The New Politics of Class

The New Politics of Class

The Political Exclusion of the British
Working Class

Geoffrey Evans and James Tilley

1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Geoffrey Evans and James Tilley 2017
The moral rights of the authors have been asserted
First Edition published in 2017
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2016949255
ISBN 978–0–19–875575–3
Printed in Great Britain by
Clays Ltd, St Ives plc
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Acknowledgements

Over the course of this project, we have received many valuable suggestions
from colleagues that have helped us improve the analysis and the arguments
presented in the different chapters of this book. We would like to extend a
special thanks to Rob Ford and Oliver Heath who gave us very helpful com-
ments on several chapters, as well as John Goldthorpe for his warm support
and encouragement.
The key to a project of this sort is being able to examine an unusually wide
range of themes in British politics and society using high-quality data. For this
purpose many people have shared data with us and we would like to thank
them for their generosity. In Chapter 2 we examine social inequalities in detail
and are very grateful to: Mark Williams for sharing with us the New Earning
Survey data on inequalities in pay; Felix Busch for supplying us with data
from the Labour Force Survey on unemployment; and Erzsebet Bukodi for the
social mobility data from the cohort studies. For Chapter 5, extracts from The
Guardian have been reproduced courtesy of Guardian News & Media Ltd,
extracts from The Times courtesy of News Corp UK & Ireland Ltd, and extracts
from the Daily Mirror courtesy of Trinity Mirror PLC. In Chapter 6 we use
a variety of data that other very helpfully gave to us. Alan Finlayson
and Judi Atkins kindly sent us Churchill’s conference speeches, which are
not publically available. Mads Thau supplied us with data on class appeals
coded directly from party manifestos. Ian Budge and Judith Bara very swiftly
supplied us with the latest 2015 manifesto data to include in our overtime
analysis. And Jennifer van Heerde-Hudson and Rosie Campbell generously
gave us their data on MPs’ biographical information (these data were collected
with the support of the Leverhulme Trust (RPG-2013-175)).
We gratefully acknowledge financial support from the Jesus College Major
Research Grants Fund and the Danish Research Council (grant number 1327-
00113) for helping to fund the 2015 British Social Attitudes module on social
class used in Chapter 3. We would also like to thank Rune Stubager for his
excellent collaboration in designing and helping to find funding for this
module. And Geoff would like to thank his colleagues in the 2015 BES team
for their patience while this book was completed, as well as support for the
Acknowledgements

inclusion of relevant questions in the face-to-face survey that are used in

Chapters 3 and 6.
We were also fortunate to have a wide range of research assistance from
many colleagues and students. Chris Prosser undertook word score coding
of both the party manifestos and party leaders’ speeches in Chapter 6. Sarah
Coombes coded the media data used in Chapter 5, and gave us many
useful suggestions for the qualitative analysis. William van Taack coded
MPs’ biographical information in Chapter 6, the occupational class codes for
the 2015 BES face-to-face survey and undertook some of the analysis for
Chapters 2 and 3. Zack Grant and Stuart Perrett coded the open-ended class
responses used in Chapter 3. Jon Mellon provided us with figures from the BES
surveys in Chapter 6 and some tables in Chapter 3. Noah Carl proofread and
indexed the book as well as providing more general assistance. Many thanks to
all of them for their invaluable help.
Finally, we would like to thank Allen and Irene Evans for providing part of
the motivation for wanting to tell this story, and Claire Vickers, Elizabeth
Tilley, and Roger Tilley for proofreading the whole manuscript and listening
to James’ grumbling.

vi
Contents

List of Figures ix
List of Tables xiii

1. Introduction 1

Part I. Social Continuity

2. Inequality 21

3. Identity 40

4. Ideology 59

Part II. Political Change

5. The ‘Papers’ 91

6. The Parties 116

Part III. Consequences

7. Class Politics Is Dead 145

8. Long Live Class Politics 170

9. Conclusion 191

10. Postscript 201

References 209
Index 233
List of Figures

1.1. Proportion of different occupational class groups 7

1.2. Proportion of different educational groups 8
2.1. Mean percentile income by occupational class 23
2.2. Proportion in top 40 per cent of the income distribution by
occupational class 24
2.3. Unemployment rate by occupational class 25
2.4. Income and unemployment by education 28
2.5. The experience of various risks by occupational class and education 30
2.6. Explanations for ‘getting ahead’ by occupational class 35
3.1. Unprompted and prompted class identities 43
3.2. Unprompted class identity by occupational class 45
3.3. Prompted class identity by occupational class 46
3.4. Working class identity by father’s occupational class 47
3.5. Class identity by education 48
3.6. Perceptions of class conflict 54
4.1. Views on ownership by occupational class 62
4.2. Views on equality by occupational class 64
4.3. Views on labour relations by occupational class 66
4.4. Union membership over time by occupational class 67
4.5. Strongly against more immigration by occupational class and education 71
4.6. Against EU membership by occupational class and education 72
4.7. Support death penalty by occupational class and education 74
4.8. Strongly support stiffer sentences by occupational class and education 75
4.9. Agree homosexuality is wrong by occupational class and education 77
4.10. Views on censorship and young people by education 78
4.11. How different groups have moved positions on the economic left–right
and social conservative–liberal dimensions 81
4.12. How different groups have moved positions on the
immigration/EU dimension 83
List of Figures

5.1. Proportion of readership supporting the Conservatives 93

5.2. Proportion of people in different groups reading different types
of newspapers 95
5.3. Proportion of people reading different types of newspapers 96
5.4. Proportion of editorials mentioning class in any form 98
5.5. Proportion of editorials mentioning class in specific ways 99
5.6. Proportion of editorials mentioning class conflict and social equality 100
6.1. The left–right position of parties 118
6.2. References to ‘the working class’ in manifestos 120
6.3. References to unions, the unemployed, and the poor in manifestos 121
6.4. References to families in manifestos 122
6.5. References to ‘the working class’ in speeches by party leaders 124
6.6. References to unions, the unemployed, and the poor in
speeches by party leaders 125
6.7. References to families in speeches by party leaders 126
6.8. MPs who previously had working class jobs 128
6.9. MPs who went to university 129
6.10. MPs who attended Oxbridge or a private school 129
6.11. Perceived differences between the parties 131
6.12. Perceptions of the extent to which parties look after the
interests of classes 132
6.13. The proportion of people in Labour constituencies who agree
that Labour is somewhat/very working class given the social
characteristics of their MP 135
A6.1. The comparative manifesto project estimate of left–right party
positions (RiLe) 141
7.1. Labour support by occupational class 149
7.2. Conservative support by occupational class 150
7.3. Differences in Labour support between working class and middle
class groups 151
7.4. Liberal support by occupational class 154
7.5. Labour support by occupational class controlling for left–right values 158
7.6. Labour support by economic left–right position 159
7.7. Liberal support by social liberal–conservative position 160
7.8. Labour support by occupational class controlling for left–right
values and class perceptions of Labour 162
8.1. Turnout over time 171
8.2. Non-voting rates by occupational class and education 173

x
 List of Figures

8.3. Non-voting rates by combined social group 174

8.4. The proportion of people who voted Labour in 1997 and think
that Labour ‘looks after working class people’ well in 2001 by
occupational class and 2001 vote choice 177
8.5. SNP support by occupational class 180
8.6. UKIP support by occupational class and education in England and Wales 183
8.7. UKIP support by occupational class and education in England
and Wales controlling for attitudes to immigration and the EU 185
9.1. Social make-up of Labour and Conservative voters 196

xi
List of Tables

1.1. Occupational class groups 4

1.2. Educational groups 6
2.1. Perceived mean earnings of different jobs 31
2.2. Perceived mean earnings of different jobs by occupational class 32
2.3. Gap between perceived and deserved earnings by occupational class 33
2.4. Beliefs about what is important for ‘getting ahead’ 34
3.1. The closeness of class identity 49
3.2. The meaning of class 51
3.3. Specific types of people associated with classes 52
3.4. Pride in being working class 53
3.5. Perceptions of differences between classes 55
3.6. Difficulty of moving between social classes 55
3.7. Awareness of social class and difficulty of having friends from
other social classes 56
4.1. Impact of occupational class on attitudes towards ownership 63
4.2. Impact of occupational class on attitudes towards equality 65
4.3. Impact of occupational class on attitudes towards labour relations 67
4.4. Impact of occupational class and education on attitudes
towards immigration 73
4.5. Impact of occupational class and education on attitudes towards the EU 73
4.6. Impact of occupational class and education on attitudes towards
the death penalty and stiffer criminal sentences 76
4.7. Impact of occupational class and education on attitudes towards
homosexuality, young people, and censorship 79
A4.1. Measurement of independent variables used in models 88
5.1. Proportion of editorials directly associating social class with
specific parties or specific policies 101
6.1. The proportion of people who think their MP is working class
by party and social background of MP 133
List of Tables

A6.1. Manifesto components included in the left–right scale 139

7.1. Impact of occupational class on vote choice in 2015 152
7.2. Impact of occupational class and education on Liberal support 156
A7.1. Impact of occupational class and education on Labour support 168
A7.2. Impact of occupational class and education on Conservative support 168
8.1. How 1997 Labour voters voted in 2001 by occupational class 175
8.2. How 2005 Labour voters voted in 2010 by occupational class 176
8.3. SNP vote share by occupational class and education 179
8.4. UKIP support by occupational class in England and Wales 184
9.1. Judgements of party leaders 195
10.1. Proportion of people voting Leave in the 2016 EU referendum 202
10.2. Proportion of people voting Leave in the 2016 EU referendum 203
10.3. Proportion of people voting 205

xiv
1

Introduction

There was a time when, odd though it may seem today, British politics was
focused on the needs and desires of the working class. Post-war Britain saw
the emergence of not just the welfare state but also a determination by
politicians of all parties never to return to the mass unemployment of the
1930s. Both these developments were implemented in the name of the
working class. A form of corporatism was adopted by both Conservative
and Labour governments so that working class interests were represented
politically and the ‘forward march of labour’ was consolidated as part of the
‘long revolution’ (Williams 1961; Westergaard 1995). It was the working
class ‘whose claims could henceforth never be ignored by governments’
(Marwick 1980, p.229). This emphasis was so engrained that The Middle
Class Vote, published in the 1950s, actually labelled the middle class ‘the
not working class’ and argued that the middle class were ‘the class without a
party’ (Bonham 1954, p.30).
This seems a long time ago now. The economic structure of Britain today is
quite different to that of the 1950s. There has been radical deregulation of the
economy, the trade union movement has become a shadow of its former self,
and, most importantly, there are simply many more people with middle class
jobs. The rise of the middle class has often been seen as indicating the end of
both class division and of the political importance of class. Shrinking divisions
between social classes are seen as a natural consequence of deindustrialization,
increased affluence, greater welfare provision and the breakdown of trad-
itional class communities. At the extreme, it is claimed that ‘the dynamism
of the labour market backed up by the welfare state has dissolved the social
classes’ (Beck and Beck-Gernsheim 2002, p.203). We are thus left with an
amorphous social structure devoid of class difference: a society in which
everyone is middle class or has no class at all.
While the social science literature of the 1950s, 1960s, and 1970s talks of
little else but the working class (Willmott and Young 1960; Dennis et al. 1956;
The New Politics of Class

Hoggart 1957; Roberts 1971),1 this is no longer the case. By the turn of the
millennium, British sociologists appeared to have turned their backs on the very
idea of class: ‘The study of class is no longer central to British sociological
analysis, and the debate on class is largely about whether this should be cele-
brated or lamented’ (Savage 2000, p.7).2 But has the class structure dissolved? We
argue that it has not. In support of this argument, we examine class in Britain
over the last fifty years and how it has both shaped and been shaped by social and
political change. We focus on the working class, not as the left’s traditional
harbingers of revolution and social transformation, but as an increasingly mar-
ginalized political group. We show while the size of class groups has changed,
there are remarkably stable class divisions in values and policy preferences. Class
division thus remains a key element of Britain’s political picture, but in a new
way. Whereas working class people once formed the heart of the class structure
and the focal point of political competition, they now lack political representa-
tion. This is because the political environment has changed. Parties have reacted
to changing class structures by changing their ideology, policy programmes,
rhetoric, and elite recruitment strategies. Vote-seeking parties now focus on the
middle class, not the working class, and it is the working class, not the middle
class, that has become Bonham’s ‘class without a party’. That, in essence, is the
argument of this book. In this introductory chapter, we lay out precisely these
mechanisms of change and the evidence that we use to show these processes at
work. First, however, we explain what we mean by class.

What Is Class?

Social class is one of the most widely discussed, and disputed, concepts in
social science. Characterizations of class position have included numerous
occupational classifications, including distinctions between manual versus
non-manual workers and owners versus employees, status rankings, income
levels, educational levels, subjective class identifications and lifestyles. Some-
times two or more of these have been combined.3 Much ink has been spilled
over whether class positions are best thought of as ‘relational’ or ‘gradational’
(Ossowski 1963) and even more on whether a given measure of class position
truly reflects the ideas of Marx or Weber.4 We have no desire to immerse the
reader in disputes about which idea of class is ‘best’. In part, this is because
different measures of class position are often closely correlated, but also
because this debate is fundamentally unresolvable. As Calvert (1982, p.214)
observed, class is an ‘essentially contested concept’, due in no small part to its
great political and social significance.
This means that no one characterization of class is definitive. Measures of
class position are useful to the degree that they allow us to demonstrate

2
 Introduction

important relationships between social position and outcomes. We do not

therefore spend time comparing different class characterizations or measures.5
We focus on occupation, and to a lesser extent education, as key measures of
where people are positioned in the class structure. Occupation matters
because it determines people’s current and future earnings, and the security
of that employment. Writing shortly after the Second World War, at the very
start of the period we are examining, Ferdynand Zweig argued that:

Security is one of the basic differences between the working class and the middle
class, but the working man not only lacks security of employment but also of
earnings. His earnings depend not only on hours of work but also on overtime,
shifts, changing conditions of work, piecework, individual and collective bonuses,
the materials and machines he handles. (Zweig 1952, pp.203–4)

These same themes of insecurity in both job tenure and earnings remain the
key aspects of occupation that concern social scientists today (Goldthorpe
2004, 2007; Rose and Pevalin 2003). It is less important to know how much
someone earns at any given point in time than how they earn that money,
and the security, promotion prospects, and job uncertainty that characterize
their conditions of employment.6 The measure of occupational class that
informs our approach is one that not only focuses on divisions between occu-
pations but is also an established and widely used measure of class position. The
Goldthorpe schema (Erikson and Goldthorpe 1992; Goldthorpe 2004, 2007)
has been carefully developed with respect to what it is about occupations
that enables them to be clustered into class positions. It has been shown to
be associated with many important aspects of people’s lives, and has been
extensively validated as a way of characterizing occupational divisions and
understanding social and political change.7
We use a slightly modified version of this schema to allow us to track the
social and political impact of changing occupational class sizes most accur-
ately. Table 1.1 shows the key groups that we focus on and how we translate
the nineteen-category socio-economic group (SEG) classification into mean-
ingful social classes.8 Table 1.1 also provides illustrative examples of both
more traditional and newer types of jobs included in the classes. Although
almost all the traditional jobs are still to be found, new occupations within
those classes have obviously emerged. For people not currently working, due,
for example, to unemployment or retirement, we use their last occupation,
and for those who have never had a job, or who were unable to be assigned to a
SEG, we use their husband’s or wife’s occupation. We separate people into
three different middle class occupational groups and one working class group.
The working class in our analyses is composed of skilled, semi-skilled, and
unskilled workers in manual occupations. It also includes agricultural workers
(but not farmers). On average, these working class jobs offer lower and more

3
The New Politics of Class

Table 1.1. Occupational class groups

SEG categories Traditional jobs Newer jobs

Old Middle Class Managers large, managers Manager, small business Self-employed website
small, self-employed with owner with employees, designer, HR manager
employees, self-employed shopkeeper, barrister, farmer
professionals, farmers
New Middle Class Employed professionals, Architect, teacher, university Occupational therapist,
intermediate non-manual lecturer, nurse, social worker dietician, paramedic,
dental hygienist
Junior Middle Class Junior non-manual Bank cashier, clerk, secretary, Legal assistant, dispatch
typist technician
Own Account Self-employed without Shop owner with no
employees employees, self-employed
own account plumber
Personal Service Personal service workers Nursery assistant, assistant, Personal care assistant,
chauffeur tour guide
Foreman Foremen and supervisors Lead hand, production Packing manager
supervisor, construction
foreman
Working Class Skilled manual, semi-skilled Machine operator, Packer, order picker, HGV
manual, unskilled manual, seamstress, warehouseman, driver, gardener, waste
farm worker quarry worker, miner, farm treatment officer
labourer

Note: People in the armed forces are assigned to the old middle class if they have management responsibilities (i.e. are
officers), and the working class if not. People with no current occupation are assigned to a class by their previous
occupation, or if no previous occupation their spouse’s occupation.

insecure incomes than middle class jobs. They also tend not to offer guaran-
teed sick pay, generous pensions, or clearly-defined promotion opportunities,
while also involving more supervised monitoring, less autonomy, less hourly
flexibility, and more unpleasant working conditions. These jobs may no
longer be heavily concentrated in traditional heavy manufacturing, but
they have many similar constraints and disadvantages. As we shall see in
Chapter 2, the differences in resources, prospects and security between people
in different classes have not changed (Felstead et al. 2015). The people who
would once have been found on a factory shop floor are now more likely to be
found in service jobs, but are still struggling to get by on erratic and insecure
incomes. Indeed class differences in pay may have actually increased (Gallie
2015) as labour markets have become polarized with fewer jobs in the middle
and more at the top and bottom (Autor and Dorn 2013; Oesch 2013).
Compared with the working class, middle class workers occupy relatively
secure salaried positions, often with occupational pensions and other benefits.
These are generally regarded as the most desirable positions in the labour
market, although there will be considerable variation between those in senior
and junior posts. We distinguish between three main middle class occupa-
tional groups, which we term the old middle class, the new middle class, and

4
 Introduction

the junior middle class. This is a slight departure from the initial formulation
of the Goldthorpe scheme.9 We do this to capture changes in the middle class
occupational structure. Importantly, we therefore separate professionals from
managers (see Hout et al. 1995).10 The old middle class was the dominant group
within the middle class immediately after the war. This group includes man-
agers, but also small employers, farmers, and self-employed professionals. The
new middle class is the dominant group within the middle class today: profes-
sional employees, ancillary non-manual workers, and non-manual supervisors.
Finally, we have the junior middle class: this comprises junior non-manual
employees, or what are typically referred to as ‘routine white collar workers’.
Finally, there are also three other smaller groups, which we include in our
statistical models, but rarely show in any tables. These are the people who fall
into the SEG categories of own account workers, foremen and supervisors, or
personal service workers. These are quite small (each around 5 per cent of
population) and very heterogeneous categories. The changing composition in
the sorts of people who are own account workers makes over-time compari-
sons particularly problematic.11 This heterogeneity is also true for personal
service workers and, to a lesser extent, foreman and technicians. For these
reasons, we pay less attention to these relatively small groups, although as a
rule of thumb personal service workers and foremen are most similar in
attitudes and behaviour to our working class group.
Occupation is important, but we are also interested in education. Educa-
tional qualifications are an increasingly differentiated source of information
about people’s capacities, transferable skills, and potential attainment. This is
especially the case among young people, whose occupational class position is
likely to be less firmly established. Education has also been shown to shape
values and political preferences differently to occupational class (Evans et al.
1996; Tilley and Heath 2007; Chan and Goldthorpe 2007). Significantly, like
the class structure, the distribution of educational qualifications has also
changed radically in a relatively short period of time. In analyses reported in
later chapters we measure education mainly through the level of qualifications
someone has obtained. Some distinctions are quite easy to make. We can
identify people with degree-level qualifications straightforwardly in our data
and can also fairly easily distinguish people who either left school at the
minimum age (fourteen until 1947, fifteen until 1972, and sixteen after-
wards), or gained no qualifications. These groups form the top and bottom
of our education categories. For the longest time span, we use a combination
of school leaving age (compared to when someone went through the educa-
tional system) and higher education level qualifications to form a five category
measure. This is the second column in Table 1.2. For all our data from 1979
onwards, we can use a better measure of highest qualification. This is shown in
the third column of Table 1.2. Generally, the key distinctions we make are

5
The New Politics of Class

Table 1.2. Educational groups

Education (5 groups) Education (7 groups)

High Degree or above Degree or above

Some higher education Some higher education
Medium Left school at 17/18 A Level or equivalent
Left school above minimum leaving age, but before 17/18 O Level or equivalent
CSE or equivalent
Apprenticeship
Low Left school at minimum leaving age No qualifications

between people with what we call high, medium, and low levels of education.
High relates to degree level education, medium to A Level equivalent educa-
tion, and low to no qualifications or leaving school at the minimum age.

The Changing Shape of the Class Structure

The last half-century has seen a pronounced growth in the size of the middle
class. Although this transition is well known, it is useful to map out what it
looks like in terms of our class categories. This is shown in Figure 1.1. The top
graph in the figure contains census data back to 1931.12 The bottom graph
presents the patterns of change captured in survey data from the British
Election Study (BES) surveys which run from 1963 to 2015, and the British
Social Attitudes (BSA) surveys which cover the period from 1983 to 2015.
Given that the latter are social survey estimates based on a few thousand
people, there is a reassuring degree of similarity in the trends from these
different sources. The growth of the middle class, and especially the new
middle class, is clear, as is the steady decline of the working class. By 2000
the new middle class had become larger than the working class in both the
BES/BSA and census series.
The rise of the new middle class is closely linked to educational changes.
Many new middle class jobs are primarily professional or semi-professional
in nature and have increasingly required formal qualifications for entry.
The attainment of educational qualifications, especially a degree, has risen
accordingly. This is, in part, because of deliberate government strategies to
improve skill levels in Britain’s increasingly knowledge-based economy. As
with occupational class, we display these trends using both official sources of
data and the survey data used for most of our analyses in later chapters of the
book. We show the changes in our high, medium, and low categories of
education in Figure 1.2 using data from the Labour Force Survey (LFS) from
1979–2015 and the BES/BSA data from 1964–2015.13 The first and last of those
groups have changed in size enormously over time. In the 1960s the number

6
 Introduction

a) Census data

60% WC

40%
NMC
JMC
20% OMC

0%
1930 1950 1970 1990 2010

b) BES and BSA data

60%

WC
40%
NMC
JMC
20%
OMC

0%
1930 1950 1970 1990 2010

Figure 1.1. Proportion of different occupational class groups

Note: The figures here show the proportion of people in different occupational class groups. The top graph shows
occupational class from UK census data; the bottom graph shows occupational class from combined BES and BSA
data (three period moving average). Four occupational class groups are displayed: old middle class (OMC), new
middle class (NMC), junior middle class (JMC), and working class (WC).
Source: UK Census 1931–2011; British Election Studies 1963–2015; British Social Attitudes Surveys 1983–2015.

of people with a degree was tiny, a few per cent of the population, and around
two thirds of people had left school at the minimum age (which for most
people then would mean fourteen). Today, around a quarter of the electorate
has a degree, and a quarter have no qualifications. As with occupation the
educational structure of Britain has been transformed.

The Politics of Class

What do these changes mean for politics? Most discussions of the changing
nature of class politics have talked about the transformation of society and the
demise of class divisions. We agree that there has been transformation, but not
because class divisions have disappeared. The changing politics of class has
resulted from the ‘top down’ influence of politicians and the media. This book

7
The New Politics of Class

a) LFS data

60% Low

40%
High
20%
Medium

0%
1960 1970 1980 1990 2000 2010

b) BES and BSA data

60% Low

40%

Medium
20%

High
0%
1960 1970 1980 1990 2000 2010

Figure 1.2. Proportion of different educational groups

Note: The figures here show the proportion of people in different educational groups. The top graph shows
education from LFS data; the bottom graph shows education from combined BES and BSA data (two period
moving average). Three educational groups are displayed: people with degree-level education (high), people
with A Level equivalent education (medium), and people who left school at the minimum school leaving age
for their cohort or have no qualifications (low).
Source: UK Labour Force Surveys 1979–2015; British Election Studies 1963–2015; British Social Attitudes Surveys 1986–2015.

develops our case with evidence on two contrasting features of British society: social
continuity and political change. By social continuity we mean objective inequal-
ities between classes, perceptions of class identities, and class divisions in social and
political attitudes. These social divisions have remained remarkably unchanged
despite Britain’s transition from an industrial to a post-industrial society. Britain
remains a class-divided society. By political change we refer to change, not among
voters, but among parties, politicians, and the media. The media’s representation of
class, the policies the mainstream parties offer, the groups that parties talk about,
and the composition of party elites have all changed dramatically.

Social Continuity
An emphasis on the decline of class is usually part of a theoretical focus on the
decline of structural divisions more broadly. Changes in the class basis of

8
 Introduction

politics are thought to have derived from general economic developments

leading to the emergence of less structured societies. The argument runs that
increased social mobility, affluence, and educational expansion have weak-
ened the distinctiveness of classes in a globalized world. This has meant that
traditional social groups, such as classes, have declined in political significance
(Clark and Lipset 1991; Pakulski and Waters 1996a; Clark 2001a, 2001b).14
Underlying much of this work is the key assumption that classes are no longer
monolithic sources of identity and interests and that ‘few individuals now
possess exclusively middle-class or working-class social characteristics, and the
degree of class overlap is increasing over time’ (Dalton 2008, p.156). At its
most extreme this process has been thought of as a form of ‘structural dissol-
ution’. This means a process of individualization, where society is composed
of individuals whose identities and interests are so multi-layered as to render
social categories redundant (Beck and Beck-Gernsheim 2002). Similarly, Beck’s
earlier (1992) formulation of the ‘risk society’ proposes that risk is ubiquitous
in modern societies and the differential experience of risk between classes has
weakened. Ultimately all of these depictions of bottom up change are about
social blurring, social atomization, and increasing social heterogeneity.
Why might these influential ideas be wrong? One answer is that the assump-
tion that classes are more alike is simply incorrect. The blurring assumed
in terms of income and quality of life has not happened. As we shall see in
Chapter 2, claims about the demise of class differences and divisions do not fit
with the plentiful evidence of continuing class-based inequalities of resources,
opportunities, and risks. In this chapter we examine trends of inequalities
in income, unemployment, health, and educational achievement, as well as
levels of inter-generational social mobility between classes. Across these many
areas an extensive body of research confirms the continuation of pronounced
class inequalities. Chapter 2 also shows that people are aware of these inequal-
ities, both in their own lives and more generally.
Interestingly, perceptions of the incomes of different occupations actually
underestimate the extent of inequality and the degree to which it has increased.
This does not, however, mean that people are unaware of class or lack a sense of
being in a class. Nonetheless, the assumption that the working class, in particular,
has lost its identity runs through many recent narratives. One strand focuses on
the idea of ‘dis-identification’ and another on the idea that everyone now thinks
that they are middle class. Given the evidence on continuing class inequality
presented in Chapter 2, it would be surprising if people did think in that way and
in Chapter 3 we show the resilience of class identities and people’s persistent
awareness of class division. Levels of class identification remain robust over more
than fifty years and people remain aware of class in both societal and everyday
contexts. There has been little change in perceptions of class conflict and barriers
between classes, or even in the difficulty of having friends from other social

9
The New Politics of Class

classes. People also seem to define classes in fairly similar terms. There is almost
the same strong emphasis on occupation as there was in the 1960s, and both
income and education are at least as important as then. In short, class awareness
remains a part of British society. Changing class sizes and increased upward
mobility have not weakened people’s sense of class position and class division.
This pattern of continued distinctive classes is also evident when consider-
ing social and political attitudes, as we do in Chapter 4. Persistent class divi-
sions in resources, risks, opportunities, and educational attainment foster
continuing differences in political preferences. That inequality, insecurity,
and limited opportunities should lend themselves to continued class divisions
in support for different economic policies is unsurprising from a rational
choice perspective. Likewise, when we consider issues and policies associated
with socially liberal values, the idea that the middle class is less authoritarian
than the working class is not new (Lipset 1959), and has been established
across many societies (Napier and Jost 2008). Overall we find that economic
issues are still related to class, with the middle classes wanting less interven-
tionist and redistributive policies, and issues such as immigration and the EU
are also powerfully divisive along class and education lines.
The message from all of these chapters is one of continuity. Objective
inequalities, perceptions of those inequalities, awareness of class position
and divisions, and the political ideologies of the different classes largely
remain unchanged in Britain. The changes in the sizes of classes, however,
have had dramatic consequences.

