige of
those occupations. The occupations presented to respondents are: secretary; nurse;
teacher; cleaner; university lecturer; artist; electrician; office manager; solicitor; farm
worker; chief executive; software designer; call centre worker; postal worker; scientist;
truck driver; accountant; and shop assistant. The count of known occupations is computed
as a sum of each occupation that the respondent knows socially (where 1 = yes). To
compute the mean prestige of respondents’ known occupations, each occupation is
ascribed a prestige score using the Australian Socioeconomic Index 2006 (McMillan,
Beavis, and Jones 2009), a validated index of occupational prestige.
The six measures of capital are collapsed into categories by quantile, to create similar
ranges of each measure while retaining the nature of the original distributions. The six cat-
egorical measures are then modelled with a polytomous variable latent class analysis
(LCA).5 This analysis regresses a latent outcome variable (with a pre-assigned number
of latent classes) onto the categorical manifest predictor variables. This analytical
approach again replicates the research design used in Savage et al. (2013).
There are, of course, limitations to such an analytical approach. As with all inferential
analyses based on cross-sectional, probability-sample derived data with a modest sample
size, there is a possibility that the data are not generalisable to the specified population or
to different time periods. Additionally, latent class analyses do not incorporate population
parameters in specifying models, making the resulting model entirely sample-based. Some
of this reduction in reliability vis-a-vis other cluster-type methods is negated by iterating
the model estimation algorithm6 through 1000 cycles to achieve internal stability of the
model. Externally, however, the nature of the model places limits on the potential for com-
parative analysis.
To overcome some of the above limitations, and expand on Savage et al. (2013), this
study also asks respondents to place themselves in a class using a response frame of
‘upper’, ‘middle’, ‘working’, and ‘not sure’.7 This measure enables comparison of respon-
dents’ self-assessment with objective measurement. The effects of self-assessed and objec-
tive class membership on respondents’ vote intention will also be compared. Additional
covariate measures – not specified in the latent class model – are used to corroborate
and add detail to the self-assessed and objective classes. These include respondent age,
gender, educational attainment, household income, occupation, and parents’ (or guar-
dians’) employment and occupational statuses when the respondent was 15 years of age.

Analysis: class membership in Australian society

We commence the latent class analysis inductively, comparing output from seven, six, five,
four, and three class models. We have no a priori reason to suggest that Australian society
506 J. SHEPPARD AND N. BIDDLE

has any specific number of social classes, but use Savage et al. (2013)’s identification of
seven classes as a starting point. Each analysis is conducted several times (nrep = 10) to
achieve maximum convergence. Table 1 displays the model fit results from each of the
five analyses. The six-class outcome variable model produces the lowest Akaike Infor-
mation Criterion (AIC) score of each model tested, and consequently is selected here as
the most valid. Additionally, the relative entropy measure for the six-class model is
0.898, suggesting clear distinctions between the latent classes (Celeux and Soromenho
1996).
The resulting six-class model produces what appear, prima facie, to be meaningfully
different latent subgroups of respondents based on distributions of the six manifest
measures included in the latent class model. Each individual in the dataset is assigned
to a class based on the highest posterior probability of membership of any of the five
classes. For instance, an individual respondent with a 0.33 probability of belonging to
class 1, 0.47 probability of belonging to class 2, and 0.20 probability of belonging to
class 3 is assigned to class 2. A binary measure is used to score probabilistic assignment
to a class, where 0 equals not assigned, and 1 equals assigned.
Based on the mean characteristics of each class, we can designate meaningful labels to
each class (Table 2). Cross-checking the six-class model against descriptive statistics of
covariate measures – including age and educational attainment8 – suggests that the six
classes have substantive meaning. Further, examination of the occupational prestige of
respondents’ parents when respondents were aged 15 provides information about the
intergenerational mobility – or stability – of one aspect of class designation. The six
classes are labelled ‘precariat’, ‘ageing workers’, ‘new workers’, ‘established middle’, ‘emer-
ging affluent’, and ‘established affluent’. They compare to Savage et al.’s (2013) seven
classes: the ‘precariat’, ‘emergent service worker’, ‘traditional working class’, ‘new affluent
workers’, ‘technical middle class’, ‘established middle class’, and ‘elite’. Although the latent
class analysis makes no assumptions based on the relative order of responses among the
six manifest variables, the substantive question under study here lends itself to some
ordering of latent classes. Accordingly, those classes are presented here along a hypothe-
tical dimension from fewest resources and advantages at one end, to most at the other.
Figure 1 presents the frequency distribution of each of the classes, in order from fewest
to most resources.
First along that dimension is the precariat, comprising 13% of the total sample.
Members of the precariat have the lowest mean household income: along the one-to-
seven point scale of household income, members of this class average 2.7. Further, they
report the fewest real estate and cash saving assets. Members of this class also have the
lowest educational attainment, the lowest participation in both highbrow and emerging

Table 1. Latent class analysis model fit statistics for different model specifications (based on outcome
variable).
AIC BIC Log-likelihood Likelihood ratio χ2
Seven-class 20,900.35 21,999.81 −10,988.93 5793.19 11,044.18
Six-class 20,892.07 21,833.73 −11,039.38 5862.10 11,443.11
Five-class 20,964.50 21,748.37 −11,110.13 6002.71 11,746.07
Four-class 21,103.84 21,729.92 −11,219.66 5917.24 12,683.38
Three-class 21,441.56 21,909.85 −11,420.01 6317.94 14,143.09
Note: n = 1026. Missing data excluded listwise. BIC, Bayesian Information Criterion.
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 507

Table 2. Mean characteristics by class membership.

Ageing New Established Emerging Established
Precariat workers workers middle affluent affluent
n = 150 n = 168 n = 292 n = 289 n = 174 n = 127
Social capital
Total occupations Mean 3.3 6.9 12.42 9.73 14.66 16.7
known socially SD 1.6 1.07 0.95 1.14 0.94 1.37
Mean social contact Mean 7.54 18.72 37.15 28.14 44.37 52.19
prestige SD 4.95 3.28 3.11 3.69 2.43 2.66
Cultural capital
Emerging activities Mean 3.77 4.14 5.09 4.89 5.41 5.57
SD 1.62 1.46 1.67 1.71 1.67 1.61
Highbrow activities Mean 1.82 2.14 2.42 2.35 2.88 3.04
SD 0.84 1.09 1.08 1.18 1.11 1.03
Economic capital
Income Mean 2.72 3.09 4.74 4.35 4.89 4.97
SD 1.83 1.73 1.83 2.05 1.92 2.14
Assets Mean 4.82 5.08 5.54 5.3 5.38 6.8
(property + savings) SD 2.17 2.37 2.44 2.66 2.53 3.04
Property Mean 2.39 2.57 2.88 2.79 2.88 3.07
SD 0.81 0.81 0.8 0.9 0.9 0.88
Savings Mean 1.88 2.12 2.42 2.23 2.41 3.21
SD 1.68 1.85 2.07 2.04 2.06 2.52
Additional explanatory variables (not included in latent class analysis)
Age Mean 56.2 58.2 51.34 51.95 51.09 52.42
SD 19.96 18.44 15.73 17.9 15.16 15.44
Education Mean 6.42 7.01 8.16 7.58 8.36 9.03
SD 2.81 2.64 2.42 2.49 2.49 2.19
Mother’s occupational Mean 40.66 47.66 49.57 44.87 53.69 49.23
prestige SD 19.21 22.62 21.2 20.84 20.91 21.09
Father’s occupational Mean 39.2 43.01 48.5 46.61 48.11 48.59
prestige SD 19.97 20.86 21.25 21.33 22.44 20.66
Respondent’s Mean 43.78 52.19 53.28 51.54 54.53 64.62
occupational SD 22.57 21.68 21.65 21.39 19.47 18.86
prestige

Figure 1. Percentage distribution of class membership (based on highest probability) among