Political Change
While the differences between social classes may not have changed, the chan-
ging size of those classes has had a pronounced impact on the political parties.
Labour decided in the 1990s that it could no longer base its electoral success
primarily on working class support. This led to the party attempting both to
change perceptions of itself as a working class party, and also to move its policy
stance rightwards towards the Conservatives. This convergence process had at
its heart a strategic focus on the median voter as the basis for electoral success. In
addition, the higher ranks of the political parties have altered. Both have become
increasingly dominated by professional career politicians with middle class
backgrounds and a university education (Heath 2015).
These changes have also been accompanied by a shift within the most polit-
ical branch of the media: the national newspapers. Information about politics
and how it relates to people’s own positions in society is obtained primarily from
the mass media. How newspapers portray class and its relationship to politics is
therefore important. This is not to say that newspapers tell people what to think:
readers to a large degree select what they want to consume and inevitably this

10
 Introduction

process will be more about reinforcement than persuasion.15 But the media can
still shape the way in which people understand the relationship between politics
and class through the messages they see on a daily basis.
In this section of the book we therefore describe ‘a perfect storm’ whereby
electoral strategy has resulted in the convergence of political parties on the
middle class voter and the exclusion of the preferences of working class people
from the political mainstream, which in turn has been amplified by the
disappearance of class politics from the press. In Chapter 5 we analyse how
the newspapers have changed their representation of class. We show that they
have moved from a portrayal of the working class as the main reference group,
with an acceptance of class as a fundamental aspect of society, to the news-
papers of today that rarely mention class. In effect, the structuring of the
political world in class terms by the newspapers no longer exists.
In Chapter 6 the convergence of the main parties’ ideology and political
rhetoric is described in detail via systematic examination of manifestos and
leader speeches since 1945. These primarily quantitative analyses reveal con-
sistent patterns of change in policy and group appeals. In terms of policy, we
show that while there was similarity between the parties at the end of the
1960s, this is sometimes exaggerated, and that policy convergence in the post-
war period only really happened in the 1990s. Class group appeals show an
even clearer pattern. References to the working class were standard practice by
Labour (and even the Conservatives) in the post-war era, but started to fall
dramatically from the late 1980s onwards. After that point class effectively
disappeared from the lexicon of party politics. As with policy convergence,
these changes were part of the re-branding of Labour as ‘New’ in the early
1990s. Finally, we show that at this time political recruitment started to be
dominated on both sides of Westminster by degree-educated members of the
middle class. The two parties grew alike in policy, in their rejection of class
appeals, and in their social composition. Crucially Chapter 6 also shows that
the electorate noticed these changes as they happened. As a result of these
combined transformations, the connection among policy, party, and class is
now almost absent in mainstream British politics.
In direct contrast to the pronounced stability of economic, social, and
attitudinal divisions between classes, Chapters 5 and 6 show that the party
messages to voters, and their mediation via the press, changed enormously.
Class was consigned to political history not by the blurring of classes, but by
the actions of politicians.

The Consequences of Social Continuity and Political Change

What does the combination of bottom up social continuity and top down
political change mean for class voting? Traditionally, class occupied a central

11
The New Politics of Class

position in British voting behaviour (Alford 1964; Butler and Stokes 1969).
Indeed much effort was spent in trying to explain the specific phenomenon of
anomalous ‘working class Conservatives’ (Nordlinger 1967; McKenzie and
Silver 1968). This emphasis on the potentially powerful role of class fell
quickly under attack, however, as scholars pointed to dealignment as a distin-
guishing feature of partisan divisions in the 1970s (Crewe et al. 1977b; Sarlvik
and Crewe 1983). The intensity of the academic dispute over the role of class
in politics in the 1980s almost paralleled the political disputes in the real
world (Heath et al. 1985, 1987; Dunleavy 1987; Crewe 1987, Franklin 1985;
Rose and McAllister 1986). The conflict in academia focused on whether class
voting had actually declined, with some arguing that we had seen ‘trendless
fluctuation’, not real change. Yet, by the late 1990s, those on the trendless
fluctuation side of the argument effectively conceded that class had declined
as an influence on party choice. The question is why?
Some have argued that class no longer matters for party choice because
there are no longer meaningful class divisions. Unfortunately, this is typically
asserted with no serious analysis of the accuracy or otherwise of this assump-
tion. Clarke et al. (2004), for example, reiterate the claim that the decline of
class voting in Britain is due to the fact that ‘class boundaries have become
increasingly fluid’ (Clarke et al. 2004, p.2) without at any point demonstrat-
ing this is actually the case. We instead emphasize the role of the political elite
in the structuring of class divisions in political choices. Sometimes referred to
as the top down, or supply side, approach, the argument here is that the
strength of social divisions in political preferences derives from the choices
offered to voters by politicians and parties. This is not a new thesis. Przeworski
and Sprague (1986) were influential advocates of this idea, and Converse
argued in the 1950s that when parties failed to adopt distinct positions on
class-relevant issues voters would be unable to use their class positions as
bases for voting (Converse 1958, pp.395–9). Nonetheless, it is only recently
that there has been thorough empirical analysis of the impact of the choices
offered by parties on social divisions in voting.16
There are two aspects to political change that we might care about. The first
is whether parties take different policy positions. Policy convergence weakens
the motivation for choosing parties as a result of class interests or values. Party
policy polarization should in turn accentuate class voting. Voter responses
depend upon the choices voters are offered (the supply side), as well as the
presence of differences in ideological and value preferences within the elect-
orate (the demand side).17 Modern Britain exemplifies the consequences of a
top down political process of change. Labour’s rightward shift in policy in the
1990s resulted in a constrained set of choices for voters, and thus less class
voting. There is a second type of political change, however. Politics is not just
about policy; it is about less instrumentally obvious signals. It concerns the

12
 Introduction

kinds of people or groups that parties refer to and are seen to represent.
Labour’s shift to being a party of the middle class was not just about policy;
it was also about the party’s image. That image derives from policy stances, but
also from the rhetoric and social background of politicians.
In Chapter 7 we examine the first electoral consequence of these changes:
the demise of class voting. Combining different sources of data, we show that
levels of class voting were largely static from the 1940s to the 1990s. Only then
did the relationship between class and party collapse, and in the space of a few
years. These dramatic changes correspond with the pronounced political
changes that we discuss in Chapters 5 and 6. As shown in Chapter 6, policy,
group appeals, and party elites all changed at the end of the 1980s and
beginning of the 1990s. Following this, class voting fell for the first time
since the end of the Second World War.
Moreover we directly link these party changes to people’s individual voting
decisions. We show that the core political issues of ownership and redistribu-
tion, which Chapter 4 reveals to be still fundamentally class-based, lost much of
their power to account for vote choice during the 1990s. Equally important were
changing perceptions of the parties as class parties. The short-lived policy con-
vergence in the 1960s did not change voters’ images of the parties. People still
thought that both the main parties were class parties. This completely altered
under New Labour. In Chapter 7 we use class perceptions of the parties to predict
vote choices, and find that it is the combination of changing policy and chan-
ging perceptions of the parties that explains the decline of class voting in Britain.
Chapter 7 illustrates the fact that while parties can change policy, if
that change does not affect people’s perceptions of the parties, both in policy
and class image terms, then policy change alone will not lead to dramatic and
persistent changes in class voting. Images in that sense are ‘sticky’. Large, abrupt
changes by a party are required to reshape perceptions, and only multiple
concurrent changes produce a transformation of the political equilibrium.
In Britain in the 1990s there was such a shock to the system. Labour radically
changed its nature in a short space of time. Not unrelated to this transform-
ational signal from the party, the newspapers stopped talking about the politics
of class. As a result of these policy and image changes, class voting was funda-
mentally undermined.
There is a further consequence of Labour’s recasting as a party of the middle
classes. Chapter 8 shows how these political changes have caused a new
phenomenon in British politics: class-based abstention from voting. Although
there have always been class differences in turnout in Britain, these differences
have been very small. As we show, working class people were almost as likely
to turn out to vote in the second half of the twentieth century as middle class
people. Britain has therefore differed historically from the US, where occupa-
tion and education have traditionally had strong influences on political

13
The New Politics of Class

participation (Hout et al. 1995; Leighley and Nagler 2014).18 This began to
change in the 2000s in Britain and is directly due to the lack of political choice.
With Labour no longer representing them and their views, working class
people have increasingly chosen not to vote. Labour has followed the earlier
path of the Democrats in the US and effectively created new inequalities in
political participation. These inequalities work to reinforce the declining
representation of working class people since parties are less likely to care
about the preferences of people who do not vote.
There are some political processes that might be thought to counteract this
gloomy prognosis. As the major parties have become more similar we have
seen increasing electoral volatility and an increasing vote for minor parties.
This was brought into focus by the 2014 European Parliament elections and
2015 General Election, which saw the emergence of the United Kingdom
Independence Party (UKIP). As in other European countries, the presence of
a radical right party in Britain is providing a new political voice for the
working class (Ford and Goodwin 2014). Chapter 8 shows that the arrival of
these new political choices has indeed increased class voting, as UKIP, and to a
lesser extent the SNP in Scotland, have a strong appeal to the working class.
Nonetheless, few of the voters drawn to either insurgent party were previously
non-voters and, in that sense, these new party options have failed to renew
working class electoral participation.

The Evidence

Any book is only as good as the evidence it uses to support its arguments. To test
our ideas we need comparable long-term data on many aspects of social structure,
social identities, political attitudes, the mass media, and parties. This distin-
guishes this book from others about class in general and the working class in
particular. Journalists and social commentators such as Owen Jones (2011) and
Ferdinand Mount (2004) have illustrated the extent of continued social snobbery
and the neglect of the working class by politicians, while others such as David
Goodhart (2013) have examined the negative consequences of immigration for
the working class, but they do not systematically examine what the mass of the
population thinks, nor do they test mechanisms that explain the outcomes to
which they refer. A similar limitation characterizes renewed interest in the work-
ing class among British sociologists, which has usually involved small-scale,
impressionistic studies (Skeggs 2004; Lawler 2005; Irwin 2015).
In contrast, our evidence is systematic, comparable over time, and allows us
to test each of the explanatory processes we have proposed. It covers the
experiences, beliefs, and actions of the electorate, as well as the claims and
actions of the parties and the media. Most of the survey evidence is taken from

14
 Introduction

the BES and BSA surveys. These are high-quality, representative probability
samples of individuals living in private households in Britain. Both involve
face-to-face interviews with several thousand people in most waves. The BES
began in 1963 with a pre-election survey, and has been repeated at every
general election since, and the BSA is an almost annual survey starting
in 1983.19 Both surveys are weighted to take into account both differential
probability of selection and known patterns of non-response.
We complement these by introducing pertinent evidence from one-off
surveys such as the Marshall et al. (1988) survey on class in Britain undertaken
in 1984, and a module on social class included in the 2005 BSA survey that we
examine in our analyses in Chapter 3. We also make use of the large sample
size and high quality of the government’s Labour Force Surveys to examine
occupational changes in Chapter 2. Chapter 7 uses Gallup surveys that allow
us to track class voting back to the 1940s, with three surveys from 1945 and
1946 as well as a run of nearly annual surveys from 1955 to 1968. We have also
collected new survey data specifically for this book: this includes a module on
class and politics in the 2015 BSA survey that features in Chapter 3 and a
module of questions on class and representation embedded within the 2015
BES survey that is used in Chapter 6.
In addition to this vast array of survey evidence on the experiences, opin-
ions, and actions of the public, we have coded what politicians, parties, and
the media say. These quantitative and qualitative analyses examine references
to class in three major newspapers, party leaders’ conference speeches, and the
content of party manifestos all stretching back to 1945. We have also obtained
detailed biographical data about individual MPs across the period and link this
information with the perceptions of the voters in their constituencies.
By integrating these extensive sources of systematic comparable data we
hope to demonstrate our argument authoritatively. As noted earlier, when
social scientists have referred to the changing relationship of class and politics
in Britain they have not usually examined evidence that might explain this
change. This is no longer the case.

Conclusions

In the 1960s the working class were pre-eminent demographically, and to a

large degree politically. Their views were those of the majority and the polit-
ical parties competed for their affections. Since then many have argued that
class no longer matters, that the distinctions between classes have blurred and
that individualization is the key attribute of modern, post-industrial societies.
As a result, parties no longer need to connect with differences in interests
between classes: the blurring of social divisions has produced a blurring of

15
The New Politics of Class

political divisions. In direct contrast, we demonstrate in the chapters which

follow that the most significant feature of the post-industrial class structure is
not its disappearance, but the changing sizes of its classes. Inequalities have
not only survived but have in some respects actually increased.
Whereas the point of division in the past was between a large and fairly
homogenous working class and a small middle class, now it is between a larger
group of middle classes and a much smaller working class. At its simplest our
thesis is that class divisions in social attitudes and political preferences remain
robust. It is the political parties that have chosen not to represent these class
differences. This has led to a decline in class voting, but also an accompanying
accentuation of class divisions in non-participation. This is a new class divide
that is unlikely to be reversed.

Notes

1. These were complemented by numerous studies of working class images of society

(Bulmer 1975) and oral histories (Roberts 1984). Jackson and Marsden’s Education
and the Working Class (1966) told of the often uncomfortable relationship between
children from working class backgrounds and grammar schools, and Willis’s Learning
to Labour (1977) focused on less academic groups. Work-place ethnographies also
focused on working class job experiences (Fraser 1968/9; Beynon 1973), while the
landmark study of ‘the affluent worker’ (Goldthorpe et al. 1969) convincingly dis-
confirmed the idea that the working class was becoming more like the middle class.
2. Since then there has been renewed interest in class amongst British sociologists,
though often with respect to subjective elements of class identification and typic-
ally using small-scale ethnographic studies: e.g. Skeggs (2004), the various contri-
butions to a special issue on ‘Class, Culture and Identity’ (Sociology 2005), a report
from the Runnymede Trust (Sveinsson 2009), the work of Savage and colleagues
(Savage et al. 2010) and, most recently, Irwin’s (2015) interviews with people about
job insecurity and the cost of living.
3. Arguably, measures of socio-economic position that combine education and occu-
pation, and often income, into a composite measure of class raise as many ques-
tions as they answer in order to understand the various interconnections between
them (Evans and Mills 1998).
4. Most recently, the BBC’s ‘Great British class survey’ (Savage et al. 2013) garnered a
lot of attention. Most commentators were of the view that class divisions were
pronounced and of great public interest, but there was also scepticism about
whether the classes defined in the BBC study were valid (Mills 2015a).
5. This also avoids the problems of subjectivity that plague self-reported measures of
class position. We treat class identity as a possible product of occupational class
position and education (see Chapter 3).
6. High levels of income churn from year to year also mean that income is un-
likely to measure consistently more or less advantaged social positions within

16
 Introduction

the economic structure. Occupation-based measures are more stable (Connelly

et al. 2016).
7. The Goldthorpe schema class categories have consistently been shown to be related to
differences in employment conditions, job autonomy, income, and life-time expected
earnings (Evans 1992, 1996; Evans and Mills 1998, 2000; Goldthorpe and McKnight
2006). This has led to its adoption as the primary component of the National Statistics
Socio-economic Classification (NS-SEC) now used in the UK Census (Rose and Pevalin
2003), and as the basis for cross-national European research using the European Socio-
economic Classification (Harrison and Rose 2006). It has been used extensively to
study the class basis of social and political preferences in Britain (Heath et al. 1985,
1991, 2001; Marshall et al. 1988; De Graaf et al. 1995; Evans et al. 1999; Evans and
Tilley 2012a, 2012b) and elsewhere (Weakliem 1989; Nieuwbeerta 1995; De Graaf
et al. 2001; Evans and Whitefield 2006; Brooks et al. 2006; Elff 2007, 2009; and various
contributions to Evans 1999 and Evans and De Graaf 2013).
8. SEGs are an occupational classification that measures employment status rather
than skill or social standing. SEGs are derived from a combination of occupational
groups, employment status (self-employed or employee) and size of establishment.
9. Where there have been criticisms of the Goldthorpe schema they have often been in
relation to divisions within the middle class and the rise of a potentially more left-
wing, liberal group of socio-cultural specialists (Butler and Savage 1995; Oesch 2006;
Guveli et al. 2007a). These criticisms can be overstated with respect to the validity of
the schema itself (Evans and Mills 2000), and there has been a debate on whether
these distinctions identify class divisions or sectoral ones (Goldthorpe 1995), but
they are useful for understanding the occupational bases of political preferences.
10. Others have proposed various versions of the old and new middle classes. The termin-
ology and exact occupational specification vary—see Parkin (1968), Kriesi (1989), and
Heath et al. (1991) among other early discussions. Recent work on this subject (Oesch
2006; Guveli et al. 2007b) has focused mainly on distinctions between ‘technocrats’
(typically managers) and ‘socio-cultural specialists’ (typically professionals).
11. Indeed recent British research finds that growth in own account work has typically
been among people who were previously unemployed. This has been described as
‘bad self-employment’, which functions as a more or less short-term alternative to
employment when jobs are scarce (Baumberg and Meager 2015).
12. The census data from 1971, 1981, 1991, 2001, and 2011 is available directly, and
for 1971, 1981, and 1991 the SEG can be used to code our class categories. For 2001
and 2011 we use the NS-SEC measure of occupation to code into our class categories.
For 1931, 1951, and 1961 the figures come from Price and Bain (1988). We have
applied a small correction of 5 per cent, calibrated against their estimates for 1971
and 1981, to the Price and Bain proportion of ‘manual workers’ for those years as
they contain people who we would have classified as personal service workers.
13. For the LFS data, we include those with ‘trade apprenticeships’ as people with no
qualifications as they are not categorized separately for some early surveys. We also
combine ‘A Level equivalent’ and ‘some higher education’ for the medium category
for both datasets as it maintains more consistency over time for both the LFS and
the combination of the BES and BSA data. The proportions of people in educational

17
The New Politics of Class

categories not shown in the figure are, in total, quite static from the early 1980s
onwards at a bit over a third of the population. There are some changes within
these other categories, however. Taking the 1985–2015 end points of the BSA for
which we have the full range of educational qualifications coded, the number of
people with apprenticeships drops from about 7 per cent of people to 2 per cent
and the numbers with CSE equivalent qualifications increases from around 4 per
cent to over 7 per cent. The main group not shown in the figure, those with O Level
equivalent qualifications, forms a stable quarter of the population.
14. There are many others who have taken related positions, from early writings on the
end of ideology and the end of class politics (Bell 1960; Nisbet 1959) through
Lipset’s reformulation of the politics of class in his revised edition of Political Man
(1981), Inglehart and Rabier’s move from class-based to post-materialist politics
(1986), Eder’s ‘new politics of class’ (1993), and Kitschelt’s analysis of the rise of
centralist social democracy (1994), to name but a few.
15. It is generally accepted that the media help to shape the public’s opinions by agenda
setting (focusing on some issues at the expense of others) rather than by directly
changing people’s views. Early studies of media effects (Lazarsfeld et al. 1948; Berelson
et al. 1954; Klapper 1960) emphasized its role in partisan reinforcement. Most research
since has also identified an indirect campaign influence reinforcing prior opinions
(Gelman and King 1993; Andersen et al. 2005), or modest effects on areas of political
behaviour such as turnout (Iyengar and Simon 2000; Goldstein and Ridout 2004).
16. Evidence on these propositions using British data was originally presented in Evans
et al. (1999) and Evans and Tilley (2012a, 2012b, 2013). Oskarson (2005) and Elff
(2009) have expanded this approach cross-nationally, while Janssen et al. (2013)
demonstrate the relationship across a large number of countries.
17. Some have argued that voters’ preferences are shaped by the way parties frame
choices and talk about politics. If parties adopt certain positions, voters will tend to
follow suit. Sartori (1969, p.84) even argued that the politicization of class divisions
by parties not only influences voters’ party choices but actually produces class
consciousness and an awareness of class-related economic interests. This ‘prefer-
ence shaping’ approach to understanding class politics implies that parties directly
influence the attitudes of their supporters. In this case we would expect class
differences in ideology and values to reflect the shifting positions taken by the
parties associated with different social classes. This implies that working class voters
should have followed ‘their’ Labour party to the centre and become more like the
middle classes. As with other recent studies (Baldassarri and Gelman 2008; Adams
et al. 2012), we find limited evidence for this and, as we show in Chapter 4, the classes
remain distinct in their policy preferences even when the parties converge.
18. Unsurprisingly, American political scientists developed the influential ‘resource model’
of turnout, in which participation is facilitated by the reduced costs of gaining infor-
mation about politics and of voting itself, for middle class, higher-income, and more
highly educated people (Verba and Nie 1972; Verba et al. 1978; Verba et al. 1995).
19. The BSA was not fielded in 1988 and 1992. For some of the analyses presented in
later chapters we are unable to use the BES from 2005 and 2010 as it did not ask
occupational class in a way that can be fully compared with previous surveys.

18
Part I
Social Continuity
2

Inequality

As the twentieth century ended many believed that social classes were a thing
of the past. Social divisions based on economic inequality were disappearing
as ‘post-material’ values concerning quality of life issues grew. Yet, at the same
time as some talked of the death of class, others were revealing that class
divisions not only remained but that social inequalities were becoming more
pronounced. The United States was in the midst of the greatest expansion of
income inequality for a century and although many advanced post-industrial
societies witnessed economic growth this was accompanied by pronounced
increases in inequality (Bartels 2008; Pierson 2001; Beckfield 2003; Alderson
et al. 2005; Fischer and Hout 2006; Neckerman and Torche 2007; Moller et al.
2009; Reardon and Bischoff 2011; Piketty 2014). As a result of globalization
and regional integration, the real wages of low-skilled workers stagnated or
even declined (Brady 2005; Beckfield 2006).
In Britain there has also been a resurgence of interest in class inequality.
Governments and NGOs have expressed repeated concern about their con-
tinued presence. A report from the Runnymede Trust in 2009 suggested that
working class people are increasingly separated from the majority by inequal-
ities of income, housing, and education (Sveinsson 2009). The Sutton Trust
(2009) exposed the role of educational institutions in inhibiting upward
mobility and reproducing class inequality. The Labour government formed
the National Equality Panel, followed a few years later by the Conservative/
Liberal government’s appointment of Alan Milburn to chair the Social Mobil-
ity and Child Poverty Commission (SMCPC). On taking up the post, the
former Labour minister expressed the view that ‘sadly, we still live in a country
where, invariably, if you’re born poor, you die poor. Just as if you go to a low-
achieving school, you tend to end up in a low-achieving job’.1 The SMCPC,
still on-going, has an explicit brief to monitor progress in improving social
mobility and reducing child poverty in the hope that Mr Milburn’s observa-
tions will eventually be shown to be wrong. As we shall see, there is little sign
of this hope being realized in the foreseeable future.
The New Politics of Class

In this chapter we focus on this contemporary renewal of research into class

inequality in Britain. We include studies of trends in inequalities in income,
unemployment, health, and educational achievement, as well as levels of social
mobility between classes. We show both the continuation, and in some cases
exacerbation, of class inequalities. Evidence of ‘the death of class’ is in short supply.
We also look at how ordinary people make sense of these inequalities. Are people
aware of the extent of inequality and how it has developed over time? How do
people explain unequal outcomes? Such subjective perceptions complement
objective inequalities. First, we consider research on trends in objective inequal-
ities between occupational classes: economic, health-related, and educational.

The Reality of Inequality

Class and Pay
It is common knowledge that wage inequality has been increasing in Britain in
recent decades. General interest has focused on changes at the top end of the
salary range. For example, the National Equality Panel report concluded that
‘between 1999 and 2007 the real earnings of the CEOs of the top 100 com-
panies more than doubled (reaching £2.4 million per year)’ (National Equality
Panel 2010, p.42). However, showing that earnings have spiralled at the top is
not in itself evidence that wider class inequalities have become exacerbated.
For this we need to look more broadly at class divisions in earnings.
Various pieces of evidence are available to examine this question. Gallie
(2015) has recently conducted a detailed examination of changing class
inequalities in pay between 1986 and 2012 using the Skills and Employment
Surveys. These constitute six national samples of between approximately 2,000
and 5,000 people. To measure class position he uses the National Statistics
Socio-Economic Classification (NS-SEC) which was developed from, and is
very similar to, the Goldthorpe class schema (Rose and Pevalin 2003). Using
gross hourly pay adjusted to 2012 prices, Gallie finds that pay inequalities
polarized across all classes when compared to higher managerial and profes-
sional workers. The working class, classified as semi- and unskilled routine
workers using NS-SEC terminology, were clearly the most disadvantaged at
the start of the period and became increasingly more so over time.
A more nuanced picture of changing patterns of earnings inequality can
also be obtained from the New Earnings Survey/Annual Survey of Hours and
Earnings (NES) which we have for the period from 1975–2008. This covers all
of the period of marked growth in pay inequality identified by Gallie, and also
allows us to examine trends from the 1970s to the 1980s. Most significantly,
the NES is a compulsory survey covering more than 1 per cent of the labour
force. This vast array of data provides arguably the most robust available

22
 Inequality

80%
NMC
+ OMC

60%

40% JMC

20% WC

0%
1970 1980 1990 2000 2010

Figure 2.1. Mean percentile income by occupational class

Note: The figure here shows mean percentile income by occupational social class for full-time and part-time
workers, aged 18–65 whose earnings were not affected by absence (three period moving average). Three occupa-
tional class groups are displayed: middle class (MC), junior middle class (JMC), and working class (WC).
Source: New Earnings Survey.

evidence on earnings and class in modern Britain. What does it tell us?
Figure 2.1 shows mean percentile wage positions of the different classes over
time. The mean percentile wage is the proportion of workers in each social
class who are paid more than the average hourly wage. This is intrinsically a
relative measure as it is estimated in relation to the average wage in a given
year. For example, approximately around a third of the working class were
paid more than the average hourly wage in 1975, but this dropped to around a
quarter by 2008. We are forced to combine the old and new middle classes
here, and show the comparison between them, the junior middle class (the
intermediate class in NS-SEC) and the working class (semi- and unskilled
routine workers in NS-SEC). In essence, occupational class differences in
income are fairly stable across more than three decades, although the gap
between the working class and the junior middle class moderately increased
between the 1970s and the 1990s.2
We can also see the extent to which these class differences in income are
evident in the BES and BSA datasets that form the core of our analysis of
political attitudes and behaviour in later chapters. These surveys have meas-
ures of household income. This means that they do not allow hourly rates to
be calculated and we are not necessarily linking people’s own jobs to their own
incomes. The sample sizes are also far smaller, and the measurement categor-
ies of income are far cruder. Nonetheless, they show a similar pattern of
stability to the NES data. Figure 2.2 shows the proportion of each class that
falls within the top 40 per cent of the income distribution since the 1960s.
The proportion of the old and new middle occupational classes in the top

23
The New Politics of Class

80%
OMC
NMC

60%

JMC
40%

20%
WC

0%
1960 1970 1980 1990 2000 2010

Figure 2.2. Proportion in top 40 per cent of the income distribution by occupational class
Note: The figure here shows the proportion of people in the top two quintiles of the income distribution by
occupational social class (three period moving average). Four occupational class groups are displayed: old middle
class (OMC), new middle class (NMC), junior middle class (JMC), and working class (WC).
Source: British Election Studies 1963–1979; British Social Attitudes Surveys 1984–2015.

40 per cent of the income distribution remains stable and very high, about 70 per
cent across fifty years. The proportion of the working class with high incomes
remains much lower than those in the new and old middle classes. Someone in
the new or old middle class is about three times as likely to be in the top 40 per cent
of the income distribution as someone in the working class. The junior middle
class falls somewhere in between, and has actually edged closer to the working
class in recent years, possibly since a higher proportion of junior middle class jobs
are now part-time and, unlike in Figure 2.1, we do not measure hourly rates.

Labour Market Insecurity

Income is only one indicator of advantage and disadvantage. One of the key
distinctions between occupational classes is in their provision of other desirable
features of jobs, such as employment security, sick pay, pensions, health insur-
ance, and other benefits. For Beck (1992), the notion of the ‘risk society’ involves
the spreading of risks across the class structure. The insurance approach (Iversen
and Soskice 2001; Moene and Wallerstein 2001) likewise focuses on cleavages
based on risk exposure that are thought to be uncorrelated with traditional class
divisions. While not necessarily claiming that class inequalities have disappeared,
these approaches suggest that risk, especially of unemployment, is now more
evenly spread across the different classes. However, as we will see, evidence on
class differences in unemployment rates actually shows little change over time.
Figure 2.3 shows unemployment rates by class position using the LFS and
BES/BSA.3 We can see that the old and new middle classes have extremely low

24
 Inequality

a) LFS data
20%

WC
10%
JMC
NMC
OMC
0%
1975 1985 1995 2005 2015

b) BSA and BES data

20%
WC

JMC
10%

OMC
NMC
0%
1975 1985 1995 2005 2015

Figure 2.3. Unemployment rate by occupational class

Note: The figures here show the proportion of people who are unemployed as a percentage of those in the labour
force (including those who are in education or training) by occupational class. The top graph shows unemploy-
ment rates from LFS data; the bottom graph shows unemployment rates from BSA and BES data (three period
moving average). Four occupational class groups are displayed: old middle class (OMC), new middle class (NMC),
junior middle class (JMC), and working class (WC).
Source: Labour Force Survey 1992–2014; British Social Attitudes Surveys 1984–2015; British Election Studies 1974–1983.

unemployment rates, always fewer than 5 per cent in the LFS and around that
level in the BES/BSA surveys. The junior middle class is similar. In contrast, the
gap between the middle classes and working class is always substantial. In fact,
unemployment rates for the working class are on average three times higher
than those of the new middle class. Beyond this, it is clear that unemployment
rates among the working class move up and down with the economic cycle.
Both sets of data indicate that changes in overall unemployment rates dispro-
portionately affect the working class.

Health
Continuing class differentials in health are evident across a variety of meas-
ures, including self-rated general health (Drever et al. 2004), physiological
gauges such as blood pressure (Atherton and Power 2007) and healthy diet

25
The New Politics of Class

and exercise (Roberts et al. 2013). The same holds true for mortality. White
et al. (2003) analysed class differentials in mortality during the 1990s using
data from the ONS Longitudinal Study: a 1 per cent sample of the UK popu-
lation for which data from censuses has been linked to death registrations. They
observed that mortality fell in all classes; but relative class differentials actually
increased for men.4 Similarly Langford and Johnson (2010) examined trends
in class differentials in male mortality during the 2000s finding, again, that
while mortality fell in all classes over time, relative class differentials increased
slightly. Johnson and Al-Hamad (2011) report a similar pattern for women.
Psychological measures such as depressive symptoms (Atherton and Power
2007) also show pronounced class differentials. A recent analysis of children’s
dysfunctional behavioural conduct, negative emotional symptoms, and
hyper-activity comparing three nationally-representative UK birth cohort
studies (1958, 1970, and 2001/2) finds growing class inequalities over several
decades (Anderson 2016). The general pattern across all three outcomes is of
striking rises in class inequality since 1969, which was the starting point for
comparison (at this point the 1958 cohort were ten to eleven years of age).
Given the links between such symptoms and educational success (McLeod
and Kaiser 2004; Kantomaa et al. 2010), these also carry implications for
continuing class inequalities in educational attainment.

Educational Attainment
The contemporary labour market places greater emphasis on educational
credentials than was typically the case in the mid-twentieth century. The dra-
matic expansion of higher education since the 1980s has brought with it an
increasing tendency for employers to emphasize degree-level qualifications as
entry hurdles for many posts. Previous generations may not have needed to
succeed in education to gain promotion, but we live in a more credentialist society
than we did in the 1960s and 1970s. Differential educational success is likely
to play a larger role in consolidating class inequalities than used to be the case.
It is well established that there are sizable and persistent class differences in
educational attainment (Erikson et al. 2005; Crawford 2014; Stuart et al.
2014). This inequality starts early. The government’s own report (Stuart et al.
2014) into the outcomes of white British children who are eligible for free
school meals (FSM) finds that they are consistently the lowest-performing
group. Even at age five, only 32 per cent of children on FSM achieve the
expected benchmark for their age, compared with 56 per cent of children
who are not on FSM. By sixteen the gap is even wider: 32 per cent on FSM
achieve five C-grade GCSEs or above, compared to 65 per cent not on FSM
(IPPR 2014). Mills’ (2015b) analysis of data from the Labour Force Survey
found that the class gradient in GCSE performance decreased slightly between

26
 Inequality

the mid-2000s and 2012, but just as with the FSM analysis, it is white British
children from working class backgrounds that are the worst performers.5
As an aside, most academic studies focus on the state education sector, but a
remarkably effective supplier of recruits to the best positions in our society is
not part of that system. English public (i.e. fee-paying) schools have excellent
international reputations for their ability to carve out success for their pupils.
For good reason: although just 7 per cent of people attend independent
schools, they make up 70 per cent of High Court judges and 54 per cent of
CEOs of FTSE 100 companies (Sutton Trust 2009). Moreover, as we have shown
elsewhere, the extent of the social divisions associated with private schooling
runs far beyond differences in access to rarefied elite jobs (Evans and Tilley
2011). Even taking into account many other differences, private schooling is a
key predictor of subsequent occupational and educational attainment.6
More importantly for our purposes, a further area of continued class inequal-
ity is in advancing beyond the minimum school leaving age and going on to
further and higher education. Children whose parents have working class jobs
are less likely to progress to A Levels, even conditional on their performance at
the GCSE stage (Erikson et al. 2005). This inequality continues to university
admissions. Bolton (2010) reports that university intake is heavily influenced
by parental occupational class. Changes since the 1970s have been very mod-
est, although from 2002–2008 there was a slight increase in the relative repre-
sentation of children from working class backgrounds.7
Overall, class divisions in educational outcomes show little sign of weaken-
ing. Unsurprisingly, this differential educational attainment is consequential
for inequalities in outcomes such as income and unemployment. People who
complete more years of education go on to earn substantially more during the
course of their working lives, and there are also pay premiums from postgradu-
ate degrees over and above that conferred by an undergraduate degree (Leary
and Sloane 2005; PWC 2005; PWC 2007; Kirby and Riley 2008; Conlon and
Patrignani 2011).8 People who complete an undergraduate degree earn about
double the wage of those who do not attain any formal qualifications (Conlon
and Patrignani 2011). Analysis of unemployment rates from 1992–2014 by
educational level finds similarly strong effects (Busch 2015). Busch shows that
throughout this period the unemployment rate for those with degrees or other
forms of higher education was no more than 5 per cent. People educated to
GCSE level were significantly more likely to experience unemployment,
though they were far less likely to do so than those with no educational
qualifications. Moreover, the differences in risks of unemployment between
people with different educational levels widened after the economic crisis of
2008. From 2009 onwards, people with no qualifications had unemployment
rates of around 17 per cent while unemployment rates among those with
degrees remained only 3 per cent.

27
The New Politics of Class

a) Proportion in bottom 40 per cent of the income distribution

80%

60% Low

40%

Medium
20%

High
0%
1980 1990 2000 2010

b) Proportion unemployed

20%
Low

10%
Medium

High

0%
1980 1990 2000 2010

Figure 2.4. Income and unemployment by education

Note: The top graph here shows the proportion of people in the bottom two quintiles of the household income
distribution by educational level (two period moving average). The bottom graph shows the proportion of people who
are unemployed as a percentage of those in the labour force (including those who are incapacitated and in education or
training) by educational level (three period moving average). Three educational groups are displayed: people with degree
level education (high), people with A Level equivalent education (medium), and people with no qualifications (low).
Source: British Social Attitudes Surveys 1985–2015.

As we might therefore expect, in the BSA data differing levels of education

are associated with substantial income and unemployment disparities. Figure 2.4
shows the proportion of people in the bottom 40 per cent of household incomes
and those who are unemployed by the three levels of education identified in
Chapter 1. As with social class, our survey data indicate a pattern of fairly constant
and predictable educational inequalities in income. The evidence on unemploy-
ment also mirrors the patterns for occupational class: those with the lowest level
of educational qualifications are more vulnerable to unemployment during diffi-
cult economic times, and the differences between educational groups are marked.

Social Mobility
Education is the single most important influence on occupational attainment
(Breen and Jonsson 2005). Educational reform has therefore been seen as a major
lever for governments attempting to increase social mobility. However,

28
 Inequality

educational changes, of all types, appear to have proved ineffective in amelior-

ating class differences in mobility chances (Bukodi and Goldthorpe 2016;
Goldthorpe 2016).9 Given that educational disadvantages may be more import-
ant in the contemporary labour market than they were before higher education
expansion, we might even expect social mobility to have fallen. This has cer-
tainly been prominent in the arguments of influential economists: Blanden
and Machin (2007) and Ermisch and Nicoletti (2007) have claimed that inter-
generational income mobility has declined in recent decades. Interestingly,
however, this research focuses specifically on intergenerational income mobility,
not occupational class mobility. And the case for decline is not straightfor-
ward. Goldthorpe and Mills (2008) argue that intergenerational occupational
mobility (defined as relative chances of upward and downward mobility) has
in fact been fairly stable over the last thirty years.10 Goldthorpe and Mills
contrast this with the middle decades of the last century when upward mobil-
ity actually increased.
Upward mobility chances for both men and women have also been esti-
mated over the long term using the 1946, 1958, and 1970 Birth Cohort
studies (Bukodi et al. 2015). For men the picture is clear: there are constant
levels of upward mobility with no significant change across the three
cohorts. For women, there is evidence of an increase in upward mobility
in the most recent cohort.11 However, this is only found among women
who work part-time. Among full-time workers the picture is the same as it is
for men.
Putting this together, there is little evidence that parental class has a weaker
impact on people’s own occupational trajectory. Indeed, the debate is largely
between those who think social mobility has decreased versus those who
think it is largely stable. This is broadly true of class inequality across the
many areas that we discuss. Occupational class and education shape the
reality of people’s lives, in terms of incomes, instability, health, and social
mobility much as they did fifty years ago.

The Interpretation of Inequality

In some respects class inequality appears to have got worse and in others it has
remained stable. However, the social and political responses to such objective
inequalities are likely to be conditioned by how people interpret them. In this
section we examine evidence on beliefs about inequality and how unequal
outcomes are achieved.
We should first note that objective evidence of class inequalities in pay, job
security, and health are clearly apparent to the people who experience them:
working class people are more fearful of unemployment, enduring poverty,

29
The New Politics of Class

a) Occupational classes
Become Not enough Inadequate
unemployed money healthcare

40% WC

30% JMC

NMC
20% OMC

10%

0%

b) Education
Become Not enough Inadequate
unemployed money health care
40%

30% High

Medium
20%
Low

10%

0%

Figure 2.5. The experience of various risks by occupational class and education
Note: The figures here show the proportion of people who think that it is likely that they will suffer the following over the
next twelve months: unemployment, not having enough money for household necessities, and not receiving necessary
health care in the event of becoming ill. The top graph shows this by occupational social class; the bottom graph shows
this by educational level. Four occupational class groups are displayed: old middle class (OMC), new middle class (NMC),
junior middle class (JMC), and working class (WC). Three educational groups are displayed: people with degree-level
education (high), people with A Level equivalent education (medium), and people with no qualifications (low).
Source: European Social Survey 2008.

and facing health problems than middle class people. This can be seen from a
recent BSA survey that asked people whether they thought it was likely that
they would be unemployed, not receive adequate health care if they became
ill, or not have enough money to cover household necessities. Figure 2.5
presents responses by occupational class and education.
Almost 40 per cent of the working class think that it is likely they will
become unemployed in the next twelve months, compared with less than
20 per cent of the middle classes. Likewise over 45 per cent of the former are

30
 Inequality

concerned about not having enough money to cover necessities compared

with less than 25 per cent of the latter.12 Concerns about adequate health care
are connected with class position in similar, though less pronounced, ways.
That they are still clearly linked despite the presence of the National Health
Service could well relate to the ability of middle class people to afford private
health care, or perhaps to live in areas where the postcode health lottery is
more favourable. Educational differences in such perceived risks again mirror
the patterns of differences between classes. The more education that someone
has, the lower their level of perceived risk.
These perceptions tell us something about the fears associated with vulner-
able labour market positions: the psychological costs of the risk of unemploy-
ment and the risk of not having enough money to make ends meet,
accompanied by concerns about inadequate health care should the worst
happen. But what of people’s beliefs about how inequality works in society
more generally? First, we examine whether people are aware of how unequal
society has become with respect to pay levels. We then look at how people
explain why some end up with the best jobs.

Awareness of Pay Inequality between Classes

Earnings inequality has been growing, but are people aware of this? The BSA
surveys in 1987, 1999, and 2009 asked respondents to estimate the amounts
earned by people in a range of jobs, including middle class (or above) positions
such as GPs, company chairmen, or cabinet ministers, as well as working class
jobs such as unskilled factory workers or shop assistants. People are asked how
much people in each type of job ‘usually earn each year before taxes’. Table 2.1
presents the answers to these questions.
We can see that people underestimate the actual salaries of high earners
though less so in some cases than others. Cabinet ministers’ salaries in 2009
were £144,500, and GPs in England typically earned around £110,000.13
People underestimated these amounts, but nowhere near as much as they

Table 2.1. Perceived mean earnings of different jobs

1987 1999 2009

Chair of large national company £90,903 £179,870 £224,283

Cabinet minister £39,455 £71,641 £109,352
GP £21,187 £38,105 £75,957
Unskilled factory worker £6,072 £10,859 £15,313
Shop assistant – £9,265 £13,297

Note: These figures are the mean amount each type of occupation is believed to ‘usually earn each year before taxes’.
Source : British Social Attitude Surveys 1987, 1999, and 2009.

31
The New Politics of Class

underestimated the salaries of heads of large national companies. In 2008 the

average remuneration package of the CEOs of the FTSE top 100 companies was
£2.4 million a year, while that of the next largest 250 companies was £1.1
million (Heath et al. 2010). By comparison, at the lower end of the wage scale
perceptions are more accurate. The 2009 LFS indicates that the average earn-
ings of female shop assistants were roughly £13,000 per year and unskilled
male factory workers earned around £15–16,000 per year.
The perceived gaps between the earnings of occupations have changed over
time, but only modestly. Whereas cabinet ministers were seen as earning 6.5
times a factory worker’s wages in 1987, in 2009 it was just over 7. Whereas GPs
were seen as earning 3.5 times a factory worker’s wages in 1987, in 2009 it was
5 times. The earnings ratio for CEOs and factory workers was effectively
unchanged over twenty-two years.14 Awareness of the growing inequality
discussed earlier in this chapter is surprisingly muted.
Are these perceptions shared across the class structure? Table 2.2 presents
estimates of wages by class position. The earnings estimates for working class
people given as a percentage of the three middle classes are shown at the
bottom of the table. From Table 2.2 it appears that people from different
classes have rather similar perceptions of the wages for different jobs, and
there is little evidence of change over time.
The closeness of the different classes’ estimates of earnings for the working
class jobs may well reflect the limited variance in earnings among such jobs, as
well as a ‘floor effect’. However, working class respondents are generally more
likely to underestimate how much those in middle class professions earn. This
downward bias may serve to mitigate resentment about highly paid middle
class jobs by working class people. This possibility can be examined by

Table 2.2. Perceived mean earnings of different jobs by occupational class

1987 2009

CEO Cab Min GP US Man CEO Cab Min GP US Man Shop Asst

Earnings in £1,000s
Old middle class 101 39 21 6 257 103 81 14 14
New middle class 90 35 22 6 252 110 80 18 14
Junior middle class 91 39 22 6 207 103 77 14 13
Working class 87 42 21 6 188 114 71 14 13

Comparisons to Working Class

WC/OMC 86 108 100 100 73 111 88 100 93
WC/NMC 97 120 95 100 75 104 89 78 93
WC/JMC 95 107 97 96 91 111 93 102 101

Note: Entries in the first four rows are the mean amounts each occupation is believed to ‘usually earn each year before
taxes’, rounded to the nearest thousand. Entries in the last three rows are group comparisons given as percentages.
Source: British Social Attitude Surveys 1987 and 2009.

32
 Inequality

Table 2.3. Gap between perceived and deserved earnings by occupational class

1987 2009

CEO Cab Min GP US Man CEO Cab Min GP US Man Shop Asst

Differences in £1,000s
Old middle class 38 9 –1 –1 82 24 7 –3 –3
New middle class 36 7 – –1 103 33 3 –1 –4
Junior middle class 35 11 –1 –1 65 36 10 –4 –4
Working class 45 14 – –2 73 51 1 –4 –4
Differences as ratio
Old middle class 0.4 0.2 – –0.2 0.3 0.2 0.1 –0.2 –0.2
New middle class 0.4 0.4 – –0.2 0.4 0.3 – – –0.3
Junior middle class 0.4 0.3 – –0.2 0.3 0.4 0.1 –0.3 –0.3
Working class 0.5 0.4 – –0.3 0.4 0.4 – –0.3 –0.3

Note: Entries in the first four rows are differences between the average amounts it is believed each occupation earns each
year versus the average amount it is believed that occupation should earn, rounded to the nearest thousand. Entries in
the last four rows are fractions of over/underpayment. Negative values indicate underpayment.
Source: British Social Attitude Surveys 1987 and 2009.

comparing beliefs about what jobs are perceived to be paid, with beliefs about
what they should be paid. This gives us a measure of the perceived injustice of
occupational differences in pay (Kelley and Evans 1993).
Table 2.3 indicates that the gap between perceived and deserved pay is sur-
prisingly constant across different classes. Levels of deserved pay are shown in
the top half of the table and the disparity between that and perceived pay is
shown as a proportion in the bottom half. All classes think that CEOs are
substantially overpaid; all think that the two working class jobs are underpaid.
All see GPs’ salary as about right. The only class differences are in the tendency
for the working class to believe that cabinet ministers are overpaid to a greater
extent than do the middle classes, though this is only a matter of degree.
Generally, the classes are in remarkable agreement on both the levels, but
particularly the fairness, of pay between different types of occupations.

Explaining Success
Maybe more important than perceptions of the extent of class differences in
incomes is the way in which they are explained. Explanations for success and
failure are argued to be shaped by an individualistic ideology pervasive in
Western societies (Ichheiser 1949; Huber and Form 1973; Kluegel and Smith
1986). This holds that people believe society to be relatively open and that
equality of opportunity, if not of outcome, is widespread. In short, people are
responsible for their own fate. Accordingly, success is attributed to personal
characteristics regardless of social origins. Such attributions are correlated with a

33
The New Politics of Class

set of beliefs emphasizing personal responsibility and the perception that

people ‘get what they deserve’, including ideas of a just world (Cozzarelli et al.
2001), the Protestant work ethic (Furnham 1988), social dominance (Lemieux
and Pratto 2003), and political conservatism (Zucker and Weiner 1993). This set
of beliefs, which Huber and Form (1973) refer to as ‘the dominant ideology’,
and which Ichheiser (1949) talks of as ‘the success ideology’ has been thought
to provide a legitimation of widespread social inequalities in Western societies
(Huber and Form 1973; Kluegel and Smith 1986).15 Its generic presence has led
to it being labelled ‘the fundamental attribution error’ (Ross 1977).
The BSA also has a set of questions that look at how people explain who
‘gets ahead’ in life. These ask how important different factors are in ‘getting
ahead’, namely: coming from a wealthy family; having well-educated parents;
knowing the right people; having a good education yourself; having ambition
and hard work. ‘Ambition’ and ‘hard work’ attribute success to the individual
rather than social circumstance, and could be thought to be endorsing merit-
ocracy,16 whereas ‘coming from a wealthy family’, having ‘well-educated
parents’, and ‘knowing the right people’ emphasize the role of social back-
ground and connections. These are typically thought of as structural, or
non-meritocratic, factors. As we have seen, the chances of obtaining a good
education are in fact strongly influenced by class background. This could
perhaps also be seen as non-meritocratic. It does not appear that people
understand it in this way, however, as ‘having a good education’ correlates
far more closely with individualistic attributions than structural ones.
It is clear from Table 2.4 that people believe that ‘hard work’, ‘ambition’,
and ‘having a good education’ are the most important explanations of success.
On average, three quarters of people or more think these are very important or
essential, a level which has stayed fairly constant over time despite the notice-
ably worse economic circumstances in which the later survey was undertaken.

Table 2.4. Beliefs about what is important for ‘getting ahead’

% saying essential or very important 1987 1999 2009

Individualistic or meritocratic
Hard work 84% 84% 84%
Ambition 79% 74% 71%
A good education 72% 74% 74%
Structural or non-meritocratic
Wealthy family 21% 15% 14%
Knowing the right people 39% 35% 33%
Well-educated parents 27% 28% 31%

Note: These are the proportions of people who said these characteristics are either ‘very important’ or ‘essential’ ‘for
getting ahead’.
Source : British Social Attitude Surveys 1987, 1999, and 2009.

34
 Inequality

Structural factors are seen to be important by only a minority of people. There

has also been a decline since the 1980s in the proportion believing that these
are very important or essential, although this is accompanied by a slight
increase in the proportion believing that ‘having well-educated parents’ is
an important factor. This reflects, perhaps, the extensive growth of higher
education during this period and its consequences for access to good jobs.
Generally, however, people believe that individuals themselves, rather than
their circumstances, are responsible for their success. But are these views

a) Structural explanations

100% 100% 100%

Wealthy family Educated parents Knowing the
right people
75% 75% 75%

50% 50% 50%

25% 25% 25%

0% 0% 0%
NMC

NMC

NMC

NMC

NMC

NMC
OMC

WC
OMC

WC

OMC

WC
OMC

WC

OMC

WC
OMC

WC
1987 2009 1987 2009 1987 2009

b) Individualistic explanations

100% 100% 100%

Hard work Ambition Education

75% 75% 75%

50% 50% 50%

25% 25% 25%

0% 0% 0%
OMC

WC
OMC

WC

OMC

WC
OMC

WC

OMC

WC
OMC

WC
NMC

NMC

NMC

NMC

NMC

NMC

1987 2009 1987 2009 1987 2009

Figure 2.6. Explanations for ‘getting ahead’ by occupational class

Note: The figures here show the proportion of respondents who said each characteristic is either ‘very important’ or
‘essential’ ‘for getting ahead’. The top graphs display structural characteristics; the bottom graphs display individ-
ual characteristics. Three occupational class groups are displayed: old middle class (OMC), new middle class
(NMC), and working class (WC).
Source: British Social Attitudes Surveys 1987 and 2009.

35
The New Politics of Class

shared across the class structure? While we might expect the socially disad-
vantaged to reject individualistic accounts of success, this is not the case.
Figure 2.6 shows there is very little difference at all in the weight given to
the individualistic factors of education and ambition. The working class are
less likely to emphasize ambition and education than the middle classes in
2009, but the differences are outweighed by the similarities: more than 60 per
cent of all classes thought that these attributes were very important or essen-
tial for getting ahead. Interestingly, in 1987 respondents were also asked about
the importance of ‘natural ability’ for explaining who gets ahead. The working
class were actually more likely to emphasize the importance of natural ability
than the middle classes. Among the structural explanations for success, the
most noticeable class difference, in the higher importance given to ‘knowing
the right people’ by working class people in 1987, had weakened by 2009. The
increase in emphasis on educated parents in 2009 was most marked in the
new middle class, yet among all classes structural factors were far less likely to
be seen as important or essential for getting ahead than individual attributes.
Explanations of who gets ahead seem to a fair degree to be shared across
society, and do not appear to have changed very much. Heath et al. (2010)
conclude that even personal mobility experiences have only a ‘very modest’
relationship with people’s explanations of success. Using an experimental design
and indirect measures of attributions, Evans (1997) found a similarly closely
shared view of the importance of structural versus individualistic influences on
social mobility between people on the left and right of the political spectrum.17

Conclusions

In the introduction to the 1989 edition of The Road to Wigan Pier, Richard
Hoggart notes Orwell’s disdain for those who thought class divisions in 1937
were dead, and points out that the same assertions continued to be incorrectly
made. As he rightly said, ‘each decade we shiftily declare we have buried class:
each decade the coffin stays empty’ (Hoggart 1989). His words are no less
resonant another thirty years later. The fact is that modern Britain has strong
and pervasive class inequalities. Pay differences between the classes have not
diminished, and may have increased. All our data shows that patterns of
unemployment are strongly related to class and education, and those groups
that are more likely to be unemployed are also those hit harder by fluctuations
in the economy.
Major studies of health inequalities point to widening not declining class
differences in relative mortality and in psychological indicators of dysfunctional
behavioural conduct, negative emotional symptoms, and hyper-activity. Class
inequalities in educational attainment start at age five and as children go

36
 Inequality

through the education system the gap gets wider. Even when controlling for
differences in GCSE performance, working class children fail to progress relative
to their middle class peers and under-representation at higher educational levels
shows no sign of abating. Education itself retains its significant and substantial
impact on earnings, risk of unemployment, and other aspects of life-chances
including, of course, the attainment of class positions themselves. Social mobil-
ity opportunities have remained at best constant, and some have argued they
are in decline. In short, it is better to be middle class than working class, and it is
better to be highly educated than poorly educated.
People are also aware of these social inequalities. Working class people
perceive more insecurity in their lives, while people in all social classes think
pay differences between some highly paid jobs and typical working class jobs
are too high. At the same time, these views on inequality are accompanied by a
general perception that the achievement of unequal outcomes is a result of
ambition and effort rather than social advantage. Such beliefs about achieve-
ment underpin to some degree the perception that class position is based on
merit; one implication being that inequality between classes is acceptable as
long as mobility between classes is possible.
These patterns of stability in both inequality and perceptions of inequality
differ from the US experience. Upward social mobility in the US has not
increased, but 40 per cent of Americans believe it has (Scott and Leonhardt
2005). Thus large increases in class inequality in the US have been accompan-
ied by the strengthening of belief in the American Dream: that it is possible to
start out poor, work hard, and become rich. While British people do endorse
aspects of this philosophy, for example most people think hard work is
important for success, there is little evidence of change over time. Inequalities
remain constant, and perceptions of those inequalities remain constant.