population.
508 J. SHEPPARD AND N. BIDDLE

cultural activities, the lowest social contact score, and their contacts have the lowest occu-
pational prestige. Finally, members of this class have not received the intergenerational
advantages typical of other classes: their parents worked in comparatively low-prestige
jobs. The GBCS (Savage et al. 2013) also identifies a precariat class, which comprises
15% of the British population.
Second is the ‘ageing worker’ class, with the highest mean age (58 years) of the five
classes. This class comprises 14% of the Australian population. Members of this class
possess slightly more resources than members of the established working class, and
have come from slightly more advantaged backgrounds: their parents’ occupational pres-
tige is higher than that of the precariat, and comparable to the new worker class. Rates of
educational attainment are higher than among members of the precariat, and rates of
social and cultural capital are slightly higher. This class reports moderate rates of house-
hold income, savings, and property assets. The most comparable class in British society is
Savage et al.’s (2013) traditional working class, which also constitutes 14% of that
population.
By comparison, members of the ‘new worker’ class (24% of the population) are on
average six years younger than ageing workers (at 52 years). They report slightly higher
rates of educational attainment, higher household incomes, savings and property assets,
and higher rates of social and cultural capital. The mean occupational prestige of this
class is slightly lower than that of the ageing worker class, but it has been parlayed into
greater material wealth. The British study identifies two comparable classes: a ‘new afflu-
ent worker’ class and a ‘technical middle’ class, which combined represent 21% of the
British population. These two British classes are distinguished by substantial differences
in social capital (Savage et al. 2013), where the Australian data show no such cleavage
in this group.
The ‘established middle’ class (24% of the population) has a similar mean age to the
new workers (51 years), slightly higher household income, and slightly higher levels of
educational attainment. They report marginally higher rates of both social and cultural
capital, and their parents had higher occupational prestige during the respondents’
youth. In every regard, they enjoy greater advantages than members of the new worker
class. The equivalent class in the British population reports similar distributions of
capital, and constitutes almost exactly the same proportion of the population (25%)
(Savage et al. 2013).
The next most advantaged group identified – the ‘emerging affluent’ class – is the equal
youngest (mean age of 51 years). In terms of economic, social, and cultural capital,
members of the emerging affluent class report greater advantages than the established
middle (and preceding) classes, but less than the more prosperous ‘established affluent’
class. This class reports higher household incomes than the established middle class,
but lower wealth accumulation (in the form of property and savings), justifying the ‘emer-
ging’ label.
The final class identified in this analysis is the ‘established affluent class’. To the extent
that the three preceding classes – the new workers, mobile middle, and emergent affluent
classes – can be broadly grouped as ‘the middle’ of Australian society, the established afflu-
ent class is distinguished by very high rates of occupational prestige (at 10 points higher
than the emerging affluent class), income, and accumulated wealth. Members of this class
report the highest rates of educational attainment, and the highest rates of both social and
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 509

cultural capital. Moreover, they are the most likely to possess both emerging and highbrow
cultural capital: distinct forms of capital more commonly associated with different classes
(Prieur and Savage 2013; Savage et al. 2013). The British project labels its comparable class
the ‘elite’, presumably to denote the substantial differences between it and the closest
group, the established middle class (Savage et al. 2013). However, the existence of the
intermediary emerging affluent class in Australia situates the most advantaged group
closer to the ‘rest of the pack’; consequently, labelling this class as ‘elite’ would infer
greater disparity than shown in the data.

Analysis: employment and occupational characteristics of the six classes

The latent class analysis model used to identify the six social classes does not include
measures of individuals’ employment status or occupation, per the Savage et al. (2013)
model. This represents perhaps the largest break between the first and second waves of
social research, preoccupied with the centrality of capital-labour relations and occu-
pational status respectively, and the third wave approach, which de-emphasises occu-
pation in favour of social and cultural capital. Instead, we examine the distribution of
employment and occupation by social class as a post hoc check on the validity of the ident-
ified classes. In this section, we report notable differences in both the extent and type of
work undertaken across the six classes, supporting both the validity of the identified
classes and the explanatory power of the Savage et al. (2013) capital-based model.
Data on the employment status of Australians offer insight into the observed distri-
bution of social, cultural, and economic capital between the classes. Only 18% of precariat
members are in full-time employment (Table 3). More than one third – 36% – are pen-
sioned retirees, and a further 9% are self-funded retirees. Five per cent are looking for
work, and 4% are students (the highest rate among any class). Similarly, 35% of
members of the ageing worker class are pensioned retirees, with 31% in full-time employ-
ment. These two classes – the least advantaged, and among the oldest – are the least likely
to be engaged in employment, and consequently are excluded from the social and cultural
opportunities afforded by workplace engagement.
By contrast, almost half – 45% – of the new worker class is in full-time employment,
and 19% in part-time work. Only 13% are pensioned retirees, while 8% are self-funded

Table 3. Employment status by class membership.