Notes

1. Alan Milburn talking to the BBC in April 2011 (http://www.bbc.co.uk/news/

mobile/uk-politics-12962487).
2. There has been more income polarization by class in the USA, where Weeden et al.
(2007) use the US Current Population Survey to show that not only is the amount
of earnings inequality increasing between classes but the share of total inequality
occurring within occupations has declined. As they note: ‘the well-known take off
in inequality has generated a “lumpier” earnings distribution with relatively stron-
ger class and occupational distinctions’ (Weeden et al. 2007, p.702).
3. It is not possible to directly create our class categories using the LFS data, so old
middle class in Figure 2.3 for the LFS data refers to NS-SEC analytic category 1.1
(large employers and higher managerial), new middle class refers to 1.2 (higher
professional), junior middle class refers to 3 (intermediate), and working class refers
to 6 and 7 (semi-routine and routine).

37
The New Politics of Class

4. When assessing changes in class differentials in death rates over time, a distinction
can be drawn between absolute inequality and relative inequality (Langford and
Johnson 2010; Johnson and Al-Hamad 2011). The former refers to the absolute
difference in number of deaths (age standardized) per 100,000 per year between two
classes; the latter refers to the ratio. For example, if class X has a death rate of 600 and
class Y has a death rate of 400, the absolute difference is 200, while the ratio is 1.5.
5. These class differentials in educational achievement occur even when taking into
account differences in ability. On the basis of extensive UK birth-cohort data, Bukodi
et al. (2015) find that class differences in cognitive ability have only modest effects
on the continued impact of parental occupational class on educational attainment.
6. Given that a privately-educated person is also seven times more likely to be married
to another privately-educated person than is someone who is state-educated, and
65 per cent of these privately-educated couples send their children to fee-paying
schools, it is unsurprising that private schools are seen as ‘perpetuating the apart-
heid which has so dogged education and national life in Britain since the Second
World War’ (Seldon 2008).
7. On entering higher education yet another set of class disparities emerge. Crawford
(2014) found that individuals from higher socio-economic groups were less likely
to drop out, more likely to graduate, and more likely to graduate with an upper
second or first-class degree.
8. There does not appear to be much evidence that the pay premium from an
undergraduate degree changed during the period of higher education expansion
that began in the late 1980s (PWC 2007, p.6; Conlon and Patrignani 2011,
pp.37–9). One possible explanation for the apparent stability of the undergraduate
pay premium over the last few decades is that countervailing forces have cancelled
one another out. On the one hand, the automation of mid-skilled service jobs, the
outsourcing of comparatively high-paying manual jobs, and the influx of low-
skilled immigrants may have increased the undergraduate pay premium. On the
other hand, the rising higher education enrolment rate, and the corresponding
decrease in the skill level of the marginal graduate may have dampened it.
9. Goldthorpe (2016) makes the point that educational attainment is a ‘positional
good’: where someone stands in the educational hierarchy relative to others is
likely to be more consequential for their occupational attainment than simply
their level of qualifications. It is not how much education an individual has but
how much he or she has relative to others. In terms of people’s ranking in the
educational hierarchy, class inequalities have remained largely constant over time.
10. We should note that these contrasting findings are not necessarily incompatible if,
as we have observed earlier, there have been changing patterns of earnings within
occupational classes. Erikson and Goldthorpe (2010) consider the discrepancies
and present the case against using income in more detail.
11. Using Bukodi’s data we find that long-range, upward mobility from the working
class (classes 6 and 7 in the Goldthorpe schema) into the middle class (classes 1 and
2 in the Goldthorpe schema) is stable for men: in each of the cohorts it is,
respectively, 26 per cent, 24 per cent, and 28 per cent. For women the equivalent
figures were: 17 per cent, 17 per cent, and 30 per cent. Heath et al. (2009) use

38
 Inequality

somewhat different data, taken from the General Household Survey for 1987, 1992,
and 2005, and from the British Household Panel Study for 1999, to estimate
measures of long-range and short-range upward mobility. They report modest
increases in upward mobility between 1987 and 2005. However, they used occu-
pational data from respondents aged eighteen and over, which might well lead to
an underestimate of some aspects of intergenerational reproduction as many
people in the professional and managerial classes do not attain destination jobs
until much later in their careers.
12. Whether members of these classes have different expectations with regard to what
might be thought of as necessities cannot be ascertained from these questions, but
we might expect a higher level of expectation among the more advantaged classes,
so these are likely to be rather conservative estimates of differences in the likeli-
hood of equivalent financial hardship.
13. This figure refers to contractor GPs (partners at a practice). Salaried GPs earn
considerably less (£58,000 in 2009 on average).
14. As Heath et al. (2010) note, comparing change over time in actual earnings for
specific jobs is not straightforward. However, we know that cabinet ministers’
salaries increased by 30 per cent between 1999 and 2009. GPs’ salaries increased
substantially following the Labour Government’s re-negotiation of their contracts
in 2004: NHS information centre figures that are available for the period 2002/3 to
2008/9 show an increase of approximately 50 per cent in just six years. Similarly,
the National Equality Panel report (p.42) indicates that the remuneration of CEOs
of large companies rose by much more between the 1990s and 2009 than did that
of the average employee, which was generally static.
15. Such attributions are not limited to explanations of success. In one of the first large-
scale studies of attributions for poverty, Feagin (1975) found that individualistic
attributions were supported more strongly than other explanations, a finding that
is indicative of a tendency to view poverty as a sign of personal and moral failure
(Katz 1989; Shirazi and Biel 2005). Feagin’s work was replicated in Britain by
Furnham (1982, 1988). He found that even during a period of large-scale
unemployment people tended to hold the poor responsible for their own fate.
16. The term meritocracy was coined by Michael Young in The Rise of the Meritocracy
(1958), a dystopian vision of the future referring to the negative consequences that
follow when achievement is based purely on a formula of ‘IQ plus effort’. Typically,
a meritocracy is now seen as a society where rewards derive from indicators of
achievement such as educational success, and motivational factors expressed
through hard work. This is generally contrasted with a society in which factors
such as family background, such as parents’ income and social connections, or
characteristics ascribed at birth, such as race or sex, are more important for success.
17. Although expectations associated with implicit assumptions about class have a
pervasive influence on people’s understanding of who gets where in society, who
becomes unemployed, who votes for which party, and even upon who marries
whom (Evans 1993a), these assumptions do not fundamentally alter the individu-
alistic bias in explicit attributions for success and failure, again testifying to the
power of this ideology.

39
3

Identity

In this chapter we follow up on Chapter 2’s examination of inequalities

between the classes by considering class identities: whether people still see
themselves and others in class terms. Following the influential approach of
Cantril (1943), Centers (1949), Hodge and Treiman (1968), Jackman and
Jackman (1983), we distinguish between ‘class identification’ and ‘class aware-
ness’.1 Class identification refers simply to the tendency for people to place
themselves in social classes. This involves recognizing the existence of classes,
but does not require people to attribute importance to these classes. Class
awareness is usually seen as separate from class identification (Vanneman and
Cannon 1987). If people are class aware, they should have a broader under-
standing of how class position influences people’s lives. An awareness of class
at the very least implies a connection between someone’s objective occupa-
tional class and their subjective class identification.
So do people in Britain still have class identities and are they aware of the
impact of class on their own and others’ lives? Our focus on these questions is
a response to claims that class position no longer influences how people see
themselves or others. In the mid- and late twentieth century researchers
typically found that most people understood class labels and believed that
social classes existed (Butler and Stokes 1969; Bulmer 1975; Jackman and
Jackman, 1983; Vanneman and Cannon, 1987; Marshall et al. 1988; Argyle
1994). People were able to describe their own and others’ social class positions.
They placed far greater weight on characteristics such as occupation, educa-
tion, and income, than on characteristics such as race, gender, marital status,
and age when judging social status (Coleman and Rainwater 1979). Recently,
however, it has been argued that such expressions of class identity and aware-
ness are more historical than contemporary. Many social and political scien-
tists now believe that even the most elementary aspect of class-related beliefs,
class identity, is a ‘relic of a bygone age’ (Eidlin 2014, p.1045).
One theme centres on the idea that, regardless of the persistence of inequal-
ity, the distinctiveness and implications of class identity have been lost.
 Identity

Traditional notions of class simply do not resonate with ordinary people’s

experience of social life. As a result some researchers (Savage 2000; Skeggs
1997; Savage et al. 2001; Irwin 2015) have argued that class position no
longer generates a deep sense of identity and belonging. Apparently we
have seen a ‘paradox of class’ (Bottero 2004), in which the continued role
of class position in shaping life chances is nevertheless accompanied by a
declining level of class identification. This work has been interpreted
through Beck and Beck-Gernsheim’s (2002) theory of individualization
and the risk society, which asserts that individuals have been ‘disembedded’
from traditional communities and the inherited identities associated with
them. People now choose their own identities so that if ‘you are interested
in what is going on in people’s minds and the ways of life they are leading,
you have to get away from the old categories’ (Beck and Beck-Gernsheim
2002, p.207). Notwithstanding their nuances, all of these arguments centre
on the idea that class identity is divorced from class position and has lost its
meaning.
Other researchers have claimed that rather than class identity disappearing,
middle class identification has now become the new norm in modern society.
This change is principally attributed to increasing affluence leading to people
in working class jobs seeing themselves as middle class. The benefits of eco-
nomic prosperity in the post-war era have been spread more equally across
social classes, while the welfare state has also served to reduce the more
extreme hardships associated with class inequality (Andersen and Curtis
2012). Inglehart’s influential interpretation of value change asserts that afflu-
ence has ‘brought a shift in the political agenda throughout advanced indus-
trial society . . . a shift from political cleavages based on social class conflict
towards cleavages based on cultural issues and quality of life concerns’
(Inglehart 1997, p.237). Class matters less because class-related poverty and
inequality are no longer such pressing concerns.
A somewhat different version of the affluence thesis relies on ideas taken
from reference group theory (Shibutani 1955; Merton 1957; Siegel and Siegel
1957) and assumes that perceptions of social structure are conditioned by the
character of the immediate social environment. The literature on ‘working
class images of society’ (Bulmer 1975) took this perspective as its starting
point, as did Runciman’s (1966) influential analysis of working class attitudes
towards inequality in the early 1960s. More recently, however, it has been
used to understand the prevalence of a middle class identity. Pahl et al. (2007)
argued that by the end of the last century the relatively narrow income range
of the population led to most people seeing themselves as in the middle. This
was argued to reflect a ‘reasonably accurate view that the material lifestyle of
the households geographically and socially close to them is simply not that
different’ (Pahl et al. 2007, p.18).

41
The New Politics of Class

Kelley and Evans (1995, Evans and Kelley 2004) generalized this propos-
ition, arguing that people consider their own position in comparison with
those around them and this homogeneity of reference groups shapes percep-
tion of the class structure and where people place themselves in this structure.
Although their thesis is tested across many countries, it suffers from a rather
serious limitation in that it does not measure class identity. It employs an
11-point scale of ‘position in society’. This scale is labelled only ‘top’ and
‘bottom’ and makes no reference to what a self-placement of ‘4’ or ‘7’, for
example, might mean. Nor is it clarified what ‘top’ and ‘bottom’ might refer
to, so it is not clear what their scale is measuring. By comparison, research
employing measures of class identity found that in most countries there are
predictable patterns of working class and middle class identification (Evans
1993b; Andersen and Curtis 2012; Curtis 2016). Even in the US, Hout (2008)
shows that just over two thirds of people think of themselves as belonging to
one of those two classes, with no trend in this proportion between the 1950s
and the 2000s. There was a predictable shift towards a higher proportion of
middle class identifiers in line with changes in class sizes, as might be
expected, but nowhere near enough to suggest that everyone now sees them-
selves as middle class.2
So has class identity in Britain changed? The dis-identification argument
suggests that despite continued inequality class is simply no longer meaning-
ful to people, the affluence argument predicts increasing middle class identi-
fication during periods of growth but a fall during downturns, while the
reference group argument suggests that predominantly middle class percep-
tions are both generated and insulated by comparison processes.3 In the rest of
this chapter we refute each of these arguments. We show that people still see
themselves as belonging to social classes and that these identities still relate to
people’s objective characteristics, both past and present. Moreover, working
class identities are still prevalent. We also show that people see others in
class terms and still believe that class has an important impact on people’s
lives. There is as much social continuity in class identities as there is in class
inequalities.

How People See Themselves

Class identity was one of the central concerns of the early BES surveys (see
Butler and Stokes 1969) and questions about this have appeared in almost all
of the BES surveys that followed, as well as a few of the later BSA surveys. First
we examine the most basic aspect of class identity: how many people were
willing to say yes when asked whether they thought of themselves as belong-
ing to a class. The left-hand graph in Figure 3.1 breaks people down into three

42
 Identity

a) Unprompted class identity b) Prompted class identity

Working
None
60% 60%

40% 40% Middle

Working

20% 20%

Middle None

0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 3.1. Unprompted and prompted class identities

Note: The figures here show the proportion of individuals who identify with particular classes. The left-hand graph
show unprompted class identity and the right-hand graph a combined unprompted and prompted class identity.
Three class identities are displayed: middle class identity (middle), working class identity (working), and no
identity (none).
Source: British Election Study 1964–2015; British Social Attitudes Surveys 2003–2015.

categories: those with no identity, those who said they thought of themselves
as working class, and those who thought of themselves as middle class. There
has been no decline in these unprompted class identifications. In 1964, when
class was ostensibly the defining theme of British society, just under 50 per cent
of people volunteered that they were middle or working class; in 2015 the
figure was almost identical. Both middle class and working class identities are
also very stable over time. There has been a slight increase in the number of
people saying that they were middle class (about 5 per cent between 1964 and
2015), but it is still only 20 per cent of the population. Clearly, we are not all
middle class now.
Those who did not volunteer a class identity were prompted to choose
between being working class or middle class. Specifically, they were asked:
‘Most people say they belong to either the middle class or to the working class.
Do you ever think of yourself as being in one of these classes?’ Almost no one
appears to have found this excessively difficult to answer across the entire
period. The right-hand graph in Figure 3.1 shows class identity after this
prompting is included. Again we find little support for the idea that class
identity has declined, let alone disappeared. In 1964 fewer than 10 per cent
of people chose not to give a class identity, and this figure was, if anything,
lower in 2015. Interestingly, as with the unprompted question, although the

43
The New Politics of Class

proportion of working class jobs has declined over time, the proportion of
people who identify themselves as working class is still clearly a majority.
Being class aware is usually seen as more than simply being able to use labels
for self-identification. There are various ways of examining this issue, but the
most commonly adopted method is to examine the link between objective
and subjective class position. If class identification connects with occupa-
tional class, it is likely to be more firmly grounded in experience. Figure 3.2
presents evidence on changes in the relationship between occupational class
and unprompted class identity across time. We can see that among all the
occupational groups, levels of overall class identity remain fairly stable.
Among those who have an identity, people with working class jobs clearly
express a working class identity, which remains constant from the 1960s to
2015: at no point do more than a trivial proportion of the working class see
themselves as middle class. Equally, people in middle class jobs are clearly
more likely to regard themselves as middle class. Levels of identification in the
old and new middle classes are very similar to each other, although rates of
middle class identification are somewhat lower for the junior middle class
group. It is also interesting to note that there is actually a pattern of gently
increasing working class self-identification among all three of the middle class
occupational groups.
If we examine levels of class identification after combining prompted and
unprompted class identification we see similar patterns. As Figure 3.3 shows, a
working class identity is both constant and dominant among people with
working class jobs. Over more than fifty years around 80 per cent of people
with working class occupations have consistently identified themselves as
working class. The dominance of a middle class identity is less pronounced
among those with middle class jobs, and again we see a pattern of gradually
increasing working class self-identification among the middle classes.
So contrary to the ‘we’re all middle class now’ thesis, middle class identifiers
are the ones who are in short supply. Substantial minorities of people with
middle class jobs see themselves as working class and have always done so.
Although the working class have retained a solidly working class identity, the
middle classes have actually become slightly more likely to express a working
class identity. The reasons for people with middle class jobs seeing themselves
as working class are worth examining a little further. One possibility is that
class identities are shaped in childhood and are as much to do with where
someone comes from as where they end up. This is suggested by various
studies of class identity that have found an effect of occupational class back-
ground in addition to people’s current jobs (Heath et al. 2009; Jackman and
Jackman 1983; Pérez-Ahumada 2014; Curtis 2016), and can be seen in
Figure 3.4 which shows the percentage of people who identified themselves
as working class or middle class by their father’s occupational class.

44
 Identity

a) Old middle class b) New middle class

None
60% 60%
None

40% 40% Middle

Middle

20% 20%

Working Working
0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

c) Junior middle class d) Working class

None
60% 60%
None

40% 40%
Working
Working

20% 20%
Middle Middle

0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 3.2. Unprompted class identity by occupational class

Note: The figures here show unprompted class identity by occupational social class. The top-left graph shows
values for members of the old middle class; the top-right shows values for the new middle class; the bottom-
left shows values for members of the junior middle class; the bottom-right shows values for the working class.
Three class identities are displayed: middle class identity (middle), working class identity (working), and no
identity (none).
Source: British Election Study 1964–2015; British Social Attitudes Surveys 2003–2015.

These effects are very stable. Class background matters for someone’s cur-
rent class identity just as much now as it did in the 1960s. The effect of
someone’s own occupational class has about the same effect on identity as
their father’s occupational class. This means that across all years, people with
middle class jobs whose fathers had working class jobs are quite likely to think

45
The New Politics of Class

a) Old middle class b) New middle class

80% 80%

Middle Middle
60% 60%

40% 40%

Working Working
20% 20%
None None

0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

c) Junior middle class d) Working class

80% 80%
Working
Working
60% 60%

40% 40%
Middle
Middle
20% 20%
None
None
0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 3.3. Prompted class identity by occupational class

Note: The figures here show prompted class identity by occupational social class. The top-left graph shows
values for members of the old middle class; the top-right shows values for the new middle class; the bottom-left
shows values for members of the junior middle class; the bottom-right shows values for the working class.
Three class identities are displayed: middle class identity (middle), working class identity (working), and no
identity (none).
Source: British Election Study 1964–2015; British Social Attitudes Surveys 2003–2015.

of themselves as working class. In 1964 44 per cent of people with middle

class (new and old) occupations whose fathers had a working class job had a
working class identity compared with 22 per cent of similar people
whose father had a middle class job. In 2015 these figures were 58 per cent
and 29 per cent respectively. The effect of class background on class identity is

46
 Identity

80%
WC

60%
JMC

40% OMC

NMC

20%

0%
1960 1970 1980 1990 2000 2010

Figure 3.4. Working class identity by father’s occupational class

Note: The figure here shows the proportion of respondents identifying as working class (prompted and
unprompted combined) by father’s occupational social class. Four occupational class groups are displayed
for fathers: old middle class (OMC), new middle class (NMC), junior middle class (JMC), and working
class (WC).
Source: British Election Study 1964–1997; British Social Attitudes 2005 and 2015.

remarkably constant across fifty years; in both 1964 and 2015 exactly twice as
many middle class people who have a working class father also see themselves
as working class compared with those who have middle class fathers.
Background clearly matters and helps us to understand why so many people
with middle class jobs see themselves as working class. It does not help explain
growing levels of working class identification among people with middle class
jobs however. The proportion of people in middle class (old and new) jobs
who have a working class background has remained constant (35 per cent in
1964; 38 per cent in 2015), as has the proportion of people in working
class jobs from middle class (old and new) backgrounds (14 per cent in 1964;
15 per cent in 2015). A more likely explanation for the increase in working
class identity among some groups is the dramatically changing distribution of
educational qualifications. Education has long been recognized as an import-
ant influence on class identification (Robinson and Kelley 1979). However,
as we saw in Chapter 1, far more people now go on to further and higher
education than was the case in the mid- and late twentieth century. One
implication of this is that the importance for social identity of having
obtained further and higher educational qualifications has declined as educa-
tional expansion reduces the social distinctiveness of higher levels of educa-
tion. This can be seen in Figure 3.5, which shows the class identity of people
with low, medium, and high educational qualifications over time.

47
The New Politics of Class

a) High education b) Medium education

Middle
80% 80%

Middle
60% 60%

40% 40% Working

Working
20% 20%
None None

0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

c) Low education

80% Working

60%

40%
Middle

20%
None

0%
1960 1975 1990 2005

Figure 3.5. Class identity by education

Note: The figures here show combined prompted and unprompted class identity by education. The top-left shows
values for those with a high level of education; the top-right shows values for those with a middle level of
education; the bottom shows values for those with a low level of education. Three class identities are displayed:
middle class identity (middle), working class identity (working), and no identity (none).
Source: British Election Study 1964–2015; British Social Attitudes Surveys 2003–2015.

People with the lowest level of education are solidly working class in their
identity throughout the period examined. Among people with medium levels
of education and degree-level qualifications, there has been a shift towards a
more working class identity in recent decades. It could be that the growth in the
proportion of the population obtaining educational qualifications has reduced
the social distinctiveness of having these qualifications. It is not because the

48
 Identity

income returns to education have declined. As we saw in Chapter 2, they have

not. Controlling for any changes in income levels by different educational
levels has no noticeable effect on the pattern of overtime change in education’s
effects on class identification. Educational attainment simply appears to have
become less consequential than it used to be for a person’s sense of being
middle class.4 This same idea could perhaps also explain why levels of middle
class identification have declined slightly among the occupational middle
classes. If the social distinctiveness of educational qualifications has declined,
so has the social distinctiveness of having a middle class job. If most people are
middle class, maybe it is less easy to feel that sense of community and identity
that characterized the much smaller middle class of the 1960s.5

Class Closeness
Overall we find a great deal of continuity in the relationship between occupa-
tional class and subjective class identification. We can, however, examine a
further aspect of class identity: closeness to others with the same identity. In
the 1963 pre-election wave of the BES, Butler and Stokes tried to measure the
strength of belonging to a class community: in other words, a sense of close-
ness to one class and distance from another. This was followed up in 2005 and
in the latest wave of the BSA in 2015 to give a long-term, if not very regular,
insight into such perceptions. People were asked whether they felt that ‘they
have a lot in common with other people of their own class’? People could say
that they felt ‘pretty close to other [middle/working] class people’ or that they
didn’t ‘feel much closer to them than to people in other classes’.
The patterns are shown in Table 3.1. There is some change over time. This is
most noticeable in the proportion of people seeing themselves as ‘working
class and close to other working class people’. This drops by 16 percentage
points between 1963 and 2005, although it then increases again by 5 percent-
age points in 2015. Among the middle class there is an increase in the number

Table 3.1. The closeness of class identity

1963 2005 2015

Close to middle class 15% 15% 15%

Middle class but not close 12% 22% 23%
Neither 8% 6% 5%
Working class but not close 25% 35% 29%
Close to working class 39% 23% 28%

Note: The numbers here are percentages giving each type of response to the question of whether ‘they have a lot in
common with other people of their own class’.
Source: British Election Study 1963; British Social Attitudes Surveys 2005 and 2015.

49
The New Politics of Class

of people who feel middle class ‘but not close’ between 1963 and 2005 that
does not decline in 2015. This change could be seen as a result of the rapidly
growing size of the middle class groups. In both 1963 and again in 2015, the
average level of closeness of people who identify with the working class is
greater than it is for those who identify with middle class.6 Overall, expres-
sions of class identity seem to be a little less about class solidarity than they
were in the 1960s, but these are rather modest changes.7

How People See Others

So far we have focused on possible changes in the levels and sources of a

person’s own class identity. But it is also important to see whether people’s
views of others have changed. In his post-war study, Zweig noted: ‘As soon as a
man opens his mouth everyone knows to which class he belongs’ (Zweig
1952, p.204). Is this still true? Unfortunately we do not have the same long-
running series of questions here, but in some respects we can still compare
responses informatively across long periods of time.
What sorts of people are thought to belong to the middle and working
classes? If the meaning of class has changed in the ways suggested by the
various authors discussed previously, we might expect people to have changed
their beliefs about what it is that divides classes. Indeed, they might not have
views on classes and how they are distinguished at all. Typically, twentieth-
century research into this issue found a consistent range of criteria were used
by people to understand where people stood in the class structure. In the US
the early work of Centers (1949), showing that occupation, education, and
income were key, was replicated and developed by Coleman and Rainwater
(1979) and Jackman and Jackman (1983). In British studies, Goldthorpe et al.
(1969), Butler and Stokes (1969), Moorhouse (1976), and Marshall et al. (1988)
also show that these same three criteria dominate, a similarity further con-
firmed by comparative work (Bell and Robinson 1980). Evidence on possible
changes in the meaning of class can be obtained by simply asking people what
sorts of attributes they think of when they think about people in the middle
and working classes (Centers 1949; Goldthorpe al. 1969; Moorhouse 1976;
Bell and Robinson 1980).
We have four waves of these questions that ask people ‘what sort of people
would you say belong to the middle class?’, and ‘what sort of people would
you say belong to the working class?’. These appear in the 1963 BES (Butler
and Stokes 1969), the 1984 Class in Britain survey (Marshall et al. 1988), and
the 2005 and 2015 BSA surveys. The response options for these questions are,
unusually for survey research, open-ended. People provide their own answers
which are recorded verbatim. They are then coded post hoc by researchers. The

50
 Identity

coding of all such data can always be open to dispute, although some of these
data are publicly available for re-analysis.8 Moreover, they allow us to see if
people’s ideas about what sorts of people are working class and middle class
have changed without constraining their options.9 Table 3.2 presents the
main categories of responses: occupation, income, education, and ‘cultural’
characteristics. These are the proportions of people who mentioned any of
these characteristics. Since some people mentioned more than one thing, and
we do not show all characteristics mentioned here anyway, these percentages
do not add up to 100 per cent.
We can see that in 1963 62 per cent of people thought that occupation
defined middle class people, and 71 per cent thought that occupation defined
working class people. Jump forward forty or fifty years and we can see that
occupation remained the primary characteristic associated with class mem-
bership, but less so. In general, people seem a little less inclined to judge class
on the basis of occupation, but not noticeably a lot more willing to use other
factors like education or income except with respect to a possible increase in
references to income in relation to being working class. The same three criteria
also remain dominant. In 1963 the combined total of references to occupation,
income, and education was 89 per cent for the middle class and 83 per cent for
the working class. Similar figures for 2015 were 79 per cent and 80 per cent.
Although interesting in themselves, these broad categories do not tell us
what sorts of jobs, education, and income status are associated with being
middle class or working class. We can look more closely at the responses to the
Marshall et al. and BSA surveys from 1984 and 2015. For these we have
comparable disaggregated coding schemes and can separate out the sorts of
characteristics associated with class. We distinguish between high and low
levels of education, high and low levels of income, and different types of jobs.
We also show two other common responses: ‘workers’ and ‘ordinary people’.
The distributions of these in 1984 and 2015 are shown in Table 3.3.

Table 3.2. The meaning of class

1963 1984 2005 2015

Middle Working Middle Working Middle Working Middle Working

class class class class class class class class

Occupation 62% 71% 52% 80% 50% 64% 44% 54%

Income 22% 9% 19% 18% 36% 27% 26% 20%
Education 5% 3% 3% 1% 10% 6% 9% 6%
Manners and morals 5% 6% 5% 4% 2% 2% 5% 5%

Note: The numbers here are percentages of people who mentioned particular types of characteristics when describing
people in the working and middle classes.
Source: British Election Study 1963, Social Class in Modern Britain 1984; British Social Attitudes Surveys 2005 and 2015.

51
The New Politics of Class

Here we see very clear-cut patterns of association. Middle class people are
associated with white-collar jobs and good incomes. The opposite is the case
for the working class: associations here are with manual workers and those
with low incomes. Education has a similar pattern of class polarization to
income and occupation, but this is not a pattern that features as strongly in
these associations. Jobs and money dominate in very clearly class-structured
ways. Again we can see the weakening of occupation as a defining character-
istic, especially for the working class, but little evidence of its replacement
with anything else. This declining reference to work as a characteristic of class,
and thus the relative increased importance of income and education, might
hint at a declining status of the working class, as might the increase in
references to income noted in Table 3.2. If being working class is a more
‘stigmatized’ identity associated with low incomes, we might expect people
to find it less easy to acknowledge class pride. Unfortunately we do not have
over time survey evidence on this topic. We do, however, have two questions
that ask about this issue in the 2015 BSA survey and which explicitly contrast
the current situation with that of earlier periods. The first asks, ‘How proud do
you think people are of being working class nowadays?’ and the second,
‘Thinking about when you were growing up, how proud were people of
being working class then?’
Table 3.4 suggests that there has been a decline in the perceived pride
associated with being working class. Among everyone, as well as those with
a working class identity, there was a drop in ‘very proud’ responses for the
working class now compared to when they were growing up. These differences
are not dramatic, but they are consistent with the idea of work, and its associated
merits, being a less salient indicator of working class status than it used to be.

Table 3.3. Specific types of people associated with classes

1984 2015

Middle class Working class Middle class Working class

Business owner 17% 1% 4% 0%

White-collar job 44% 8% 32% 2%
Blue-collar job 2% 42% 1% 25%
High income 18% 0% 23% 2%
Low income 1% 17% 1% 18%
High education 3% 0% 6% 1%
Low education 0% 1% 0% 5%
‘Worker’ 8% 26% 6% 17%
‘Ordinary people’ 10% 11% 14% 11%

Note: The numbers here are percentages of people who mentioned particular types of characteristics when describing
people in the working and middle classes.
Source: Social Class in Modern Britain 1984; British Social Attitudes Survey 2015.

52
 Identity

Table 3.4. Pride in being working class

All Working class identifiers

Nowadays When you were Nowadays When you were

growing up growing up

Very proud 20% 30% 23% 38%

Quite proud 57% 53% 58% 51%
Not very proud 21% 16% 17% 11%
Not at all proud 2% 1% 2% 1%

Note: The table presents responses to two questions. The first asks: ‘How proud do you think people are of being working
class nowadays?’, and the second: ‘Thinking about when you were growing up, how proud were people of being
working class then?’ The left-hand side presents responses to these questions for all respondents and right-hand side just
for those expressing an unprompted working class identity.
Source: British Social Attitudes Survey 2015.

How People See Society

The literature on the dissolving of class assumes that class divisions do not
matter any more to people. But there is little actual survey evidence on this.
We have therefore replicated questions first asked many years ago in the early
BES surveys, to see if responses have changed on these issues. We start with the
boldest of claims, that class conflict itself is in some sense inevitable. At
various points in the BES and also in the 2015 BSA people were asked whether
they think ‘there is bound to be some conflict between different social classes’
or whether ‘they can get along together without any conflict’? Figure 3.6
shows the percentage of people who think that there is bound to be class
conflict. The primary conclusion across the fifty years covered by these ques-
tions must be one of no clear change. There is certainly no evidence of a
decline in perceptions of the inevitability of class conflict.
Class conflict is arguably a quite extreme and rare event. However, as we
have seen in Chapter 2, divisions between classes can still exist in many forms
and constitute more or less permanent constraints and barriers in society. It is
therefore helpful to look at people’s beliefs about such differences between
classes, as well as perceptions of how those differences have changed. In 1970
the BES asked people ‘how wide are the differences between social classes in
this country’ and whether they thought that ‘these differences have become
greater or less or have remained about the same’. We replicated these ques-
tions in the 2015 BSA. The responses in the two years are shown in Table 3.5.
In 2015 77 per cent thought class differences were fairly or very wide, com-
pared with 51 per cent in 1970. This is a fairly substantial change. If we look
at perceptions of change, we likewise see that people believe class divisions
are becoming more polarized in 2015 (31 per cent) than in 1970 (12 per cent).

53
The New Politics of Class

100%

75%

50%

25%

0%
1960 1970 1980 1990 2000 2010

Figure 3.6. Perceptions of class conflict

Note: The figure shows percentages of respondents agreeing that ‘there is bound to be some conflict between
different social classes’.
Source: British Election Studies 1963–1997; British Social Attitudes Surveys 2005 and 2015.

On this evidence, in 2015 people believe we live in a more class-divided

society than we used to almost half a century earlier. If the classes are
divided, does this make traversing class boundaries difficult? If transitions
between classes are commonplace, then it might also be less consequential
that there are wide class divisions. It is important to know whether people
believe it is hard to move between classes and whether classes are thought to
be more permeable than in the past. We examine this issue via a question in
the 1963 BES that was replicated in the 2005 and 2015 BSA which asked
people, ‘how difficult would you say it is for people to move from one class
to another?’ The responses are shown in Table 3.6. It seems that people
think it is, if anything, harder to move classes than they did ten years ago.
Responses are distributed somewhat differently in 1963, perhaps as a result
of question framing, but the 58 per cent who thought moves between
classes were hard in 1963 had increased to 65 per cent in 2005 and 73 per
cent in 2015.
There are also more day-to-day indicators of class boundaries in society in
terms of how people interact with people from other social classes. This is
what we might call the micro-world of class divisions. There are a couple of
very useful questions here too. Firstly, in 1979 the BES asked, ‘When you first
meet someone, how aware are you of their social class?’ and ‘how easy do you
think it is to have friends in other social classes?’. Again we replicated these
questions in 2015. Table 3.7 shows the results. In both 1979 and 2015 only
one third of people say they never really notice someone’s class on first

54
 Identity

Table 3.5. Perceptions of differences between classes

The differences between classes are . . . 1970 2015

Very wide 27% 24%

Fairly wide 24% 53%
Not very wide 40% 21%
No differences 9% 2%

The differences between social classes have become . . . 1970 2015

Greater 12% 31%

About the same 59% 44%
Less 28% 24%

Note: The table shows responses to questions on ‘how wide are the differences between social classes in this country’ and
‘whether these differences have become greater or less or have remained about the same’ (‘no differences between
classes’ includes an additional category of ‘not wide’ in 1970).
Source: British Election Study 1970; British Social Attitudes Survey 2015.

Table 3.6. Difficulty of moving between social classes

Moving between classes is . . . 1963 2005 2015

Very difficult 30% 19% 21%

Fairly difficult 28% 46% 52%
Not very difficult 42% 36% 26%

Note: The table shows responses to the question: ‘how difficult would you say it is for people to move from one class to
another?’
Source: British Election Study 1963; British Social Attitudes Surveys 2005 and 2015.

meeting. And even this is not quite a statement that they ‘don’t notice a
person’s class’ on first meeting. Either way, there is little evidence of change.
Most people are aware of class when meeting people. With regard to friends,
most people in both years say there is ‘no difficulty’ in having friends from
other classes. But there is certainly no evidence that this number is increasing;
in fact, it is falling. In 2015 there are 10 per cent fewer people who say cross-
class friendships are easy than in 1979.
Despite the social desirability biases that are likely to affect responses
to both of these questions, the difference between the years tells us that
awareness of class barriers is, if anything, a little stronger now than it was
in the late 1970s. This appears generally true of class perceptions. Some
perceptions suggest more constancy: class conflict and social contact. Some
perceptions suggest a little more class division: perceptions of wider gaps
between classes and the greater difficulty of moving between classes. None
suggests less division.

55
The New Politics of Class

Table 3.7. Awareness of social class and difficulty of having friends from other social classes

Aware of someone’s class . . . 1979 2015

Usually notice 27% 22%

Sometimes notice 40% 45%
Never really notice 33% 33%

Having friends from other classes is . . . 1979 2015

Hard 6% 7%
A little difficult 23% 32%
No difficulty 71% 61%

Note: The table shows responses to two questions: ‘when you first meet someone, how aware are you of their social
class?’, and ‘how easy do you think it is to have friends in other social classes?’
Source: British Election Study 1979; British Social Attitudes Survey 2015.

Conclusions

In this chapter we have shown the resilience of class identities and the
persistence of people’s awareness of class as a source of division. Most people
still express class identities and these show no evidence of decline over the last
fifty years. Moreover, unlike research in the US that has found a growing
proportion of the public holding middle class identities, British people with
middle class jobs are quite likely to see themselves as working class. In part,
this is because class identity reflects where someone comes from as well as
where they end up, and many people in middle class jobs come from working
class backgrounds. Over time this tendency has even increased slightly, as
educational qualifications have lost some of their power to convey a sense of
being middle class. Despite fifty years of transition from an industrial to post-
industrial society, expressions of a working class identity remain important.
The idea that we are ‘all middle class now’ is very far removed from the reality
of class identity in contemporary Britain.
Furthermore, it is not just in their self-perceptions that people express the
persistence of class. Evidence on beliefs about the meaning of class and the
extent of class divisions is less extensive, but the various indicators we have
examined echo the resilience of class identities: the nature of classes, the
likelihood of class conflict, whether people notice a person’s class, the impact
of class on friendships, the extent of class boundaries, the possibility of
moving between classes, are all at least as pronounced in 2015 as they were
back in the 1960s and 1970s.10
People are a little less likely to consider class membership to be defined
by the job that someone does. This may reflect the growth in income inequal-
ity between the working class and others documented in Chapter 2. If
the working class are becoming ‘the poor’ rather than ‘the workers’, then it

56
 Identity

might mean that the term could also be socially stigmatizing. Some journalists
seem to have assumed this is strongly the case, and have written as though the
word ‘chav’ is broadly synonymous with ‘working class’ (Jones 2011). But this
seems a dramatic overstatement. The changes that we find are small, and in
fact, working class identities, and class identities more generally, appear to be
remarkably robust over time.
In summary, class awareness is, if anything, higher now than it was forty or
fifty years ago. Changing class sizes and increased upward mobility have not
weakened these indicators of social division. This pattern of continued dis-
tinctiveness is also evident when considering policy attitudes, as we do in
Chapter 4. Persistent class divisions in resources, risks, opportunities, and
educational attainment foster continuing differences in political ideology.

Notes

1. These are distinct from ‘class consciousness’, in which ‘an awareness of common
interests . . . leads to action through political representation’ (Oddsson 2010: 293).
Typically, this form of class-based political action has been expressed via class
voting—‘the democratic class struggle’ (Anderson and Davidson 1943). We exam-
ine this in Chapters 7 and 8.
2. In the 1950s 40 per cent of people who gave a class identity in the US placed
themselves in the middle class and 60 per cent in the working class. By 2000 59 per
cent said middle class and 41 per cent working class.
3. It is unclear whether the recent recession might have influenced social compari-
sons. Evans and Kelley (2004) argue that higher unemployment might exert a
downward force on self-placement in the social hierarchy as people feel worse
about their position because of employment insecurity. Others have suggested
that experiences of distress are cushioned because people feel themselves ‘to be in
the same boat’ (Ragnarsdottir et al. 2013), while Oddsson (2010) believes the
pressures associated with recession might increase class awareness.
4. This pattern is confirmed by multivariate logistic models of class identity that
include people’s own occupational class and education, and their father’s occupa-
tional class. Across the period as a whole, the effects of these three factors are more
or less equivalent in affecting class identity. However, by 2015 the effects of
education had declined. In 2015 the impact of higher education compared to a
low level of education fell to only two thirds of the overall average for the com-
bined surveys. By contrast, the effects of own occupational class and fathers’ class
were indistinguishable from the average for the combined dataset.
5. A further interesting possibility is that the high proportion of working class identi-
fiers in the middle classes may be due to such people seeing society as composed of a
small privileged elite that is distinct from everyone else, including themselves, who
are by comparison ‘working class’. Systematic evidence on such images of the
structure of society is unfortunately rather limited. However, we were able to look

57
The New Politics of Class

at this issue using quite recent BSA data (2009), in which respondents were pre-
sented with several different distributions of inequality and asked for their views on
which of them Britain was most like. There was an indication that people who
described themselves as working class were more likely to see society as divided
between a large disadvantaged group and a smaller privileged elite (for further
analysis see Evans and Mellon 2016b).
6. This fits with Surridge’s (2007, p.213) recent analysis of class identity using data
from 2003, which found that working class identities are a little more salient than
middle class ones when asked as part of a larger battery of different social identities:
‘24 per cent selected a working-class identity as one of their three primary iden-
tities, a very similar proportion to those who said they were working class in the
“standard” identity question. However, only 8 per cent gave middle class as one of
their three identities, compared with 19 per cent on the more usual identity
question’.
7. We should also be slightly cautious about interpreting some of these changes
because the 1963 answers are based on an initially prompted class question and
the 2005/2015 answers are based on an initially unprompted question. This means
that more people say that they are middle or working class in 1963 after the first
question and may therefore feel that they ‘should be’ close to their class.
8. Sadly the original uncoded responses to the 1963 survey are no longer obtainable.
9. These open-ended questions also assume that people are able to access and articu-
late their beliefs. This assumption reflects the standard problem of recall versus
recognition in survey measurement (Vanneman and Cannon 1987, pp.102–6). As
it happens, indirect approaches to eliciting class-based expectations employed to
address this concern (Evans 1993a, 1997; Stubager et al. 2016) produce reassuringly
similar results.
10. Heath et al. (2009) found some evidence of decline in some aspects of class identity
in their comparison of data from 2005 with earlier periods. It is possible that this
reflected the optimism of the long-disappeared economic boom of the ‘noughties’.
If so, it appears to have been quickly reversed.

58
4

Ideology

This chapter moves away from looking at what class means to people, both
objectively and subjectively, to how class shapes political ideology. Given the
way in which class influences people’s lives at every stage, it is unsurprising
that people’s occupation and education are a strong determinant of how
people think that the world should be organized. In this chapter we show
that occupational class and educational attainment provide a structure to
views on long-standing economic issues, like redistribution and public own-
ership, but also non-economic moral issues and conflicts over EU integration
and mass immigration. Moreover, we also show that most of this structure has
remained unchanged over the last fifty years. Class, in its broader sense, still
determines how people think about fairness, what the role of the state should
be, and how they define what is right or wrong.
As always, there is some devil in the detail. First, while differences between
classes over economic policy are largely stable there have been some changes.
For economic issues, we show that differences were most pronounced during
the 1980s. This means that while there was a degree of convergence during the
1990s, this still left class differences much as they were in the 1960s. Second,
while economic issue positions are largely driven by occupational class, edu-
cation is as important for predicting non-economic positions. It is the inter-
twined nature of education and occupation that generates positions towards
EU integration and immigration. Other non-economic attitudes are largely
divorced from people’s economic situation. Issues of moral ‘rights’ and
‘wrongs’, such as attitudes towards homosexuality and child rearing, are
almost exclusively driven by education.
Nonetheless, while some of the trees might be a bit different to one another,
the shape of the wood is clear: people’s position in society affects how they
view political issues in a way that has not changed very much over the last five
decades. People today may be generally less keen on privatization than they
were in the 1980s, or for that matter less keen on criminal executions than
they were in the 1960s, but the class distinctions that mark both of these
The New Politics of Class

issues are just as prevalent. At the end of this chapter we show that the broad
positioning of the classes, defined in terms of occupation and education, on
economic left–right issues and social authoritarianism-liberalism is remark-
ably consistent over time. The professional classes are centrist economically
and socially liberal; the managerial and bourgeois classes are right-wing
economically and socially authoritarian; the working class is left-wing eco-
nomically and socially authoritarian. Just as people’s experiences and views
of class have not really changed, nor have class positions on the big political
issues.
This chapter is organized by issues and we start with the dominant conflict
within post-war British politics: the organization of the economy. Where do
people stand on public ownership versus private ownership; on whether
income should be redistributed; and how conflicts between employers and
employees should be resolved? The second section looks at issues that are
broadly separate from this major economic left–right dimension: attitudes
towards immigration; the EU; law and order; and ‘moral’ issues such as toler-
ance of homosexuality and attitudes towards child rearing. The final section
then imposes a structure on these disparate attitudes to show how class
generally matters for political ideology.

A Bit to the Left, a Bit to the Right

Many political issues boil down to issues of ownership and equality. Should
industries be privately owned? Do workers get a fair deal compared to owners?
Should wealth be redistributed from those who currently own it to those who
do not? These are typically what we think of as economic left–right issues and
are thought to be informed by self-interest. People with less want more, and
people with more want to hang on to what they already have.1 These are
therefore precisely the kind of issues that we might expect to be shaped by
social class. Indeed there is plenty of evidence, from Britain and other coun-
tries, that this is the case. People in working class jobs are typically more
economically egalitarian than the middle classes (Evans 1993c; Bartels 2008;
Houtman et al. 2008; Weakliem and Heath 1994; Corneo and Gruner 2002;
Hayes 1995; Linos and West 2002; Kalmijn and Kraaykamp 2007). As Lipset
puts it: ‘In all democratic nations . . . there has been a correlation between
socioeconomic status and political beliefs and voting. The less privileged
have supported parties that stood for greater equality and welfare protection,
through government intervention, against the strain of a free enterprise econ-
omy’ (Lipset 1991, p.208).

60
 Ideology

The problem is that most of this previous work treats these economic left–
right values and attitudes rather crudely and most importantly examines just a
single point in time.2 Moreover there are divisions within the middle class
that are often not taken seriously (see, however, Kalmijn and Kraaykamp
2007). People more closely connected to the marketplace (employers) and
people in positions of authority (managers) are more likely to side with free
market principles than the professional middle classes who are more isolated
from the marketplace and managerial responsibilities. Here we take a more
nuanced approach and look at the same, or very similar, questions over long
periods of time, carefully identify different occupational groups, and finally
separate these economic issues into three distinct elements. The first is about
ownership and control. How do people think industries should be run and
who benefits from capitalist institutions? The second is about equality. How
do people think income and wealth should be distributed and what kind of
power does wealth bring? The third is about labour relations. What role
should trade unions play in society and to what extent are management and
employee interests opposed to one another?

The Means of Production

We start by looking at questions of ownership and control. We focus on
occupational class here, as it is by far the strongest predictor of left–right
attitudes in Britain (as it is comparatively, Kalmijn and Kraaykamp 2007).
Throughout this chapter, and mainly to keep the number of lines on graphs
to a manageable level, we just show the old middle class, new middle class,
and working class groups in all the figures. Tables also show the other major
middle class group: the junior middle class. As in Chapter 3, we use a combin-
ation of data from the BES surveys (1963–2015) and the BSA surveys
(1983–2015).
The left-hand graph in Figure 4.1 shows how many people support further
privatization3 from 1963 through to 2005 for the old middle class, the new
middle class, and the working class. There are two obvious points to make.
Support for privatization is highest in the 1980s, particularly 1983. Second,
the differences between the class groups are very constant, albeit exaggerated
somewhat in the early 1980s when the issue is most salient. The working class
are consistently the most opposed to privatization (even in 1983 fewer than a
third support further sell-offs of publically owned industries) and the old
middle class are the most supportive. There is a very similar pattern when
we look at who is seen to benefit from the capitalist economy. The right-hand
graph in Figure 4.1 shows the proportion of people who think that ‘big
business benefits owners rather than workers’. The working class are quite

61
The New Politics of Class

a) More privatisation b) Workers benefit more than owners

60% OMC
40% OMC

40%
NMC

20%
20%
NMC WC

WC
0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.1. Views on ownership by occupational class

Note: The left-hand graph here shows the proportion of people who agree that there should be ‘more privatization
of companies by government’. The right-hand graph shows the proportion of people who disagree that ‘big
business benefits owners rather than workers’ (two period moving average). Three occupational class groups are
displayed: old middle class (OMC), new middle class (NMC), and working class (WC).
Source: British Election Studies 1963–2005; British Social Attitudes Surveys 1986–2015.

obviously the least likely to think that workers benefit, and the old middle
class the most likely to think that workers benefit. Again these differences are
remarkably constant over time.
Overall, it is quite clear that there are substantial differences in how
people view the costs and benefits of public ownership and identify the
winners and losers from the free market. These differences are primarily
driven by occupational class. Table 4.1 shows results from a logit regression
analysis, pooling the data by decade, for the effects of occupational class
(this time for all four major occupational class groups) holding constant
education level and a large number of other variables. These are: region, age
group, trade union membership, sector of work, agricultural employment,
housing tenure, religious denomination, gender, year, and ethnicity.4 Con-
trolling for all these other factors, including education, occupation still
obviously matters for both attitudes towards privatization and more general
views of private ownership.5 The old middle class are most supportive of
private enterprise and the working class least supportive, with the new and
junior middle class groups somewhere in between. These differences are
greatest in the 1980s, but are clearly present before and after then. Holding
constant lots of other facets of people’s lives that might be related to their
occupational class does not get rid of these consistent, and substantial,
differences between people in different class positions on how the economy
should be organized.

62
 Ideology

Table 4.1. Impact of occupational class on attitudes towards ownership

Want more privatization 1960s 1970s 1980s 1990s 2000s

Old middle class 24% 42% 50% 50% 18%

New middle class 18% 41% 40% 39% 17%
Junior middle class 21% 44% 40% 39% 15%
Working class 17% 32% 30% 30% 12%
Old middle class - working class 7% 10% 21% 19% 6%

Think workers benefit from big business 1980s 1990s 2000s 2010s

Old middle class 52% 37% 39% 33%

New middle class 40% 25% 29% 25%
Junior middle class 39% 24% 26% 20%
Working class 30% 19% 21% 19%
Old middle class - working class 22% 17% 17% 14%

Note: The numbers here are predicted probabilities from logit regression models using pooled data by decade that predict
agreement with the statement that there should be ‘more privatization of companies by government’ and disagreement
with the statement that ‘big business benefits owners rather than workers’. As well as occupational class, these models
include controls for education, housing, trade union membership, gender, age, region, agricultural employment, religion,
race, and year. Employment sector is also included in the second set of models. The predicted probabilities are for a white
Anglican man in his forties, who is a homeowner, has middling educational attainment, lives in the south east of England,
and is not a trade union member. Year is set as close as possible to the middle of each decade.
Source: British Election Studies 1964–2005; British Social Attitudes Surveys 1986–2015.

Some Animals Are More Equal Than Others

The second pillar of left–right ideology concerns equality, and especially
equality of income and wealth. To what extent are there clear class differences
in how people see inequality and in support for policies designed to rectify
those inequalities? The left-hand graph in Figure 4.2 shows how opposition to
the redistribution of income and wealth has changed since 1974.6 Again it is
striking how large the differences are between the occupational class groups.
The old middle class is typically about 20 per cent more likely to oppose
redistribution than the working class, with the new middle class generally in
between. There is some evidence of small changes here as well. The gaps between
groups after the mid-1990s are somewhat smaller than in the 1970s and 1980s.
There is less change over time when we look at underlying notions of inequality
and unfairness. The middle and right-hand graphs in Figure 4.2 show responses
to questions about whether people agree that there ‘is one law for the rich and
one law for the poor’ and whether ‘ordinary people get a fair share of the wealth’.
There are large differences between occupational groups from when the data
starts in 1986 through to 2015. Interestingly, the differences here are much more
clearly between the working class and the two middle class groups.
Table 4.2 shows estimates from a logit regression model which allows us to
look at the effects of occupational class over time, controlling for all sorts of
other variables. Again the data is pooled separately for each decade in order to

63
The New Politics of Class

a) Want less redistribution b) Same law for poor and rich c) Ordinary people get fair share

60% 60% OMC 60%

OMC
OMC
NMC
40% 40% 40% NMC

NMC
WC
20% 20% 20% WC
WC

0% 0% 0%
1960 1985 2010 1960 1985 2010 1960 1985 2010

Figure 4.2. Views on equality by occupational class

Note: The left-hand graph here shows the proportion of respondents who disagree that ‘income and wealth should
be redistributed to ordinary people’ (two period moving average). The middle graph shows the proportion who
agree that there is ‘one law for the rich and one law for the poor’ (two period moving average). The right-hand
graph shows the proportion who agree that ‘ordinary people get a fair share of the wealth’ (two period moving
average). Three occupational class groups are displayed: old middle class (OMC), new middle class (NMC), and
working class (WC).
Source: British Election Studies 1974–2015; British Social Attitudes Surveys 1985–2015.

allow us to track any changes, and we include the same long list of control
variables as in the previous models in Table 4.1. The story here is rather similar
to the one regarding ownership. Holding everything else constant, occupa-
tional class still clearly matters. People with working class jobs consistently
perceive more inequality, and want more measures to reduce inequality, than
people in either the old or new middle class. While these differences are fairly
stable, it is interesting that they also appear to peak in the 1980s, just as they
did for attitudes towards privatization.