Ageing New Established Emerging Established
Precariat workers workers middle affluent affluent
n = 150 n = 168 n = 292 n = 289 n = 174 n = 127
Full time 18.1 31.0 44.8 33.3 50.0 48.4
Part time 13.4 10.7 19.0 19.4 22.1 17.5
Unemployed 4.7 4.2 2.1 6.3 2.3 2.4
Retired (pensioner) 36.2 35.1 13.1 19.8 8.1 11.9
Retired (self-funded) 8.7 6.0 8.3 9.0 7.6 8.7
Retired 3.4 7.1 3.1 2.1 3.5 2.4
(combination)
Student 4.0 3.0 2.4 2.8 1.2 3.2
Keeping house 3.4 1.2 2.8 3.5 4.7 2.4
Other 8.1 1.8 4.5 3.8 0.6 3.2
Total 100.0 100.0 100.0 100.0 100.0 100.0
510 J. SHEPPARD AND N. BIDDLE

retirees. This class reports the lowest rate of unemployment among the classes, with only
2% looking for work in July 2015. Members of the new worker class have the highest rates
of workforce participation, and access to associated social and cultural capital. This, along-
side age, distinguishes this class from the less active ageing worker class.
The established middle class reports lower rates of workforce participation, but members
of the established middle class – with higher rates of savings and assets – appear more
entrenched and comfortable in their status than the new worker class. An exact third of
this class are in full-time employment, with a further 19% working part-time. Twenty per
cent of class members are pensioned retirees, while 9% are self-funded retirees. On the
other hand, this class has the highest rate of unemployment (6%) at July 2015.
Members of the emergent affluent class are the most likely to be in full-time employ-
ment (50%). A further 22% of this class is employed part-time; only 8% are pensioned
retirees, 8% self-funded retirees, and 2% looking for work. The class also has the lowest
percentage of students, at only 1%. Finally, almost half – 48% – of the established affluent
class is employed full-time; 17% work part-time, and only 2% are unemployed. A com-
paratively low 12% are pensioned retirees, and 9% are self-funded retirees.
The analysis of the distribution of occupations among the six classes provides
additional evidence for the validity of the latent class model (Table 4). The modal occu-
pations of the identified classes broadly reflect the hierarchy inferred by the Goldthorpe
(and similar) schema, while measures of social and cultural capital provide additional
dimensionality. The most common occupations among members of the precariat are pro-
fessional staff, trades and technicians (including skilled trade workers), and labouring
(including cleaners, construction and mining labourers, and factory process workers).
This class possesses the highest rate of sales workers, at 15% of members.
Among members of the ageing worker class still participating in the workforce, 29% are
employed in professional sectors, and 20% in trade and technical sectors. A further 15%
work as managers, and 14% in the community and personal service sectors (including
health care provision). The modal occupation among members of the new worker class
is professional (31%), followed by management positions (15%), and technical and
trade occupations (14%). A further 13% work in community and personal service pos-
itions. Members of the established middle class have the highest rate of employment in
the community and personal service sectors of all the classes, at 15%, while 28% of
members are employed in professional sectors.

Table 4. Main occupation by class membership.

Ageing New Established Emerging Established
Precariat workers workers middle affluent affluent
n = 150 n = 168 n = 292 n = 289 n = 174 n = 127
Managers 6.5 15.4 14.8 13.9 23.1 26.6
Professionals 21.7 29.2 31.2 27.8 27.3 49.4
Technicians/trades 21.7 20.0 13.7 12.6 10.7 3.8
Community/personal 8.7 13.9 12.6 14.6 11.6 5.1
service
Clerical/administrative 2.2 6.2 10.9 11.3 16.5 5.1
Sales 15.2 4.6 5.5 8.0 3.3 7.6
Machine operators/ 4.4 1.5 5.5 2.7 3.3 0.0
drivers
Labourers 19.6 9.2 6.0 9.3 4.1 2.5
Total 100.0 100.0 100.0 100.0 100.0 100.0
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 511

The emerging affluent class is likewise characterised by employment in professional

sectors (27%), followed closely by management (23%). Additionally, this class contains
the highest number of clerical and administrative workers, at 17% of all working respon-
dents. Finally, almost half – 49% – of the established affluent class is employed in the pro-
fessional sector, with a further 25% employed in management. This class is substantially
more ‘white-collar’ than the other five classes, again supporting the use of the Savage et al.
(2013) model to validly measure social class even without relying on occupational prestige
as an input measure.

Analysis: objective and self-assessed class membership

Respondents were also asked with which social class they identify, from an option of
‘upper’, ‘middle’, ‘working’, or ‘not sure’. Of the sample of 1200, only 71 did not place
themselves in one of the three classes named; the overwhelming majority appear both con-
scious of their own class identity and willing to self-report in a telephone survey. Of the
remaining 1129, more than half (57%) identify themselves belonging to the middle
class. A further 41% identify as working class, leaving only 2% of the valid sample – 21
respondents in total – identifying as belonging to the upper class.
Cross-tabulating self-assessed class membership as a factor of objective class membership
(Figure 2) indicates that subjective class does bear some relationship to the latent class analy-
sis results. Among members of the precariat, 2% describe themselves as being upper class;
60% as middle class; and 34% as working class. Four per cent of the precariat did not
place themselves in any class. The relatively diffuse distribution of self-assessed classes
among the precariat may be due to the high number of retirees in this class. Traditional
class identifiers such as occupation and education likely have a lasting effect on some Aus-
tralians’ subjective class identity, even as their economic, social, and cultural capital decline
in later years; accordingly, someone in a precarious economic or social position may still
reasonably identify as a class associated with their previous occupational status. In contrast,
while 9% of the ageing worker class do not place themselves in a class, 40% describe them-
selves as middle class, and 51% in the working class. This distribution suggests that working
class Australians have a strong and accurate sense of their class identity.
The new working class is more likely to self-describe as middle class (64%), with only
31% describing as working class, 1% as upper class, and 4% providing no answer. The stark
contrast with the ageing worker class with regard to subjective middle and working class
identity adds a further dimension to the distinctions between the two objective classes:
new workers’ greater propensity to identify as middle class suggests ambition toward
upward mobility that is much less evident among ageing workers. Similarly, members
of the established middle class – despite reporting more economic, social, and cultural
capital than new workers – overwhelmingly identify as working class (56%). Only 37%
of the established middle class identify as middle class. In fact, the self-assessed middle/
working divide is almost completely inverse between new workers and the established
middle class, with new workers much more likely to identify with the ‘higher’ class.
Members of the emerging affluent class predominantly identify as middle class (55%),
although 39% describe themselves as working class: a similar percentage as among the pre-
cariat or the new worker class. As with the new worker class, the ostensible contrast
between high levels of economic, social, and cultural capital and the propensity to identify
512 J. SHEPPARD AND N. BIDDLE

Figure 2. Self-assessed class membership by objective class membership.

with the ‘lowest’ class is perhaps indicative of the emerging affluent class’ upwardly mobile
and ambitious nature. Two per cent of the emerging affluent class identify as upper class,
and 4% provide no answer. Finally, members of the established affluent class are the least
likely to self-identify with a class, with 11% providing no response. They are the most
likely to identify as upper class (6%), while 58% identify as middle class, and only 25%
(the lowest among the classes) identify as working class. Of all of the classes, members
of the established affluent display the most accurate sense of their class identity, although
the greatest unwillingness to place themselves within the three listed categories. Across the
population generally, Australians are quite acutely aware of their class identity, with self-
assessed class membership reflecting the relative capital and mobility of the objectively
measured classes.

Concluding comments
Australian society is often characterised as comparatively egalitarian, having forsaken the
class hierarchies of its British antecedent. However, increasing evidence of socio-economic
stratification has made timely a re-examination of the existence of social class in
Australian society. Both the measurement and analytical concepts in this paper contribute
to what has been described as the third wave of studying social class, and constitutes
substantial advancement over previous measures such as the Alford index and Goldthorpe
schema (Savage et al. 2013). We have identified six distinct, meaningful classes in
Australian society, based on Australians’ economic, social, and cultural capital: the precar-
iat, the ageing worker class, the new worker class, the established middle class, and the
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 513

established affluent class. The precariat is characterised by high numbers of retired pen-
sioners, the ageing worker class the highest mean age, and the new worker class by its
low rate of unemployment. The established middle class accounts for one quarter of the
adult population, while the emergent affluent class has the youngest mean age, and the
established affluent is the most advantaged on every measure of capital.
We have also showed Australians are acutely aware of their class identity, with self-
assessed class membership reflecting the relative capital and mobility of the objectively
measured classes. Those who belong to the precariat, ageing worker, and established
middle classes are the most likely to self-describe as working class. Those who belong
to the upwardly mobile new worker and emerging affluent classes, as well as those in
the established affluent class, are more likely to identify as middle class. Further,
members of the established affluent class were the most likely (among a small number)
to identify as upper class. To the extent that Australians may perceive their society as
being relatively ‘classless’, at the individual-level they display a finely-honed sense of
their own relative class position.
The results presented in this paper have a number of implications. The results of the
latent class analysis demonstrate clear and substantive differences between social classes
in Australia. However, the identification of the six-class model suggests that Australian
society is stratified beyond just occupational categorisation or socio-economic status. Aus-
tralian social strata are nuanced, and incorporating non-economic measures of capital
provides vital information. Second, Australians are not as reticent to identify as being
in a particular class as might be assumed; this may suggest something about the lingering
effects of British settlement on our social structure, or else something about the willingness
of survey respondents to self-describe even in ambiguous contexts.
The final implication is perhaps the most important. As already noted, social class has
dropped off the academic discourse in Australia, despite its recent appearances in media
coverage and public debate. In the academic literature, researchers now commonly talk
about intergenerational mobility, social cohesion, social capital, and entrenched disadvan-
tage: each an important concept on its own. However, the results presented in this paper
show that objective and subjective measures of social class are still very much present in
Australian society. The study of Australian class stratification deserves much more
ongoing exploration.