I’m Alright, Jack

Finally we turn to labour relations. The left-hand graph in Figure 4.3 shows
how people’s views of trade union power have changed since 1963.7 And they
certainly have changed. When people were asked whether trade unions had
too much or too little power in the 1960s, almost everybody thought unions
had too much power, but over the course of the 1980s, people’s perceptions
changed rapidly as the unions were defanged by a combination of mass
unemployment and government legislation. In the 1960s and 1970s there
was also a clear distinction between the classes as to whether unions were too
powerful. The working class were considerably less likely than the two middle
class groups to think unions had too much power. This divide narrowed con-
siderably in the 1980s and although it remains the case that the old middle class

64
 Ideology

Table 4.2. Impact of occupational class on attitudes towards equality

Want less redistribution 1970s 1980s 1990s 2000s 2010s

Old middle class 35% 51% 48% 53% 47%

New middle class 31% 42% 38% 46% 40%
Junior middle class 31% 40% 39% 43% 37%
Working class 18% 28% 27% 36% 32%
Old middle class - working class 17% 23% 21% 17% 15%

Ordinary people get a fair share 1980s 1990s 2000s 2010s

Old middle class 70% 61% 63% 62%

New middle class 64% 52% 55% 55%
Junior middle class 62% 53% 52% 49%
Working class 48% 43% 46% 47%
Old middle class - working class 21% 17% 18% 15%

Rich and poor treated equally 1980s 1990s 2000s 2010s

Old middle class 63% 51% 61% 59%

New middle class 54% 45% 55% 54%
Junior middle class 50% 44% 51% 49%
Working class 39% 34% 45% 44%
Old middle class - working class 24% 17% 17% 16%

Note: The numbers here are predicted probabilities from logit regression models using pooled data by decade that
predict a) disagreement that ‘income and wealth should be redistributed to ordinary people’, b) agreement that there is
‘one law for the rich and one law for the poor’, and c) agreement that ‘ordinary people get a fair share of the wealth’. As
well as occupational class, these models include controls for education, housing, trade union membership, gender, age,
region, agricultural employment, religion, race, and year. Sector of employment is also included in the second and third
set of models. The predicted probabilities are for a white Anglican man in his forties, who is a homeowner, has middling
educational attainment and lives in the south east of England, and is not a trade union member. Year is set as close as
possible to the middle of each decade.
Source: British Election Studies 1974–2015; British Social Attitudes Surveys 1986–2015.

are more worried about union power than the working class, the differences
today are smaller. If we look at more underlying attitudes to employer/employee
relations, we see almost no change, both in terms of absolute beliefs and also
divisions between the classes. The right-hand graph in Figure 4.3 shows whether
people disagree that ‘management will always try to get the better of employees
given the chance’. Very few working class people disagree with this, but a sizable
minority of the old middle class take a more benevolent view of how those at
the top of the hierarchy deal with those at the bottom. There is little change in
this difference when we compare 1985 (when this question was first asked) with
today. In fact, the 25 per cent gap between the old middle class and the working
class in 1985 had actually increased to 28 per cent by 2015.
This divergence in underlying attitudes about labour relations from specific
views on union power is not surprising, given the radical change in the nature
and scale of unionization in Britain over the last fifty years. Figure 4.4 shows the
proportion of people who are members of a union by their occupational class
from 1963 to today. In the 1960s over 40 per cent of working class people were in

65
The New Politics of Class

a) Unions have too much power b) Management treats employees well

100%
OMC
OMC
80% 40%
NMC

60% NMC

40% WC 20%
WC
20%

0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.3. Views on labour relations by occupational class

Note: The left-hand graph here shows the proportion of people who agree that ‘trade unions have too much
power’. The right-hand graph shows the proportion of people who disagree that ‘management will always try to
get the better of employees given the chance’ (two period moving average). Three occupational class groups are
displayed: old middle class (OMC), new middle class (NMC), and working class (WC).
Source: British Election Studies 1974–2005; British Social Attitudes Surveys 1985–2015.

a union, but less than a quarter of the new middle class were unionized and
very few members of the old middle class were affiliated with a union.
Given the relative size of the working class, this meant that most union
members had working class jobs. In fact, if we include foremen as well, the
working class made up well over three quarters of union members in 1964.
By the end of the 1990s this position was completely changed. Today,
around half of union members are located in the new middle class. It is
not surprising then that people outside the working class are less likely to
see unions as too powerful, since any power that they do retain is increas-
ingly used to help their predominantly middle class, and often degree-
educated, membership.
In fact, once we control for education and, particularly, trade union mem-
bership, we see much less change in how unions are perceived. Table 4.3 shows
percentages, derived from a model that holds constant the long list of other
variables mentioned previously, of people thinking unions have too much
power and who think that management does not take advantage of employees
by occupational class. For the latter, the differences between the occupational
groups shrink, but remain large and constant over time. More interestingly,
once we hold constant some of these other factors (and, unsurprisingly, trade
union membership is a particularly good predictor of how people view
unions), the differences between the classes’ views of union power change a

66
 Ideology

WC
40%

NMC

20%

OMC

0%
1960 1970 1980 1990 2000 2010

Figure 4.4. Union membership over time by occupational class

Note: The figure here shows the proportion of people who state that they are currently a member of a trade union or
staff association (three period moving average). Three occupational class groups are displayed: old middle class
(OMC), new middle class (NMC,) and working class (WC).
Source: British Election Studies 1963–2015; British Social Attitudes Surveys 1983–2015.

Table 4.3. Impact of occupational class on attitudes towards labour relations

Think unions have too much power 1960s 1970s 1980s 1990s 2000s 2010s

Old middle class 80% 88% 69% 39% 32% 35%

New middle class 84% 88% 61% 33% 24% 32%
Junior middle class 80% 87% 64% 34% 25% 39%
Working class 66% 77% 56% 33% 18% 26%
Old middle class - working class 14% 11% 13% 6% 15% 9%

Think management does not take advantage of employees 1980s 1990s 2000s 2010s

Old middle class 45% 36% 40% 41%

New middle class 38% 24% 28% 30%
Junior middle class 34% 24% 22% 24%
Working class 21% 14% 16% 15%
Old middle class - working class 24% 22% 24% 26%

Note: The numbers here are predicted probabilities from logit regression models using pooled data by decade that
predict a) agreement that ‘trade unions have too much power’ and b) disagreement that ‘management will always try to
get the better of employees given the chance’. As well as occupational class, these models include controls for education,
housing, trade union membership, gender, age, region, agricultural employment, religion, race, and year. Sector of
employment is also included in the second set of models. The predicted probabilities are for a white Anglican man in his
forties, who is a homeowner, has middling educational attainment and lives in the south east of England, and is not a
trade union member. Year is set as close as possible to the middle of each decade.
Source: British Election Studies 1963–2015; British Social Attitudes Surveys 1985–2015.

67
The New Politics of Class

lot less over time. The gap between the middle classes and the working class
now looks almost constant. Nonetheless, it does seem that the new middle class
is slightly more favourable towards unions now than it was fifty years ago. This
should not surprise us given that unions look more and more like a series of
professional organizations than the labour movement of old.
Different aspects of left–right economic ideology have been more or less
important, and more or less divisive, at different points over the last fifty years.
This means that while there does appear to be some convergence in the
attitudes of occupational class groups since the 1980s, this is probably a return
to the differences that existed in the 1960s and 1970s. In fact, this pattern of
gently increasing difference from the 1970s to 1980s and then gently decreas-
ing differences from the 1990s onwards is true for both support for privatiza-
tion and support for redistribution. Nonetheless, we should not let these
changes overshadow the most important point. What is most notable is not
change, but the consistent, large differences between occupational class
groups on these attitudes, even when holding constant a huge variety of
other characteristics such as housing, region, age, and, importantly, educa-
tion. To argue that classes are now indistinguishable from one another in
terms of attitudes towards the free market is simply incorrect.

Beyond Left and Right

Although British politics is often thought to have been dominated by eco-

nomic left–right issues, there are other divisions within society and within
the political system. These are issues that are often thought to form a second
dimension in most political systems, sometimes called the ‘new politics’ or
the authoritarian–libertarian dimension (Kitschelt 1994, 1995; Flanagan
1987). We divide these other issues into three different policy areas. The
first element concerns issues that are still related to economic concerns, but
are also clearly separate from the classic capitalism versus socialism debate.
This includes attitudes to EU integration and immigration. Positions on both
are related to economic interests, but also wider notions of national identity
and culture.
The second and third policy areas are largely separate from the economic
world, however. Here we are less concerned with people’s views on the rights
and wrongs of the market economy and more concerned with how they see
right and wrong in a moral sense. Stubager (2008, 2009, 2010) makes an
extremely useful distinction between attitudes towards social hierarchies
and attitudes towards social tolerance. Our second policy area taps into ideas
about hierarchy and authoritarianism. How much importance should we
attach to order and how much power should we give to those who uphold

68
 Ideology

that order? This particularly includes attitudes towards people who break the
law and subvert that order. In other words, to what extent should the state
focus on punishing criminals for their crimes? The third of our policy areas is
about social tolerance. This is perhaps best thought of as moral conservatism.
Should people with ‘unconventional’ lifestyles be tolerated or should they be
forced to conform?
These are all important issues in the political realm, and there is a great deal
of evidence that occupation and education matter for how people weigh their
pros and cons. For EU membership and immigration, these are differences that
are partially motivated by the same economic concerns as we have already
seen: further immigration and EU integration is less appealing to people who
do not economically benefit. More generally, there is a long-standing body of
research that suggests that occupation and education matter for both authori-
tarianism and tolerance. Lipset’s famous maxim that ‘the more well-to-do are
more liberal, the poorer are more intolerant’ (Lipset 1959, p.102) sums up
much early research that built on Adorno et al.’s classic, albeit now somewhat
discredited, work The Authoritarian Personality (1950). There is further nuance
to Lipset’s findings, however. He actually shows repeatedly the now standard
finding that, on a ladder of tolerance and liberalism, it is those in professional
jobs, our new middle class, who are at the top, with the working class at the
bottom and the old middle class somewhere in the middle. Later work has
questioned the universalism of this relationship, often suggesting that educa-
tion, not occupation, is the most important factor (Dekker and Ester 1987;
Houtman 2003; Ray 1983; Weakliem 2002).
There are different models for how education shapes attitudes, although it
seems that socialization within the educational system is most important
(Stubager 2008). It is less clear exactly how this happens. Some argue for a direct
impact of the educational curriculum. Supporting this claim, there is evidence
that field of study is important in shaping attitudes; people with humanities and
social sciences degrees are more socially liberal due to the nature of their courses
than those with natural science or business degrees (van de Werfhorst and de
Graaf 2004; van de Werfhorst and Kraaykamp 2001; Surridge 2016). Others
claim that most of these education effects are due to selection into different
educational trajectories by people with different views (Lancee and Sarrasin
2015). There is undoubtedly also a reinforcing effect of tutors’ predominantly
liberal views and conformity with fellow students’ liberal views (Jacobsen 2001).
Regardless of the exact mechanisms, there is no doubt that education, and
maybe occupation, are linked to social liberalism. In the next section we
disentangle the different aspects of this dimension of politics into three:
attitudes towards other nationalities and supra-national cooperation; atti-
tudes towards authoritarianism, especially with regards to crime; and moral
conservatives, especially attitudes towards tolerance of particular groups.

69
The New Politics of Class

Bloody Foreigners
We start by looking at attitudes towards immigration and EU membership.
These are in some ways closest to economic issues. After all, they involve
economic costs and benefits that affect different types of people in different
ways. Indeed, historically, attitudes towards the EU were generally thought to
be shaped by people’s levels of human capital. This meant that people in
professional jobs with lots of education (who were well placed to benefit from
the market opportunities that flowed from the EU integration process) were
broadly supportive of the EU, while people with fewer marketable skills
who occupied a more vulnerable economic position were more sceptical
about EU integration (Anderson and Reichert 1995; Gabel and Palmer 1995;
Gabel 1998a; Gabel 1998b; Balestrini 2012; Hobolt 2014). This is a direct way of
thinking about how self-interest, embodied in social class, affects attitudes.
There is also a more indirect way, however, because while attitudes towards
these issues are affected by economic concerns, ‘cultural’ concerns are also very
important. It is now widely recognized that support for EU integration is
partially a function of group loyalties and cultural threat (McLaren 2002,
2004, 2006, 2007; Hooghe and Marks 2004, 2005; Garry and Tilley 2009; van
Klingeren et al. 2013). Since views of national identity and cosmopolitanism
are often produced by educational and occupational experiences,8 both these
external economic and cultural threats to people are likely to be shaped by their
social class. Very similar arguments about both economic and cultural threats
are typically made for how people think about immigration policy.9
The above points to a greater role for education in determining these attitudes.
Occupational class certainly matters for economic threat, but less so for cultural
threat, whereas education matters for both economic and cultural threat.
Figures 4.5 and 4.6 therefore show how attitudes towards these two issues have
changed since the 1960s by both occupational class and education. As there has
been no continuously asked series of questions over this period of time, Figure 4.5
uses a number of different questions to measure opposition to immigration.10
This means that changes in the average level of opposition to immigration are not
necessarily easy to interpret. Even if it were the same question over time, since the
most common formulation asks whether there is ‘too much immigration’, this
could simply reflect changes in actual levels, or types of, immigration.
More important for our purposes are the differences among groups. The
occupational class markers we use are the same as those discussed previously,
but we also show the effects of education. As discussed in Chapter 1, we
mainly show the differences among the three main groups of high education
(people with a degree), medium education (left school at 17/18 or A Level
equivalent qualifications), and low education (left school at minimum leaving
age or no qualifications).

70
 Ideology

As Figure 4.5 shows, people’s views on immigration are related to both

education and occupational class. People in working class jobs with low levels
of education are most opposed to immigration, and people in new middle
class jobs with degrees are the least opposed. Note that it is the new middle
class group that is most enthusiastic about immigration, as we expected. These
differences are relatively large, especially for education, and very constant.
While opposition to immigration has generally risen over time, whether due
to real increases in immigration rates, changing views about immigration, or
changed question wording, the differences among groups have remained the
same or actually increased.
There is a similar story for opposition to EU membership, shown in Figure 4.6.
Here we show opposition to joining the Common Market prior to the 1970s and
then subsequently support for withdrawal from the EEC/EC/EU.11 Clearly levels
of opposition to EU membership vary quite dramatically over time, peaking in the
1970s and early 1980s, then falling dramatically, before rising again towards the
end of the 2000s. While it is almost always the case that the same groups are most
opposed to EU membership (the working class and the least educated), the
differences between people do change a little over time. In particular the gaps
between class and educational groups are larger in the early 1980s and 2010s.

a) Occupational class b) Education

80%
Low
70% WC
60%

OMC
50% Medium
40%

30%
20%
NMC High

10% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.5. Strongly against more immigration by occupational class and education
Note: The figures here show the proportion of people who strongly agree that immigration should be reduced, or a
close variant of this question (two period moving average). The left-hand graph disaggregates the results by
occupational class; the right-hand graph disaggregates the results by education. Three occupational class groups
are displayed: old middle class (OMC), new middle class (NMC), and working class (WC). Three educational groups
are displayed: people with degree-level education (high), people with A Level equivalent education (medium), and
people who left school at the minimum school leaving age for their cohort (low).
Source: British Election Studies 1963–2015; British Social Attitudes Surveys 1986–2013.

71
The New Politics of Class

a) Occupational class b) Education

60% 60%
WC Low

Medium
40% 40%

20% OMC 20%

High
NMC
0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.6. Against EU membership by occupational class and education

Note: The figures here show the proportion of people who agree that Britain should leave the EU, or not join the
EEC in earlier years (three period moving average). The left-hand graph disaggregates the results by occupational
class; the right-hand graph disaggregates the results by education. Three occupational class groups are displayed:
old middle class (OMC), new middle class (NMC), and working class (WC). Three educational groups are displayed:
people with degree-level education (high), people with A Level equivalent education (medium), and people who
left school at the minimum school leaving age for their cohort (low).
Source: British Election Studies 1963–2015; British Social Attitudes Surveys 1986–2013.

Tables 4.4 and 4.5 back up these conclusions. Here we show the effects of
occupation controlling for education, and vice versa, on opposition to
immigration and EU membership respectively. We also hold constant the
litany of other factors mentioned previously such as ethnicity, trade union
membership, and age group. Both occupational class and education matter,
although here it is education that dominates, and both have a constant
effect over time. It is the two groups that are shrinking—the working class
and those with low levels of education—that are most opposed to immi-
gration and EU membership. These patterns are fairly constant, especially
for immigration, but it is also clear that the divisions today between edu-
cational and class groups are some of the largest that we have seen in the
post-war period.

Crime and Punishment

Although Lipset (1959) makes little direct reference to how class shapes atti-
tudes towards law and order, the clear implication is that the working class
and people with less education have more punitive views towards criminals.
Many authors have conceptualized authoritarianism in terms of attitudes
towards criminals (Ray 1982). In Britain, Heath et al. (1985, 1991) argue that

72
 Ideology

Table 4.4. Impact of occupational class and education on attitudes towards immigration

Want less immigration 1960s 1970s 1980s 1990s 2000s 2010s

Old middle class 44% 40% 67% 54% 63% 57%

New middle class 42% 34% 69% 51% 48% 54%
Junior middle class 42% 43% 68% 54% 52% 59%
Working class 42% 44% 61% 56% 59% 63%
Old middle class – working class 1% 11% 9% 5% 11% 8%
High education (degree) 36% 27% 47% 32% 28% 29%
Medium education (A Level) 41% 30% 63% 43% 42% 46%
Low education (school leaving age) 53% 49% 72% 56% 61% 62%
High education – low education 16% 22% 25% 24% 33% 33%

Note: The numbers here are predicted probabilities from logit regression models using pooled data by decade that
predict agreement that immigration should be reduced (or a close variant of this question). As well as occupational class
and education, these models include controls for housing, trade union membership, gender, age, region, agricultural
employment, religion, race, and year. The predicted class probabilities are for a white Anglican man in his 40s, who is a
homeowner, has middling educational attainment and lives in the south east of England, and is not a trade union
member. The predicted education probabilities are for the same type of person in the junior middle class category. Year is
set as close as possible to the middle of each decade.
Source: British Election Studies 1963–2015; British Social Attitudes Surveys 1986–2013.

Table 4.5. Impact of occupational class and education on attitudes towards the EU

Support leaving EU 1960s 1970s 1980s 1990s 2000s 2010s

Old middle class 33% 22% 28% 22% 22% 26%

New middle class 31% 22% 32% 20% 22% 33%
Junior middle class 38% 28% 33% 24% 26% 34%
Working class 36% 34% 42% 28% 28% 38%
Old middle class – working class 3% 12% 14% 6% 6% 12%
High education (degree) 32% 38% 19% 12% 14% 19%
Medium education (A Level) 29% 41% 26% 18% 22% 27%
Low education (school leaving age) 31% 53% 42% 26% 28% 42%
High education – low education 1% 15% 23% 15% 14% 23%

Note: The numbers here are predicted probabilities from logit regression models using pooled data by decade that
predict agreement that Britain should leave the EU (or not join in earlier years). As well as occupational class and
education, these models include controls for housing, trade union membership, gender, age, region, agricultural
employment, religion, race, and year. The predicted class probabilities are for a white Anglican man in his forties, who
is a homeowner, has middling educational attainment and lives in the south east of England, and is not a trade union
member. The predicted education probabilities are for the same type of person in the junior middle class category. Year is
set as close as possible to the middle of each decade.
Source: British Election Studies 1963–2015; British Social Attitudes Surveys 1986–2013.

attitudes towards crime and the death penalty form a core part of the second
dimension of politics in Britain and education is a key part of explaining
divisions on issues within the middle class occupational groups. This is echoed
by more recent work that focuses on education and how the educational
system instils ‘in students a set of libertarian values that continue to influence
their thinking long after they have completed their education’ (Stubager 2008,
p.331; see Surridge 2016 on the British case).

73
The New Politics of Class

a) Occupational class b) Education

80% WC 80% Low

60% 60% Medium

OMC

40% 40%
NMC

High
20% 20%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.7. Support death penalty by occupational class and education

Note: The figures here show the proportion of people who agree that the death penalty should be reintroduced, or
retained in earlier years (two period moving average). The left-hand graph disaggregates the results by occupational
class; the right-hand graph disaggregates the results by education. Three occupational class groups are displayed:
old middle class (OMC), new middle class (NMC), and working class (WC). Three educational groups are displayed:
people with degree-level education (high), people with A Level equivalent education (medium), and people who
left school at the minimum school leaving age for their cohort (low).
Source: British Election Studies 1963–2001; British Social Attitudes Surveys 1983–2015.

It is therefore not surprising that Figures 4.7 and 4.8 show that support
for the death penalty12 and stiffer sentences for criminals in Britain over
time are strongly related to occupational class and education. People in the
working class are about 20 percentage points more likely to support the
death penalty than people in the new middle class, and while overall sup-
port has dropped a little since the 1990s, this gap is actually larger today
than it was fifty years ago. These differences are even bigger by education,
with 30–40 percentage point gaps between the most and least educated
groups. Similar points can be made about the proportion of people who
strongly agree with the statement, ‘people who break the law should be
given stiffer sentences’; the working class and those with the least education
consistently favour the harshest measures against criminals. These differ-
ences are large (10–20 per cent gaps between the working class and the new
middle class and 20–30 per cent gaps between the most and least educated)
and barely change over time.
Interestingly, while the differences by occupational class are reduced when
we hold constant other characteristics of people (the same long list as before
including education), the differences by education remain almost the same.
Table 4.6 shows that occupational class still affects how people view the death
penalty, but there is now less than a ten percentage point gap between the

74
 Ideology

a) Occupational class b) Education

60% 60%

40% WC 40% Low

OMC
Medium
NMC
20% 20%

High

0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.8. Strongly support stiffer sentences by occupational class and education
Note: The figures here show the proportion of people who strongly agree that ‘people who break the law
should be given stiffer sentences’ (two period moving average). The left-hand graph disaggregates the results
by occupational class; the right-hand graph disaggregates the results by education. Three occupational
class groups are displayed: old middle class (OMC), new middle class (NMC), and working class (WC).
Three educational groups are displayed: people with degree-level education (high), people with A Level
equivalent education (medium), and people who left school at the minimum school leaving age for their
cohort (low).
Source: British Election Studies 1987–2001; British Social Attitudes Surveys 1986–2015.

new middle class and the working class. The differences on sentencing by
occupational class are almost zero and we do not show them here. By
contrast education matters for both attitudes. The differences between
those with medium and low levels of education are at least ten percentage
points for both sentencing and the death penalty, and are unchanged over
fifty years. There are even larger gaps between those with high and low levels
of education.
There is some change over time. While people with degrees are consistently
the most relaxed about punishment, this group is more authoritarian than it
used to be. The gap on support for the death penalty between the least and
most educated has dropped by about 10 percentage points since the 1960s,
although it remains at nearly 30 per cent. This is almost entirely due to the
increasing illiberalism of those with degrees. This is perhaps not surprising
since those with degrees in 2013 are a larger and more heterogeneous group
than they were in 1963. If we think of the ways that higher education may
liberalize people, then it is likely to prove less effective when higher education
is no longer such an immersive experience. As more students work and study
part-time, or live at home rather than on campus, it seems likely that these
socialization effects may weaken.13

75
The New Politics of Class

Table 4.6. Impact of occupational class and education on attitudes towards the death
penalty and stiffer criminal sentences

Support death penalty 1960s 1970s 1980s 1990s 2000s 2010s

Old middle class 70% 62% 70% 71% 72% 66%

New middle class 67% 56% 63% 66% 69% 61%
Junior middle class 74% 61% 68% 71% 72% 66%
Working class 73% 65% 71% 74% 76% 70%
New middle class – working class 7% 10% 7% 8% 7% 10%
High education (degree) 36% 34% 39% 44% 46% 44%
Medium education (A Level) 65% 59% 60% 63% 63% 62%
Low education (school leaving age) 76% 71% 76% 75% 74% 72%
High education – low education 40% 37% 37% 32% 28% 29%

Strongly support stiffer sentences 1980s 1990s 2000s 2010s

High education (degree) 12% 15% 20% 17%

Medium education (A Level) 24% 25% 28% 27%
Low education (school leaving age) 33% 34% 41% 35%
High education – low education 21% 20% 21% 19%

Note: The numbers here are predicted probabilities from logit regression models using pooled data by decade that
predict a) agreement with the reintroduction (or retention in earlier years) of the death penalty and b) strong agreement
that ‘people who break the law should be given stiffer sentences’. As well as occupational class and education, these
models include controls for housing, trade union membership, gender, age, region, agricultural employment, religion,
race, and year. The predicted class probabilities are for a white Anglican man in his forties, who is a homeowner, has
middling educational attainment and lives in the south east of England, and is not a trade union member. The predicted
education probabilities are for the same type of person in the junior middle class category. Year is set as close as possible
to the middle of each decade.
Source: British Election Studies 1963–2001; British Social Attitudes Surveys 1983–2015.

Nowt as Queer as Folk

The final set of attitudes that we are interested in relate to what might be termed
moral questions: how people think about questions to do with sex and ‘trad-
itional values’. Here the recent emphasis has also been on the role of education
(for example, Lottes and Kuriloff 1994; Kalmijn and Kraaykamp 2007). None-
theless, there is also a tradition of thinking about how occupational class
matters for these attitudes, with the presumption that middle class people are
typically more tolerant of certain groups.14 There is some cross-national evi-
dence for this when it comes to attitudes towards homosexuality (Andersen and
Fetner 2008) and we find similar patterns in Britain. When people are asked
whether they agree that ‘sexual relations between two adults of the same sex is
always wrong’, there are substantial differences by occupational class and edu-
cation. Figure 4.9 shows these differences from the early 1980s until today. The
rate and scope of change is quite striking: while a substantial majority of people
thought homosexuality was always wrong in the 1980s, this compares to only a
small minority now. Of all the attitudes that we look at in this chapter it is
probably the one for which we see the most far-reaching and most consistent
change over time.15 Just as importantly, there are clear and consistent effects of

76
 Ideology

class and education. The new middle class and those with degrees are the most
tolerant, and this remains unchanged over time.
This pattern is not entirely replicated when we look at other ‘moral’ issues.
People’s views about censorship and whether young people are respectful
enough are shaped by education, but not by occupational class. Figure 4.10
therefore just shows differences by education in support for the statements
that ‘censorship of films and magazines is necessary to uphold moral stand-
ards’ and ‘young people today don’t have enough respect for traditional
British values’.16 There are huge differences by educational attainment in
relation to whether young people are respectful enough, and sizable effects
for how education shapes attitudes towards censorship. More education
means that you are less worried about young people’s lack of respect and less
willing to endorse censorship, or at least censorship of a particular kind that
‘upholds moral standards’.
Of course, all these differences seem likely to be related not just to education
but to religiosity and age which are themselves correlated with education. If
anything is a constant of history it is older people complaining about
the youth of today. Nonetheless, Table 4.7 shows that even controlling for
religion and birth cohort, alongside the long list of other variables that we
used before, differences by education for both questions remain. These

a) Occupational class b) Education

80% 80% Low

WC

OMC Medium
60% 60%
NMC
High
40% 40%

20% 20%

0% 0%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.9. Agree homosexuality is wrong by occupational class and education

Note: The figures here show the proportion of people who agree that ‘sexual relations between two adults of the
same sex is always wrong’ (two period moving average). The left-hand graph disaggregates the results by occupa-
tional class; the right-hand graph disaggregates the results by education. Three occupational class groups are
displayed: old middle class (OMC), new middle class (NMC), and working class (WC). Three educational groups are
displayed: people with degree-level education (high), people with A Level equivalent education (medium), and
people who left school at the minimum school leaving age for their cohort (low).
Source: British Social Attitudes Surveys 1983–2015.