Notes
1. Bennett, Emmison, and Frow (1999, 56) cite the British example that ‘those whose favourite
eating-out place is French restaurants tend to like Impressionism, classical music and
modern literature’.
2. Response rate calculated using the American Association for Public Opinion Research
Response Rate 3 formula.
3. http://ada.edu.au/
4. These analyses were conducted in R using the ‘FactoMineR’ (Husson et al. 2016), ‘Psych’
(Revelle 2016), and ‘GPArotation’ (Jennrich and Bernaards 2014) packages.
5. The LCA is performed using the ‘lca’ function in the ‘poLCA’ package for R (Linzer and
Lewis 2014).
6. In the case of the ‘poLCA’ software, these are expectation-maximisation and Newton-
Raphson algorithms to produce maximum likelihood estimates.
514 J. SHEPPARD AND N. BIDDLE

7. The exact question asked is ‘Which social class would you say you belong to?’, with the ques-
tion immediately following the questions on activities.
8. Educational attainment is measured in these data as an ordinal variable from 1 to 11. The
categories range from ‘Never attended primary school’ (‘1’), to ‘Postgraduate degree’ (‘11’).

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Jill Sheppard is a political scientist and lecturer in the School of Politics and International Relations
at The Australian National University.
Nicholas Biddle is an economist and the Deputy Director of the Centre for Social Research and
Methods at The Australian National University.

References
Alford, Robert R. 1973. Party and Society: The Anglo-American Democracies. Westport: Greenwood
Press.
Baxter, Janeen, Michael Emmison, John Western, and Mark Western. 1991. Class Analysis and
Contemporary Australia. South Melbourne: Macmillan Education AU.
Bennett, Tony, Michael Emmison, and John Frow. 1999. Accounting for Tastes: Australian Everyday
Cultures. Cambridge: Cambridge University Press.
Bennett, Tony, Mike Savage, Elizabeth Bortolaia Silva, Alan Warde, Modesto Gayo-Cal, and David
Wright. 2009. Culture, Class, Distinction. Oxford: Routledge.
Bourdieu, Pierre. 1984. Distinction: A Social Critique of the Judgement of Taste. Cambridge: Harvard
University Press.
Broom, Leonard, Frank L. Jones, and Jerzy Zubrzycki. 1965. “An Occupational Classification of the
Australian Workforce.” Journal of Sociology 1 (2): 1–16.
Cameron, Sarah M., and Ian McAllister. 2016. Trends in Australian Political Opinion: Results from
the Australian Election Study, 1987–2016.
Celeux, Gilles, and Gilda Soromenho. 1996. “An Entropy Criterion for Assessing the Number of
Clusters in a Mixture Model.” Journal of Classification 13 (2): 195–212.
Chetty, Raj, Nathaniel Hendren, Patrick Kline, and Emmanuel Saez. 2014. “Where Is the Land of
Opportunity? The Geography of Intergenerational Mobility in the United States.” The Quarterly
Journal of Economics 129 (4): 1553–1623. doi:10.1093/qje/qju022.
Clark, Terry Nichols, Seymour Martin Lipset, and Michael Rempel. 1993. “The Declining Political
Significance of Social Class.” International Sociology 8 (3): 293–316.
Connell, R. W., and Terence H. Irving. 1992. Class Structure in Australian History: Poverty and
Progress. Melbourne: Longman Cheshire.
Dorling, Danny. 2015. Injustice: Why Social Inequality Still Persists. Bristol: Policy Press.
Emmison, Michael, and Mark Western. 1991. “The Structure of Social Identities.” In Class Analysis
and Contemporary Australia, edited by Janeen Baxter, Michael Emmison, and John S. Western,
279–305. South Melbourne: Macmillan Education AU.
Erikson, Robert, and John H. Goldthorpe. 1992. The Constant Flux: A Study of Class Mobility in
Industrial Societies. Oxford: Clarendon Press.
Erikson, Robert, John H. Goldthorpe, and Lucienne Portocarero. 1979. “Intergenerational Class
Mobility in Three Western European Societies: England, France and Sweden.” The British
Journal of Sociology 30 (4): 415–441.
Goldthorpe, John H., and Gordon Marshall. 1992. “The Promising Future of Class Analysis: A
Response to Recent Critiques.” Sociology 26 (3): 381–400.
 AUSTRALIAN JOURNAL OF POLITICAL SCIENCE 515

Graetz, Brian, and Ian McAllister. 1994. Dimensions of Australian Society. South Melbourne:
Macmillan Education AU.
Güveli, Ayse, Ariana Need, and Nan Dirk de Graaf. 2007a. “The Rise of ‘New’ Social Classes Within
the Service Class in The Netherlands: Political Orientation of Social and Cultural Specialists and
Technocrats Between 1970 and 2003.” Acta Sociologica 50 (2): 129–146.
Güveli, AyŞe, Ariana Need, and Nan Dirk de Graaf. 2007b. “Socio-Political, Cultural and Economic
Preferences and Behaviour of the Social and Cultural Specialists and the Technocrats. Social
Class or Education?” Social Indicators Research 81 (3): 597–631.
Husson, Francois, Julie Josse, Sebastien Le, and Jeremy Mazet. 2016. FactoMineR: Multivariate
Exploratory Data Analysis and Data Mining (version 1.33). https://cran.r-project.org/web/
packages/FactoMineR/index.html.
Jackman, Mary R., and Robert W. Jackman. 1973. “An Interpretation of the Relation Between
Objective and Subjective Social Status.” American Sociological Review 38 (5): 569–582.
Jennrich, Robert, and Coen Bernaards. 2014. GPArotation: GPA Factor Rotation (version 2014.11-
1). https://cran.r-project.org/web/packages/GPArotation/index.html.
Jones, Frank L. 1974. “Social Stratification in Australia: An Overview of a Research Program.” Social
Science Information 13 (1): 99–118.
Kemp, David Alistair. 1978. Society and Electoral Behaviour in Australia: A Study of Three Decades.
Brisbane: University of Queensland Press.
Kingston, Paul W. 2001. “The Unfulfilled Promise of Cultural Capital Theory.” Sociology of
Education 74: 88–99.
Lamont, Michèle, and Annette Lareau. 1988. “Cultural Capital: Allusions, Gaps and Glissandos in
Recent Theoretical Development.” Sociological Theory 6 (2). doi:10.2307/202113.
Leigh, Andrew. 2013. Battlers & Billionaires: The Story of Inequality in Australia. Carlton: Black Inc.
Linzer, Drew, and Jeffrey Lewis. 2014. PoLCA: Polytomous Variable Latent Class Analysis (version
1.4.1). https://cran.r-project.org/web/packages/poLCA/index.html.
Marks, Gary N. 1999. “The Measurement of Socioeconomic Status and Social Class in the LSAY
Project.” Technical Paper No. 14. Longitudinal Surveys of Australian Youth (LSAY) Technical
Reports. Australian Council for Educational Research. http://research.acer.edu.au/cgi/
viewcontent.cgi?article=1027&context=lsay_technical.
McMillan, Julie, Adrian Beavis, and Frank L. Jones. 2009. “The AUSEI06 A New Socioeconomic
Index for Australia.” Journal of Sociology 45 (2): 123–149.
Mendolia, Silvia, and Peter Siminski. 2015. “New Estimates of Intergenerational Mobility in
Australia.” IZA DP No. 9394. Discussion Paper Series. Institute for the Study of Labor (IZA).
Mills, Colin. 2014. “The Great British Class Fiasco: A Comment on Savage et Al.” Sociology 48 (3):
437–444.
Mills, Colin. 2015. “The Great British Class Survey: Requiescat in Pace.” The Sociological Review 63
(2): 393–399.
Oesch, Daniel. 2008. “The Changing Shape of Class Voting.” European Societies 10 (3): 329–355.
Piketty, Thomas. 2014. Capital in the Twenty-First Century. Cambridge: Harvard University Press.
Prieur, Annick, and Mike Savage. 2013. “Emerging Forms of Cultural Capital.” European Societies
15 (2): 246–267.
Revelle, William. 2016. Psych: Procedures for Psychological, Psychometric, and Personality Research
(version 1.6.9). https://cran.r-project.org/web/packages/psych/index.html.
Robinson, Robert V., and Jonathan Kelley. 1979. “Class as Conceived by Marx and Dahrendorf:
Effects on Income Inequality and Politics in the United States and Great Britain.” American
Sociological Review 44 (1): 38–58.
Savage, Mike, Fiona Devine, Niall Cunningham, Mark Taylor, Yaojun Li, Johs Hjellbrekke, Brigitte
Le Roux, Sam Friedman, and Andrew Miles. 2013. “A New Model of Social Class? Findings from
the BBC’s Great British Class Survey Experiment.” Sociology 47 (2): 219–250.
Savage, Mike, Michael Savage, James Barlow, and Peter Dickens. 1995. Property, Bureaucracy and
Culture: Middle-Class Formation in Contemporary Britain. Oxford Psychology Press.
516 J. SHEPPARD AND N. BIDDLE