77
The New Politics of Class

a) Agree censorship upholds moral standards b) Agree young people have no respect
90% 90%

Low

70% Low 70%

Medium
High

50% 50%
High
Medium

30% 30%
1960 1975 1990 2005 1960 1975 1990 2005

Figure 4.10. Views on censorship and young people by education

Note: The left-hand graph here shows the proportion of people who agree that ‘censorship of films and magazines
is necessary to uphold moral standards’ (two period moving average). The right-hand graph shows the proportion
who agree that ‘young people today don’t have enough respect for traditional British values’ (two period moving
average). Three educational groups are displayed: people with degree-level education (high), people with
A Level equivalent education (medium), and people who left school at the minimum school leaving age for
their cohort (low).
Source: British Election Studies 1974–2015; British Social Attitudes Surveys 1986–2015.

differences are very constant over time and still large. There are about 20
percentage points separating those with high and low levels of education for
both the question about homosexuality and the question about young
people’s levels of respect.

Putting It All Together

What the above discussion, and associated tables and figures, illustrates is that
class mattered, and continues to matter, for how people view political and
social issues. Occupational class is a good predictor of how people view all aspects
of how the market should be constrained. Education, and to a lesser extent
occupation, shape how people view other political issues from immigration to
the death penalty. The public as a whole have changed their minds on some
of these issues: few people think homosexuality is wrong or that unions are
too powerful today. What remains constant are the differences in how
people think about these questions with regard to their own formative and
current circumstances. Someone’s place in the occupational structure and
their educational experiences shape how they see the world just as much
today as they did fifty years ago.

78
 Ideology

Table 4.7. Impact of occupational class and education on attitudes towards homosexuality,
young people, and censorship

Agree homosexuality is wrong 1980s 1990s 2000s 2010s

Old middle class 76% 58% 38% 35%

New middle class 73% 58% 39% 35%
Junior middle class 76% 60% 37% 38%
Working class 77% 64% 49% 49%
New middle class – working class 5% 7% 10% 14%
High education (degree) 61% 36% 24% 19%
Medium education (A Level) 77% 52% 31% 25%
Low education (school leaving age) 86% 69% 44% 37%
High education – low education 26% 32% 20% 18%

Agree young people have no respect 1980s 1990s 2000s 2010s

High education (degree) 63% 57% 69% 57%

Medium education (A Level) 75% 71% 78% 73%
Low education (school leaving age) 82% 78% 86% 78%
High education – low education 19% 21% 17% 22%

Agree censorship upholds moral standards 1970s 1980s 1990s 2000s 2010s

High education (degree) 69% 71% 56% 57% 60%

Medium education (A Level) 57% 75% 63% 64% 69%
Low education (school leaving age) 69% 78% 69% 72% 72%
High education – low education 7% 13% 15% 12%

Note: The numbers here are predicted probabilities from logit regression models using pooled data by decade that predict
a) agreement that ‘sexual relations between two adults of the same sex is always wrong’, b) agreement that ‘censorship of
films and magazines is necessary to uphold moral standards’, and c) agreement that ‘young people today don’t have
enough respect for traditional British values’. As well as occupational class and education, these models include controls for
housing, trade union membership, gender, age, region, agricultural employment, religion, race, and year. The predicted
class probabilities are for a white Anglican man in his forties, who is a homeowner, has middling educational attainment
and lives in the south east of England, and is not a trade union member. The predicted education probabilities are for the
same type of person in the junior middle class category. Year is set as close as possible to the middle of each decade.
Source: British Election Studies 1974–2015; British Social Attitudes Surveys 1986–2015.

This can be seen most clearly if we group some of these issues to make two
ideological dimensions on which people can be placed. This is a common strategy
in Britain (Heath et al. 1994; Evans et al. 1996), the US (Ansolabehere et al. 2008;
Fleishman 1988) and other Western countries (Grunberg and Schweisguth 1993;
Middeldorp et al. 1993) and typically assumes that political systems have divi-
sions along economic and ‘new politics’ lines (Kitschelt 1994; Flanagan 1987).
We reduce political placements to a position on economic issues and a position
on a set of moral/authoritarian issues. We are thus looking at the long-standing
economic cleavage, largely driven by occupational class, and the ‘new politics’
cleavage, largely driven by education. We exclude, for the moment, the issues of
immigration and EU integration.
Creating scales to measure people’s positions on these two dimensions is
not a straightforward task as there are few questions that actually span the

79
The New Politics of Class

entire time period consistently. For the 1960s–1980s, we measure someone’s

left–right position using the questions on trade unions and privatization and
their position on the second dimension using the question on the death
penalty. For 1986–2015 we have a much better set of measures. These are
composed of five questions for economic left–right ideology and five ques-
tions for social conservative–liberal ideology. Variations of these scales are
widely used and have been extensively validated (Heath et al. 1994; Evans
et al. 1996; Evans and Heath 1995; Sturgis 2002; Cheng et al. 2012).
For the left–right scale for the earlier period, we add up positive responses and
divide by two. The left–right score is thus the average number of people who
agreed with more privatization and that unions have too much power, and
the conservative–liberal score is simply the number of people who agreed that
the death penalty was wrong. High scores on the left–right measure mean
that someone is more right-wing, and high scores on the conservative–liberal
measure mean that someone is more liberal. We do something similar for the
later period. Here we have five questions that relate to left and right with
responses from ‘strongly agree’ to ‘strongly disagree’, and five questions that
relate to conservative and liberal with responses from ‘strongly agree’ to
‘strongly disagree’. The first five are all questions that we have looked at indi-
vidually: whether people want less redistribution, whether there is the same law
for rich and poor, whether ordinary people get a fair share of the wealth, whether
workers benefit more than owners and whether management treats employees
well. These are added together and transformed to give a 0–1 scale. 1 on this scale
represents people who said ‘strongly agree’ to all of these statements (the most
right-wing response); 0 represents people who said ‘strongly disagree’ to all these
statements (the most left-wing response). For the conservative–liberal scale we
again use five questions. These relate to: opposition to the death penalty, oppos-
ition to stiffer sentences, support for gay rights, opposition to censorship, and
support for young people. These are the same questions discussed earlier in the
chapter. Again we scale the answers, with 0 the score for people who strongly
disagreed with all the statements (the most conservative response) and 1 the
score for people who agreed with all the statements (the most liberal response).17
Given these scales we can track different groups over time in a
2-dimensional space. In order to make these easier to visualize we show
three groups that combine educational and occupational characteristics.
With apologies to Marx, these are the proletariat (people in working class
jobs with low levels of education), the intelligentsia (people in new middle
class jobs with high levels of education), and the bourgeoisie (people in old
middle class jobs with middling educational levels). These groups represent
different sections of society, one growing rapidly in the last few decades
(the intelligentsia) and one declining in recent decades (the proletariat).
The actual number of people in these precise groups may not always be

80
 Ideology

huge, but they allow us to see how education and occupation combine to
produce different kinds of attitudes.18 Figure 4.11 shows how these groups
have changed over time on the two scales; the first graph shows the 1960s–
1980s and the second the 1980s–2010s. Note that these figures come from
statistical models that hold constant all other variables that we have used
previously, for example, trade union membership and age.19
Focusing on the underlying attitudinal structure, across multiple
attitudes, throws the patterns that we have already found into starker relief.
Figure 4.11 shows distinct groups of people with distinct attitudes and these
never overlap. In every decade, the proletariat are always the most left-wing
and always the most socially conservative. The intelligentsia is always the
most socially liberal, and always more right-wing than the proletariat. The
bourgeoisie is always the most right-wing and is always less liberal than
the intelligentsia. These differences are remarkably consistent over a long
period of time. These differences are also large. For the more reliable measures
of both dimensions in the second graph, the difference between the intelli-
gentsia and the proletariat is almost a standard deviation on the conservative–
liberal scale and the difference between the bourgeoisie and the proletariat is
about three quarters of a standard deviation on the economic left–right scale.

a) 1960s to 1980s b) 1980s to 2010s

0.7
Liberal

Liberal

1960s
Intelligentsia
2010s
0.6 1980s
Intelligentsia
2010s
0.4 1960s Bourgeoisie
0.5 1980s
Bourgeoisie
1980s 2010s
Conservative

Conservative

1960s
0.2 Proletariat 1980s
1980s
Proletariat
1980s

0.0 0.3
0.2 0.4 0.6 0.3 0.5 0.7
Left Right Left Right

Figure 4.11. How different groups have moved positions on the economic left–right
and social conservative–liberal dimensions
Note: The figures here show predicted probabilities from linear regression models using pooled data by decade that
predict scores on left–right and social liberal–conservative 0–1 scales. The left-hand graph shows results from the
1960s to the 1980s; the right-hand graph shows results for the 1980s to the 2010s. Higher scores indicate more
economically right and socially liberal responses. As well as occupational class and education, these models
include controls for housing, trade union membership, gender, age, region, agricultural employment, religion,
race, sector of employment, and year. The three groups are low education and a working class occupation
(proletariat), medium education and an old middle class occupation (bourgeoisie), and high education and a
new middle class occupation (intelligentsia).
Source: British Election Studies 1963–1987; British Social Attitudes Surveys 1986–2015.

81
The New Politics of Class

Of course, this is not to say that nothing changes. Looking at the 1980s
onwards, the most noticeable change is that people have become a little more
socially liberal (about a third of a standard deviation of the scale). This is
despite holding constant religion and race. Interestingly, this move towards
a more liberal world view is least pronounced among the proletariat. Class
differences on this dimension have actually slightly increased. There is less
change in terms of left and right over the same period, although the 1990s
stand out as a slightly more left-wing period, especially with regard to the two
middle class groups. It is more difficult to talk about change in the earlier
period since we are reliant on very few questions. It does appear that people
became more right-wing in the 1980s compared to the 1960s, or at least they
became fonder of privatization and less fond of trade unions, but this altered
the differences between the three groups only marginally. People also became
less opposed to the death penalty between the 1960s and 1980s as well, a
change we already saw in Figure 4.7.20
This analysis leaves out two issues where we saw the most change: the EU and
immigration. While these are related to the second dimension, in that people
who are less socially liberal are typically more opposed to EU membership and
further immigration, they are also separate in terms of their policy implications
and the more economic mechanisms that underpin these beliefs. How did the
attitudes of the three archetypes discussed above change with regard to these
issues? Figure 4.12 shows this. Here we combine the two questions on the EU
and immigration to give a similar 0–1 scale, where 0 indicates people who
strongly disapprove of further immigration and wish to withdraw from the
EU and 1 indicates people who advocate neither of these positions. Again,
these figures come from regression models that hold constant the same wide
variety of other factors used before. It is interesting to examine the three groups
separately. The bourgeoisie’s position is fairly static. The intelligentsia becomes
steadily more pro-EU and pro-immigration over the period, with a slight reverse
in the 2010s. But it is the proletariat group which changes its views most
distinctively. The proletariat always have the most Eurosceptic and anti-
immigration views, but this disparity between them and the two middle class
groups is widest in the 1970s and particularly in the 2010s. On at least these two
issues, class differences actually appear to have widened slightly.

Conclusions

Chapters 2 and 3 have emphasized that people’s experiences and perceptions

of class in Britain have changed remarkably little over the last fifty years. It is
therefore not that surprising that the way in which class and education
influence political ideology has also remained the same. There is little

82
 Ideology

Pro EU/immigration
0.9
Intelligentsia

Bourgeoisie
0.7
Anti EU/immigration

0.5 Proletariat

0.3
1960s 1970s 1980s 1990s 2000s 2010s

Figure 4.12. How different groups have moved positions on the immigration/EU
dimension
Note: The figure here shows predicted probabilities from linear regression models using pooled data by decade that
predict scores on an EU/immigration 0–1 scale. Higher scores indicate greater support for EU integration and
immigration. As well as occupational class and education, these models include controls for housing, trade union
membership, gender, age, region, agricultural employment, religion, race, and year. The three groups are low
education and a working class occupation (proletariat), medium education and an old middle class occupation
(bourgeoisie), and high education and a new middle class occupation (intelligentsia).
Source: British Election Studies 1963–2015; British Social Attitudes Surveys 1986–2013.

evidence in this chapter of the declining importance of class. People in work-

ing class jobs still think that the market economy does not benefit them as
much as it does the middle classes, and are more supportive of public owner-
ship, trade unions, and redistribution than anyone else. Equally, education
and occupation still shape people’s views of right and wrong: middle class
people with more education are systematically the most socially liberal group
and most in favour of EU integration and mass immigration.
This lack of change in the differences between educational and occupa-
tional groups disguises one important fact, however. The differences
between the average person and different educational and occupational
groups have changed because the number of people in those different
groups has changed. When people who had a working class job and left
school with few, if any, qualifications were the majority of the population,
the average position of British society was similar to people in this majority.
As this group, our stylized proletariat, shrinks, it starts to look rather differ-
ent to the average person. If we assume no change over time within groups
then the average person is constantly moving away from the proletariat over
time. Assuming there is no change within groups for the high-quality scale
measures of left–right and conservative–liberal attitudes, the proletariat are
about one quarter of a standard deviation more left-wing and about a tenth

83
The New Politics of Class

of a standard deviation more socially conservative than the average person

in the 1960s. By the 2010s the proletariat group are a half of a standard
deviation more left-wing and a third of a standard deviation more socially
conservative than the average person. Differences between the groups may
have remained rather static, but the position of the working class group has
become more dissimilar from the typical voter, because the typical voter is
now more educated and more middle class. Why does this matter? It matters
because rational politicians, especially when faced with a first past the post
electoral system, should change their policies to match the average voter. As
the average voter moves away from the working class, so should the parties.
As we see in Chapter 6, this is broadly what has happened.
It is also important to make clear that there have been some changes. In
particular, there has been modest convergence between occupational classes
on some of the economic left–right issues since the 1980s. Nonetheless, this
essentially entails a return to the status quo of the 1960s, rather than a long
run convergence of views between classes. There have also been modest
increases in class and education disparities in terms of attitudes towards
immigration and the EU in recent years. In that sense, aspects of underlying
class conflict appear to be greater now than they were previously. Of course,
for that to be politically salient, parties need to do something about it, and
voters need to know about that class conflict and party responses to it. For
most people, a key source of information about society and politics is the
media. And so in Chapter 5 we look at how media coverage of class and
politics has changed, before in Chapter 6 turning to how parties themselves
have changed.

Notes

1. Some have argued that self-interest does not drive political attitudes. Rather, what
matters is people’s ‘conditioning in their pre-adult years, with little calculation of
the future costs and benefits of these attitudes’ (Sears et al. 1980, p.671). Although
this work on ‘symbolic politics’ by Sears has been very influential (Sears et al. 1980;
Sears 1975, 1983, 1993), it is in practice very difficult to separate out conditioning
and socialization factors from social characteristics. It is also worth noting that
Sears finds that education and income (a weaker measure of class than occupation
anyway) are strong predictors of attitudes, even holding constant what he terms
symbolic attitudes, such as party identification.
2. For example, in Bartels’ 2008 book Unequal Democracy, he provides various rich
sources of evidence about how income inequality affects, and is affected by, politics
in the US, but he is only able to track consistently one attitudinal question about
economic issues, specifically whether ‘the government should see to it that every
person has a job’. Even that is only for 1972–2004, and he can only compare broad

84
 Ideology

income groups over time. Those looking at Britain have typically had access to a
wider range of repeated questions, and better measures of class, but what work
there is over time in Britain is now also rather out of date. Heath et al. (1991) look
at both economic and non-economic values for the four British general election
studies from 1974 –1987, but given this time period is barely more than a decade, it
seems insufficient to make lasting judgements about long-term change. While
Evans (1993a) covers a longer time period (1964–1987), his data ends almost thirty
years ago and is a comparison of only two cross-sectional surveys.
3. The question is worded slightly differently depending on the survey, but essentially
asks whether people would prefer ‘more nationalization of companies by govern-
ment’, ‘more privatization of companies by government’, or for things to stay the
same. In some years people are also asked whether they would like a lot more or a
little more nationalization/privatization. We compare here people who say they
want more privatization to everyone else.
4. These control variables are measured slightly differently for the 1963–2005 period
that covers the privatization measure and 1986–2015 period that covers the ques-
tion about who benefits. In general, the measures that cover a longer time period
are somewhat weaker, and those from the mid-1980s onwards are somewhat
stronger. For example, the regional measure for the full 1963–2015 data has six
categories, whereas the regional measure for the period that starts in 1986 has
eleven categories. Full details of the control measures used are included in Appen-
dix A4.1.
5. Sector of work cannot be measured before the late 1970s and so is not included in
the models that predict privatization attitudes. Interestingly, if we model attitudes
towards privatization from the 1980s onwards we see that while employment
sector does shape people’s attitudes about the merits of the free market, it has less
effect than occupational class. This rather puts the lie to the claim made in the
1980s that sectoral divisions would become the most important source of conflict
in British politics (Dunleavy 1980, 1986; Dunleavy and Husbands 1985).
6. The exact question is almost identical from year to year and asks whether people
agree that ‘income and wealth should be redistributed to ordinary working people’
with ‘strongly agree’ to ‘strongly disagree’ as possible responses. We show the
proportion of people who disagree with the statement (either somewhat or
strongly).
7. The exact wording of the question varies slightly over the time period. The BES
surveys, apart from 1987, give people the options of answering that trade unions
have too much or too little power (with a ‘don’t know’ option as well), and the BSA
surveys allow people to say that the power that unions wield is far too little, too
little, about right, too much, or far too much. We combine these to just contrast
people who think unions have too much power compared to any other answer.
8. For example, Tilley and Heath (2007) show that occupational class and education,
along with religion and generation, are two of the best predictors of the strength of
pride in Britishness over the 1984–2003 period (see also Dowds and Young 1996).
9. Since most immigration to Britain is low-skill, this has a much more negative effect
on people in working class jobs with low levels of education, and since attachment

85
The New Politics of Class

to national culture is generally weakened by higher education, cultural threats are

perceived as lowest among the most highly educated (Schneider 2008; McLaren
and Johnson 2007; Sides and Citrin 2007). There is also a large, and ever-growing,
literature examining these arguments in relation to radical right party support
across Europe (for example, Oesch 2008; Ivarsflaten 2008; Lucassen and Lubbers
2012). Two recent books deal with immigration policy and views in Britain in
detail: Revolt on the Right by Ford and Goodwin (2014) talks about the factors that
affect UKIP support and Immigration and Perceptions by Lauren McLaren (2015)
focuses on how views about immigration link to national identity. We return to
this debate in Chapter 8.
10. Most questions focus on whether immigration rates should be increased or
decreased, and we look at the proportion of people who think immigration should
be decreased a lot/very strongly agree it should be decreased. For the period from
1983–1996 we also used a question that asks whether people think controls on
immigrants’ relatives joining them should be stricter and for a couple of years in
the early 2000s we used a question that asks whether people agree that immigrants
take jobs from British people. Some of these questions are asked in the same survey
in certain years and the correlation between them is typically high, suggesting that
they tap into a general underlying positive or negative view of immigration.
11. Some of the response options given to people vary over time. For example, in 1974
people were asked whether they thought Britain should ‘stay in the EU’, ‘stay and
change the EU’, ‘change the EU or leave’, or simply ‘leave’, whereas in the 1980s,
the only options given to people were to ‘stay’ or ‘leave’. In later years, people were
also given options of a ‘single European government’ or ‘increasing EU powers’.
What is common is the option to leave the EU, and that is what we report here.
12. Before 1970, the question on the death penalty refers to whether the death penalty
should be kept (the death penalty for murder was formally abolished at the end of
1969), from 1979–1985 whether it should be re-introduced, and from 1986–2015
whether ‘for some crimes, the death penalty is the most appropriate sentence’.
13. It is difficult to get good quality figures on many of these changes, but even
between 1995 and 2005 the number of students living with their parents nearly
doubled from 12 to 20 per cent (Universities UK 2006).
14. It is important to note that increased tolerance of all groups for those with more
education or in a higher occupational social class is not the expectation. It is not
tolerance in and of itself that education or a professional job brings, but rather
support and opposition to certain types of people. For example, more highly
educated people are often less tolerant of racist or religious groups, and sometimes
more willing to curtail their freedom of speech, than people with lower levels of
education (Davis 1975; McCutcheon 1985).
15. See Turner (2015, pp.215–30) for an account of changing attitudes towards sexu-
ality in 1990s Britain, which puts those changes in a political and cultural context.
Perhaps most revealing in terms of attitude change is the leader column from The
Sun in 1994 about gay men serving in the military that Turner cites: ‘the British
soldier needs to worry about the enemy ahead. Not some queer behind him’
(p.217). It is impossible to imagine The Sun running this editorial today.

86
 Ideology

16. For 1974 and 1979 we use a slightly different question that asks whether the ‘right
to show nudity and sex in films and magazines’ has gone too far. This captures
something similar, and is correlated at 0.5 with the standard question when both
are asked in 1992. The direct mention of ‘nudity and sex’ perhaps explains the
difference in how this question is answered by degree holders in the 1970s, which
could be picking up on feminist, rather than socially conservative, objections to
pornography.
17. For the whole dataset, the economic left–right five-item scale has an alpha score of
0.82 which remains very consistent over time. The conservative–liberal five-item
scale has an, again unchanging, alpha score of 0.62.
18. These are relatively common pairings of the education and full seven-category
occupational class variables. A clear majority of people in working class occupa-
tions have the lowest level of education in every decade. Equally people in the new
middle class occupational group are consistently the most likely to have a degree,
with nearly half having some sort of higher education in every decade apart from
the 1960s. The old middle class occupational group is unsurprisingly more mixed,
but this is the group in which someone is most likely to have intermediate
qualifications.
19. We run separate linear regression models for each decade and for each scale,
including survey year as a series of dummy variables along with the control
variables of region, age group, trade union membership, sector of work, housing
tenure, religion, agricultural employment, sex, and ethnicity. The points for the
proletariat are the actual values for that group as a whole in each decade, and the
points for the other two groups are thus the scores they would have if they were
identical to the proletariat group in every way apart from their education and
occupational class.
20. This shows the problems of using a single item to capture change over time on this
broad dimension of social liberalism as it seems likely that people were as, or more,
socially liberal in the 1980s than they were in the 1960s, particularly with regard to
issues of sexual morality. For example, Tilley (2005) shows that from 1974 to 1997
people in Britain tended to become more socially liberal with regard to equal
opportunities (for women and ethnic minorities), the availability of abortion,
and censorship.

87
The New Politics of Class

Appendix to Chapter 4

Table A4.1. Measurement of independent variables used in models

1963–2015 period 1974–2015 period

Occupational Old middle class; New middle class; Junior As 1963–2015.

class middle class; Personal service; Own
account workers; Foremen; Working class.
Education Degree; Some higher education; A Level; Degree; Teacher or nursing qualification;
Post-minimum school education; Left A Level or equivalent; O Level or equivalent;
school at school leaving age. CSE or equivalent; Apprenticeship or
secretarial training; No qualifications.
Housing Council housing or housing association; Council housing or housing association;
Private rental or other; Owner-occupier. Private rental; Outright owner; Owner with
mortgage; Other.
Region Scotland; Wales; North, Midlands and East; Scotland; Wales; North East; North West;
South West; London and South East. Yorkshire; East Midlands; West Midlands;
South West; East; London; South East.
Race Non-white; White. Black; Asian—Indian sub-continent;
Asian—Chinese; Mixed or other; White.
Employment Not available. Private sector; Public or charity sector.
sector
Sex Men; Women. As 1963–2015.
Age group 18–29; 30–39; 40–49; 50–59; As 1963–2015.
60–69; 70+.
Religion Anglican; Presbyterian; Methodist; Baptist; As 1963–2015.
Other non-conformist; Catholic; Other
Christian; Non-Christian; No religion.
Trade union Current member of trade union or staff As 1963–2015.
membership association; not member.
Agricultural Employed in agriculture; not in agriculture. As 1963–2015.
employment

Note: Occupational class is not available for the 2010 BES. Education is not available for the 1983 and 1984 BSA. Housing
tenure is not available in the 2001 BES. The fuller race measure is not available for the two 1974 BESs. Public sector
employment is not available in the October 1974 and 1979 BESs or the 1983 BSA. Religion is not available for the
February 1974 BES. Agricultural employment is not available for the 2010 BES.

88
Part II
Political Change
5

The ‘Papers’

Chapters 2–4 have shown how class, whether measured by occupation or

education, affects how people live their lives and how they view different
political issues. Part II of the book focuses on how the political world creates
structures within which class is, or is not, linked to party choices. Chapter 6
looks at what politicians say, where they come from, and what policies they
offer voters. But there is another important way in which structure is imposed
on people’s choices. This is via the information that voters consume about
politics: information about where people are located in society, how those
locations map on to political parties, whether class conflict is real and, if so,
whether it is the basis of party conflict. This information comes primarily from
the mass media and this chapter is an attempt to evaluate how the media
tells the story of class in Britain and how that has changed. We show that
newspapers have moved from a post-war portrayal of the working class as ‘the
people’ with class stratification simply part of the scenery, to the newspapers
of today that rarely mention class, stratification, or conflict. The structuring of
the political world through the class-tinted lenses that newspapers used to
provide no longer exists.
We focus on the British press in this chapter, and when one thinks of British
newspapers, one does not think of subtlety or neutrality. Indeed a robustly
partisan press has been a feature of British society since the early nineteenth
century. Dickens’ account of Eatanswill in The Pickwick Papers is perhaps the
most eloquent description of how press and politics intertwine: ‘it was essen-
tially and indispensably necessary that each of these powerful parties should
have its chosen organ and representative: and, accordingly, there were two
newspapers in the town—the Eatanswill Gazette and the Eatanswill Independ-
ent; the former advocating Blue principles, and the latter conducted on
ground decidedly Buff ’ (Dickens 1837, p.166). Of course ‘Buff ’ and ‘Blue’
newspapers are common to many countries, but Britain is also remarkable in
having such a small number of national newspapers that, until recently, were
read by such a high proportion of the population. In the 1960s, three quarters
The New Politics of Class

of adults regularly read a daily national newspaper, and as recently as 2003

over half of people were still regular readers. This is high by international
standards. Writing in the 1960s, Butler and Stokes (1969, p.282) note that in
1964 newspapers ‘reached over 80 per cent of households—a higher propor-
tion than in any other country’ and that ‘Britain has a smaller total of
independent morning newspapers than any comparable nation’. As we will
see the first of these generalizations is no longer true: newspaper circulation has
been rapidly falling for the last twenty years. But newspapers still remain few in
number and national in type. Moreover, the degree of partisan affiliation is
unusual. The Daily Telegraph, the Daily Mail, and the Daily Express (apart from
2015) have endorsed the Conservatives in every single post-war election. On
the other side, the Daily Mirror has endorsed Labour in every one of those
elections. The Guardian has also remained manifestly left-wing over the last
fifty years, mainly endorsing the Liberals or Labour depending on the election
(Butler and Butler 2000; see Wring and Deacon (2010) or Thomas (1998) for
detailed accounts of recent press partisanship). Some newspapers do switch.
The Sun (and in its earlier incarnation as the Herald) supported Labour until the
mid-1970s when it switched to the Conservatives; then at the 1997 election it
declared itself to be ‘backing Blair’.1 Regardless of these switches (and in fact the
Sun switched back again in 2010 to the Conservatives) newspapers do not feel
constrained by their past in lavishing praise on the party that they currently
support. Using data from the British Election Studies and British Social Attitudes
surveys, Figure 5.1 shows the percentage of readers of the main newspapers that
support the Conservatives over time.2
Knowing what newspaper people read is a good guide to their party support.
Guardian readers rarely vote Conservative, while Telegraph readers almost
always do. Well over half of Mail readers support the Conservatives, yet fewer
than a fifth of Mirror readers do. These are, with the exception of the Sun, very
long-standing relationships between partisans and partisan newspapers.
These extreme, or at least peculiar, aspects of the British press mean that
we may well expect certain types of media effects to operate on readers more
or less strongly. Typically, when we think of how the media affects people’s
views, we think of both priming and persuasion. Priming, or agenda setting, is
the way in which a media source focuses on some issues at the expense of others.
Persuasion, by contrast, is the way in which newspapers influence people,
typically by framing issues in certain ways. Naturally these overlap. After all,
the Mirror rarely mentions good economic news if the Conservatives are in
government. What matters is the overall impact of these effects, which is
arguably to reinforce people’s partisan opinions.3 Certainly the first serious
attempts to analyse the effects of the mass media in the US, in the decades
following the Second World War, claimed that this reinforcement effect was
dominant (Lazarsfeld et al. 1948; Berelson et al. 1954; Klapper 1960; Campbell

92
 The ‘Papers’

a) Broadsheets b) Tabloid and mid-market

80% 80%
Telegraph
Mail
60% 60%
Express
Times
40% 40%
Sun
Guardian
20% 20%

Mirror

0% 0%
1960s 2010s 1960s 2010s

Figure 5.1. Proportion of readership supporting the Conservatives

Note: The figures here show the proportion of people who support the Conservatives by regular newspaper
readership. The left-hand graph shows values for broadsheet newspapers; the right-hand graph shows values for
tabloid and mid-market newspapers. Regular readership is taken as a positive response to the questions ‘do you
normally read any daily morning newspaper at least three times a week?’or ‘do you read a morning newspaper
regularly?’.
Source: British Election Studies 1963–2010; British Social Attitudes Surveys 1983–2015.

et al. 1960). This view was largely shared by Butler and Stokes (1969) writing
about 1960s Britain and more recent work has also advocated the reinforce-
ment model (Ansolabehere and Iyengar 1995; Finkel 1993; Newton 2006)
or an ‘enlightening of preferences’ during election campaigns by the
media that reinforces people’s opinions (Gelman and King 1993; Andersen
et al. 2005).
Nonetheless, there is continuing debate about the degree to which the
media changes people’s opinions. This is largely due to two major difficulties
with establishing media effects. First, it is hard to know whether people
change their opinions because of their newspaper, or change their newspaper
because of their opinions. Both result in people having the same view as their
newspaper, but only the former means that the media persuades people.
Second, it is difficult to know whether readers follow their newspaper’s
views, or whether the papers follow their readers. Again, it is only the former
that implies that the media have persuasive power.4
These are difficult problems to overcome. It means that much of the litera-
ture wants to answer the question: ‘are the mass media politically powerful, in
the sense that they have an effect of their own on political attitudes and
behavior, or is their role a minimal one, restricted to reflecting or reinforcing
attitudes?’ (Newton 1991, p.51). We do not. We are not interested in whether
newspapers directly persuade people to switch parties, or indeed to adopt new

93
The New Politics of Class

policy stances. We are interested in how newspapers, as a medium, portray a

particular aspect of society, class, and its relationship to politics. Part of this is
inevitably about partisan reinforcement, or even persuasion, but it is, more
importantly, about how people’s broader views of politics and society are
subtly shaped by the messages that they see on a daily basis.