Sheppard, Jill, and Nicholas Biddle. 2015. “Social Class in Australia: Beyond the ‘Working’ and
‘Middle Classes’.” 19. ANUpoll. Canberra: The Australian National University. http://politicsir.
cass.anu.edu.au/research/projects/electoral-surveys/anupoll/social-class.
Silva, Elizabeth B. 2015. “Class in Contemporary Britain: Comparing the Cultural Capital and Social
Exclusion (CCSE) Project and the Great British Class Survey (GBCS).” The Sociological Review
63 (2): 373–392.
Throsby, David. 1999. “Cultural Capital.” Journal of Cultural Economics 23 (1–2): 3–12.
Waal, Jeroen van der, Peter Achterberg, and Dick Houtman. 2007. “Class Is Not Dead – It Has Been
Buried Alive: Class Voting and Cultural Voting in Postwar Western Societies (1956–1990).”
Politics & Society 35 (3): 403–426.
Weber, Max. 1946. “Class, Status, Party.” In From Max Weber: Essays in Sociology, edited by Weber,
Max, 180–195. London: Routledge & Kegan Paul.
Western, John. 1991. “Class in Australia: The Historical Context.” In Class Analysis and
Contemporary Australia, edited by Janeen Baxter, Michael Emmison, and John Western, 14–
37. South Melbourne: Macmillan Education AU.
Whiteford, Peter. 2014. “Australia: Inequality and Prosperity and Their Impacts in a Radical
Welfare State.” In Changing Inequalities and Societal Impacts in Rich Countries: Thirty
Countries’ Experiences, edited by Brian Nolan, Wiemer Salverda, Daniele Checchi, Ive Marx,
Abigail McKnight, István György Tóth, and Herman G. van de Werfhorst, 48–70. Oxford:
Oxford University Press.
Wilkins, Roger. 2014. “Evaluating the Evidence on Income Inequality in Australia in the 2000s.”
Economic Record 90 (288): 63–89.
Wright, Erik Olin. 1985. Classes. London: Verso.
Wright, Erik Olin. 2005. Approaches to Class Analysis. Cambridge: Cambridge University Press.

Appendix to ‘Class, capital and identity in Australian society’ Australian

Journal of Political Science 2017

This table presents the factor loadings from the MCA used to create the scale measures of ‘emer-
ging’ and ‘highbrow’ cultural capital, which are subsequently used as measures in the latent class
analysis model that identifies the six classes within Australian society. The results from this table
justifies the structure of the two measures of cultural capital.
Variable loadings for five dimensions extracted using MCA to create measures of ‘emerging’ and
‘highbrow’ cultural capital.

Dimension 1 Dimension 2
(‘Emerging’) (‘Highbrow’) Dimension 3 Dimension 4 Dimension 5
Go to opera 0.003 0.452 0.022 0.019 0.000
Listened to rock/indie 0.308 0.054 0.090 0.081 0.028
Gone to gigs 0.379 0.022 0.001 0.152 0.010
Played video games 0.206 0.077 0.197 0.090 0.015
Watched sports 0.092 0.000 0.392 0.326 0.003
Seen plays/gone to theatre 0.182 0.285 0.002 0.007 0.163
Exercised/gone to gym 0.254 0.004 0.151 0.066 0.120
Used Facebook/Twitter 0.244 0.051 0.169 0.002 0.131
Listened to classical music 0.026 0.317 0.000 0.055 0.281
Watched dance or ballet 0.139 0.276 0.010 0.140 0.000
Listened to hip-hop/rap 0.355 0.040 0.026 0.035 0.117