Who Reads What?

Different newspapers cheer on different parties and are also read by different
kinds of people. Importantly from our point of view, different newspapers are
read by people with varying levels of education and different sorts of occupa-
tion. These differences are very persistent over time. This is most obvious if we
divide the main newspapers into three categories: broadsheet (the Times, the
Telegraph, the Guardian, and the Independent), mid-market (the Mail and
Express), and tabloid (the Sun, the Mirror, and the Star). We stick with the three
stereotypical groups discussed in Chapter 4 to illustrate these differences: the
intelligentsia (people with a new middle class job and high education),
the bourgeoisie (people with an old middle class job and medium education),
and the proletariat (people with working class jobs and low education). Using
combined BES and BSA data back to 1963, Figure 5.2 shows how readership of
these three types of newspaper are shaped by occupational class and education
over time. These numbers come from statistical models that hold constant a
number of other factors that affect readership, most notably age and sex.5
What Figure 5.2 illustrates is both the way readership has changed, and
perhaps more importantly how readership is a persistent product of education
and occupation. If you are in the intelligentsia then you either read a broad-
sheet or nothing. Almost no one in this group reads a tabloid and very few
read the Express or Mail. Before the 2000s over half this group read a broad-
sheet newspaper. The proletariat group looks equally distinctive. Before 2000
around half or more of people in this group read a tabloid and almost none
read a broadsheet. There has been a notable change, however, since the turn of
the century. Readership of tabloids by the working class has almost halved and
readership of broadsheets by the intelligentsia has more than halved.
Classes have become more similar in terms of readership, not because news-
papers are drawing a more diverse readership, but rather because fewer people
are reading a newspaper. In 2015 only 31 per cent of people regularly read a
national newspaper. In 1964 it was over 80 per cent.6 Figure 5.3 shows how
aggregate newspaper readership has changed over time. Interestingly, here we
see much bigger falls in tabloid readership and much smaller falls in broadsheet
readership. Why? Because the people who generally read tabloids (people with
lower levels of education and in working class jobs) have declined in number,

94
 The ‘Papers’

a) Intelligentsia b) Bourgeoisie c) Proletariat

80% 80% 80%

Broadsheet
60% 60% 60% Tabloid
Broadsheet

40% 40% 40%

Mid-market
Mid-market
20% Mid-market 20% 20%
Tabloid
Tabloid Broadsheet
0% 0% 0%
1960s 2010s 1960s 2010s 1960s 2010s

Figure 5.2. Proportion of people in different groups reading different types of

newspapers
Note: The figures here show predicted probabilities from multinomial logit regression models using pooled data by
decade that predict type of newspaper read. Newspapers are grouped into broadsheets (Times, Telegraph, Guardian,
and Independent); mid-market (Express and Mail), and tabloid (Mirror, Sun, and Star). The left-hand graph shows
results for the intelligentsia (high education and a new middle-class occupation); the middle graph shows results
for the bourgeoisie (medium education and an old middle class occupation); the right-hand graph shows results for
the proletariat (low education and working class occupation). As well as occupational class and education, these
models include controls for sex, age, and region. Regular readership is taken as a positive response to the questions
‘do you normally read any daily morning newspaper at least three times a week?’or ‘do you read a morning
newspaper regularly?’.
Source: British Election Studies 1963–2010; British Social Attitudes Surveys 1983–2015.

whereas there has been a growth in the number of people (the more educated
and those in professional jobs) who generally read broadsheets.
What Figure 5.3 shows is therefore not just the decline of tabloid readership but
the decline of tabloid readers. This has implications for how we should interpret
Figure 5.2. At first glance, it might appear that there has been some decrease in the
class differences in newspaper readership. But that is not really the case. People in
the proletariat group are consistently about twenty times more likely to read a
tabloid than a broadsheet regardless of the year. Equally, the intelligentsia are
about eight times more likely to read a broadsheet than a tabloid in the 2010s,
exactly the same ratio as we see in the 1960s. Class differences in press consump-
tion stay constant. It is consumption of the mass press that has fallen.7

What the Papers Say

The fact that partisan class-based newspapers are less widely read is important
and we will return to it later. Equally important is what those papers say to
their readers about social class. We want to know how papers talk about class

95
The New Politics of Class

Tabloid
40%

Mid-market
20%

Broadsheet

0%
1960 1970 1980 1990 2000 2010

Figure 5.3. Proportion of people reading different types of newspapers

Note: The figures here show the proportion of people regularly reading a newspaper (two period moving average).
Regular readership is taken as a positive response to the questions ‘do you normally read any daily morning
newspaper at least three times a week?’or ‘do you read a morning newspaper regularly?’. Newspapers are grouped
into broadsheets (Times, Telegraph, Guardian, and Independent); mid-market (Express and Mail), and tabloid (Mirror,
Sun, and Star).
Source: British Election Studies 1963–2010; British Social Attitudes Surveys 1983–2015.

and how, if at all, they relate class to politics. Moreover, we want to know how
this is shaped by the nature of the readership and how this has changed over
time. These are not straightforward questions. And perhaps this is why there
has been so little investigation of how the media reinforce social cleavages via
their presentation of class politics. The way we answer them is twofold:
quantitative and qualitative. The former approach involves coding editorials
from three newspapers during every election campaign since 1945 for men-
tions of class and class politics. The latter involves grouping these editorials
over time to allow us to talk about the narratives that underpin discussion of
class. These narratives vary not just by time but by newspaper.
Which newspapers? We look at three that broadly represent three of the
main occupational groups that we consistently examined in previous chap-
ters: the Mirror, a Labour-supporting tabloid with a largely working class
readership; the Guardian, a Liberal/Labour-supporting broadsheet with a
largely new middle class readership; and the Times, a Conservative/Liberal-
supporting broadsheet with one of the highest old middle class readerships.8
In some ways, this is not an ideal spread of the newspaper market. We include
no mid-market paper and no tabloid on the right. Nonetheless, the choice of
these three newspapers allows us to look at newspapers with relatively consist-
ent, and different, readership class bases, relatively consistent partisan stances,
and newspapers that are not in vertiginous decline (such as the Express).

96
 The ‘Papers’

But which stories? We look here at editorials during the month-long formal
election campaigns.9 We pick editorials because we want to analyse the ‘voice’
of the newspaper. This allows us to hone in on what kind of readers are talked
about in editorials and how politics and class structures are discussed with
reference to those readers. We pick election campaigns for the practical reason
of making the exercise more manageable, but also because this is a period in
which we should expect linkages to be made between class groups and politics.
By looking at the papers four weeks before the election, and on the election
day itself, we have twenty-five days of newspapers for each campaign. We do
not analyse the Sunday newspapers, which are historically distinct from the
dailies. With nineteen elections since the war, that means that there are 475
possible editorials to look at from each newspaper. In practice, there are fewer
editorials than this, as newspapers were not printed on some days and not
every paper contained an editorial10, meaning that we have a total of 1,292
editorials to analyse: 390 from the Mirror, 456 from the Guardian, and 446
from the Times. Not all editorials have domestic political content either. In
fact, about 8 per cent of the editorials, relatively evenly spread across the three
newspapers, do not mention domestic politics. This reduces our sample fur-
ther to 1,190 editorials: 348 in the Mirror, 435 in the Guardian, and 407 in the
Times. It is these near 1,200 editorials that we analyse here.
How do we measure whether editorials talk about class? We separate out
three main ideas. First, we are interested in whether there is any discussion of
social class. Is there recognition of class divisions, and is there talk of people as
working or middle class? Second, we are interested in class conflict. Do edi-
torials talk about the classes as in conflict with one another? Third, we want to
know how politics and class intersect and specifically whether particular
classes are associated with particular parties.
These three areas allow us to see quantitatively how discussion of class in
different newspapers has changed over time in different ways. In this section
we will use the combined dataset of all the editorials from all three news-
papers, since this allows us to assess change over time most accurately. Later in
this chapter we will discuss in more detail how different newspapers portray
class. Here we are more interested in describing the broader media milieu that
exists at different points in time.
We start this description with class stratification. Are class differences an
accepted part of the world that the newspapers are describing? The simplest
possible measure here is to look at how many mentions of class there are in
editorials over the period. We make no distinctions here about the way in
which class is mentioned, or which class is mentioned, simply that it forms
part of the editorial discussion in that newspaper on that day. Figure 5.4 shows
exactly this.

97
The New Politics of Class

Here we present the percentage of editorials for all newspapers combined

that mention class stratification, class conflict, or either the middle or working
class.11 As we might expect, there is some overlap here: editorials that talk
about class stratification or conflict also tend to mention specific classes.12
What Figure 5.4 reveals is both the sheer number of mentions of class, espe-
cially before the 1980s, but also the clear change that has happened. Around a
quarter of all editorials mention class in some form or another, but this has
decreased substantially over time. Before 1997 over a third of editorials men-
tion class, but barely 15 per cent mention it between 1997 and 2015. While
newspapers still talk about class, they talk about it far less than they did fifty
years ago.
Figure 5.5 gives more detail on this shift, showing how mentions of class
stratification have changed and how mentions of the different class groups
have changed. The former is about an acceptance of class divisions that
structure society. Before 1992 over a fifth of all editorials mention this. To
put it another way, someone reading a paper every day would come across
editorial discussion of this idea at least once a week. From 1997 this falls to
fewer than 10 per cent. During the last five election campaigns someone
would need to read every editorial for over a fortnight before class division
reared its head.
Even more illuminating is specific mention of the different classes. The
working class was regularly mentioned as a group for the twenty years after
the war Second World War: a third of all editorials specifically mentioned the
working class during the 1955 election campaign. This fell in the mid-1960s,
and then fell again in the late 1980s and it has since remained at a fairly low

60%

40%
1945–1992
Average = 33%

1955
20%
1997–2015
Average = 15%

0%
1945 1965 1985 2005

Figure 5.4. Proportion of editorials mentioning class in any form

Note: The figure here shows the proportion of editorials that mention social class in any form for the month before
each general election.
Source: 1945–2015 coded newspaper editorials (Mirror, Times, and Guardian). N = 1190.

98
 The ‘Papers’

a) Class stratification b) Different classes

30% 30% Middle and

upper class

20% 20%

10% 10% Working

class

0% 0%
1945 1965 1985 2005 1945 1965 1985 2005

Figure 5.5. Proportion of editorials mentioning class in specific ways

Note: The left-hand graph here shows the proportion of editorials that mention class stratification for the month
before each general election (three period moving average). The right-hand graph shows the proportion of
editorials that mention different social classes directly for the month before each general election (three period
moving average).
Source: 1945–2015 coded newspaper editorials (Mirror, Times, and Guardian). N = 1190.

level. At the last five elections, fewer than one in twenty editorials mentioned
the working class. Discussion of the middle and upper classes13 has followed
an opposite trend. Fewer than 15 per cent of editorials mention them before
the 1970s, but almost a quarter of editorials contain a reference to the middle
or upper classes after the mid-1980s. Until the 1970s newspapers talked more
about the working class than other classes. After the 1980s newspapers talked
more about the other classes than they did the working class.
The nature of newspaper discussion about class has changed. Fewer editor-
ials mention class stratification and the working class. The precise timing of
these changes is more difficult to assess, but the direction of change is clear.
What about class conflict? Figure 5.6 shows how mentions of class conflict
and also broader discussion of equality and ‘fairness’ have changed over
time.14 There is a rather different pattern here. Rather than decline, we see a
rise and fall. Conflictual class relations are rarely discussed in the earliest
period, a time when class stratification was nonetheless regularly talked
about, but become more salient during the 1960s, until dramatically falling
again in 1992 to near zero levels by the 2000s. There is a similar, though more
pronounced, pattern for the discussion of equality and fairness. Here we see a
steady increase from 1945 onwards, sharply peaking during the two 1974
elections in which a third of editorials mention fairness, and then gradually
falling to very low levels by the 2000s.

99
The New Politics of Class

a) Class conflict b) Social equality

15%

30%

10%
20%

5%
10%

0% 0%
1945 1965 1985 2005 1945 1965 1985 2005

Figure 5.6. Proportion of editorials mentioning class conflict and social equality
Note: The left-hand graph here shows the proportion of editorials that mention class conflict for the month before
each general election (three period moving average). The right-hand graph shows the proportion of editorials that
mention social equality or inequality for the month before each general election (three period moving average).
Source: 1945–2015 coded newspaper editorials (Mirror, Times, and Guardian). N = 1190.

The changing way in which class features in newspapers fits with what we
might expect given the changing class structure of the country. The working
class features less and the middle class more as the former declines in number
and the latter increases in number. Equally, the reality of class stratification
features less heavily as the people who lose out from that stratification are less
numerically dominant in society. At the same time, class conflict and ‘fairness’
are less serious issues when most people are working class (and political parties
are catering for that) in the 1950s, but both become more important as the
numbers of people in the middle and working class groups become more
evenly balanced. As the working class then shrinks further from the 1990s
onwards, these issues recede again from editorial notice.15
Yet there is some evidence that this trend has reversed a little in recent
years. In particular, the coverage of the 2015 election campaign was more
focused on issues of class than it had been for over twenty years. It is notable
that this is particularly the case for the newspaper which caters for working
class readers. The Times and the Guardian had practically identical coverage
of class in 2015 as they did in 2010 and 2005, but the Mirror mentioned
class in one form or another in nearly 30 per cent of editorials in 2015,
compared to only a few per cent in 2010 and 2005. These small reversals
should not overshadow the broader trend, however. The landscape of infor-
mation about social class has changed as the numbers of people in those
classes changed.

100
 The ‘Papers’

There is a final mechanism that shapes how we see class. That is how the
parties talk about class. After all, the media is often reporting on what parties
say and do, especially during election campaigns. Although we will investigate
this in much more detail in Chapter 6, which focuses on how party policy,
party rhetoric, and party personnel have changed, and how this has affected
public opinion about parties, it is useful to see the way in which the parties are
portrayed by the media. This is particularly true when thinking about the
election campaign period. Is it the case that the newspapers talk about parties
in class terms? The short answer is yes, but not very much.
Table 5.1 shows the numbers of direct associations between class groups for
three grouped time periods: the working class consensus period from 1945 to
1970, the breakdown of that working class consensus from 1974 to 1992, and
the middle class consensus period from 1997 onwards. Only 2 per cent of
editorials directly link a political party with a social class.16 It is nonetheless
telling how even these small numbers break down over time. About one
editorial per campaign before 1992 either talked about Labour as the party of
the working class or the Conservatives as the party of the middle class. This
was particularly the case in the Mirror which articulated party politics in class
terms at around double the rate of the two broadsheets. In the five campaigns
from 1997 onwards, there are but three mentions of class and party in any
newspaper, and two of those actually associate Labour with the middle class,
not the working class.
Class and party are rarely linked explicitly in newspaper editorials, but when
they are, the pattern appears driven by the party changes that we discuss in
Chapter 6. We find the same kind of pattern when we look at how specific

Table 5.1. Proportion of editorials directly associating social class with specific parties or
specific policies

Working class Breakdown of Middle class

consensus consensus consensus
(1945–1970) (1974–1992) (1997–2015)

Parties and class

Labour and working class 2.2% 1.6% –
Labour and middle class – – 0.6%
Conservatives and middle class 0.7% 1.6% 0.3%

Policy and class

Welfare policy 1.1% 1.3% 0.3%
Employment policy 1.3% 0.5% –
Other policy 1.5% 0.3% –
Total 3.9% 2.2% 0.3%

Note: The numbers here show the proportion of editorials that link a) political parties and classes and b) party policies and
classes for the month before each general election from 1945 to 2015.
Source: 1945–2015 coded newspaper editorials (Mirror, Times, and Guardian). N = 1190.

101
The New Politics of Class

policies are linked to class. Table 5.1 also shows when class is invoked in
discussion of policy. We divide this into three areas: welfare policy, employ-
ment policy, and other areas of policy (which include the NHS and educa-
tion). Again, there is discussion of class and policy, albeit at low levels, before
1997; afterwards there is essentially none.
We can take three important points away from this discussion. First, class
forms a regular, but declining, part of editorial discussion in newspapers. The
reality of class stratification, and the existence of classes, is a common feature
of editorials in the 1950s. It is much less common by the 2000s. That decline is
even more striking when we look at how often the working class is mentioned.
Conversely, as the working class is talked about less, the middle class garners
more attention. The second point is that editorials rarely talked about class
and social conflict in the 1940s and 1950s, or from the 1990s onwards. It is the
intervening period (the late 1960s, 1970s, and 1980s) in which we see editors
deploying conflict as a motif. Finally, while it is rare for newspapers to directly
associate policy and parties with classes at any point, they almost never
associate classes with parties or policies today. Newspapers may have rein-
forced a view of Labour as a working class party with working class policies,
and the Conservatives as a middle class party with middle class policies, from
the war up to the 1990s, but they no longer do so.
This analysis is helpful, but partial. It ignores any detail of the narrative and
framing that newspapers use when talking about class and it ignores differ-
ences between the newspapers. We know that different newspapers are read
by different classes and different party supporters; how does this affect the
way in which editors shape their editorials?

The Framing of Class

In this next part of the chapter we provide a more in-depth analysis of the
newspaper editorials, focusing on themes that emerge during particular
election campaigns for particular newspapers. What we are interested in
here is not the number of mentions of class, but the specific narrative that
editorials use when talking about class. How do the newspapers frame issues
around class, and how does this vary over time? We argue here that there
are three distinct periods which map on to different types of class narratives
from all three newspapers. These are the same time periods that we used
above to look at how parties and classes are associated in the newspapers.
First, we have the working class consensus between 1945 and 1970 in which
the working class was clearly the majority group in society; then the break-
down of that dominance after 1970 and up to 1992, which meant an

102
 The ‘Papers’

intensification of class conflict as the middle class became the majority

group; and finally a middle class consensus era from 1997 onwards in
which the middle class was the dominant group. Obviously, the framing
of class issues differs somewhat from newspaper to newspaper, and espe-
cially between the broadsheet and tabloid press, but interestingly these
three periods see a similar set of frames used by each newspaper, albeit
coloured by their readership and partisan view. We thus organize this
section around these three time periods.

Working Class Consensus (1945–1970)

The most important facet of coverage up to 1970, which is especially obvious
in the 1945 and the 1950s elections, is the way in which class is simply part of
the background. Class stratification is a fact of life, and the average person is
part of the working class. This latter point is especially obvious in the Mirror.
The editorials in the Mirror are quite clearly addressing a relatively homogen-
ous working class group that is treated as synonymous with most people in
Britain. We repeatedly see references to ‘the people of Britain’ and social
policies that benefit working class people are labelled as ensuring ‘a better
life for the men, women and children of this country’ (24 May 1955). This is
explicitly related to politics as well, ‘the tradition of the Labour party is
forward with the people. All the people’ (5 May 1959). Similarly, the Mirror
asks on the day of the 1970 election:

Who stands for us? For all of us, the people of this great island. Not just . . . for the
Fortunate Few, born to great privilege . . . It is Labour, the party led by gifted men
and women who came from Us. (18 June 1970)

The Mirror is speaking to the working class, when the working class made up
the majority of society. This is combined with a partisan appeal, in that
Labour is portrayed as the party of ‘the many’ not the few, but it is ‘the
many’ that is prominent, not the party political nature of these statements.
Moreover, when politics is framed around class conflict, which is rare, the
conflict is between the upper class, ‘the few’, and the working class, ‘the
many’, as in the 1970 editorial above: those ‘born to great privilege’ are in
opposition to everyone else. In both 1955 and 1964 there was extensive
coverage of the public-school-educated Conservative cabinet with the head-
line in 1964 of ‘Etonians of the world unite! You have nothing to lose but your
privileges!’ (12 October 1964) and when invoking the ‘image of the Tory party,
1966’ Alec Douglas-Home is described as the ‘maestro of the grouse moors’ (22
June 1966). For the Mirror, society is its readership, which is the working class.
Any class conflict is between this working class and the upper class (or ‘big

103
The New Politics of Class

industrialists’ or ‘Tory bankers’). Equally, strikes tend to be presented as the

fault of a minority of irresponsible ‘strikers’, not ‘workers’. This is nicely
shown in its verdict on Ray Gunter (then Minister of Labour and former
head of a railway workers union) calling him a:

moderate with a fierce patriotism which is largely based on his faith in the virtues
of the working class. So when unofficial strikers have behaved cruelly, stupidly or
short-sightedly he feels they have betrayed the rest. (16 March 1966)

We see that same theme in the Guardian and the Times, albeit refracted
through the lens of a very different readership, both in class and party
terms. While both share a middle class readership, they differ in terms of
party support during this era. The Times, apart from in 1945, endorsed the
Conservatives or the Conservatives/Liberals, whereas the Guardian editorials
were much more supportive of Labour during this period.17 Regardless of these
differences, both the Guardian and the Times talk at some length about the
working class in editorials and recognize class stratification frequently. On the
latter there is regular mention of ‘all classes’ with regard to almost every
imaginable activity: from nutrition (Times, 25 June 1945) to education
(Times, 22 June 1945), welfare services (Times, 28 April 1955) to the declining
morals of the young (Times, 1 October 1959), productivity (Times, 10 March
1966) to the theatre (Times, 26 March 1966), and so on. On the former, both
papers also talk a lot about specific policies that will benefit the working class
and also, particularly in the Times, what issues motivate working class voters.
In 1950 the Times has a detailed, and very neutral, discussion of working class
jobs and how pay rates are motivating strike activity (26 January 1950). This is
followed a fortnight later with an equally detailed discussion about working
class housing and how subsidies are ‘grossly maladjusted’ (17 February 1950).
Similar stories can be seen in the Guardian throughout this period as well. For
example, in 1945 it is argued that the ‘first need of the people . . . will be for
houses to let at working-class rents’ (22 June 1945) and in 1950, regarding
unofficial strikes, the editorial talks of being ‘brought into touch with
working-class currents of opinion’ (22 February 1950). Overall there is con-
stant recognition of those who do jobs ‘requiring heavy physical effort under
rough or unpleasant working conditions’ (Times, 02 May 1955) and how
policies are, and to some extent should be, aimed at them.
Of course, neither the Times nor the Guardian are actually read by many of
the ‘workers’ that they commonly talk about. This is perhaps best shown by
the slightly snobbish tone that is sometimes adopted: the Times in 1959 talks,
with no hint of irony, about ‘cheaper beer for the workers’ (16 September
1959) and often adopts a rather patrician tone when discussing differences
between workers. The Times is also quite clearly addressing a readership within
the upper middle class. One of the 1955 editorials bemoans the woe of

104
 The ‘Papers’

‘sending children off to [boarding] school’18 and there are regular references to
the arcana of proctors and punting that would only be known to Oxbridge
graduates. Perhaps more importantly, the two broadsheets also diverge from
the Mirror in their slightly different discussion of class and politics.
Part of this difference is emphasis. The Mirror regularly mentions landlords
and monopolies and their negative effect on ‘ordinary people’, yet this sort of
issue is only sporadically covered in the Times and the Guardian. Similarly
strikes, industrial disputes, and inflation, which more negatively affect the
middle class, get more coverage in the broadsheets. The type of story does
differ somewhat. The other difference is in the efficacy of class politics.
Although it is rare for editorials to directly link classes with parties, there is
recognition that the class system does feature in electoral competition to some
extent. For example, in 1950, the Guardian remarks that:

The chief difference [between the parties] is that either a Conservative or a Liberal
government would be inclined to let a slightly larger share of the national cake go
to the middle classes. (17 February 1950)

This is combined to some extent with suggestions that Labour should move
away from class politics. The Guardian in 1951 calls ‘for Labour to rid itself of
its appeal to class instincts’ (25 October 1951) and the Times in 1959 describes
Labour as the ‘party of jealousy; of a stubborn adherence to the class war . . . of
a pathetic belief that the few can be made to pay for the many’ (7 October
1959). The underlying premise of much of this type of discussion is that class
politics benefits the dominant group of the working class and harms the
middle class readers of Times and Guardian editorials. This is probably most
apparent in discussion of trade unions. The Mirror talks of ‘strikers’ disrupting
industry, whereas both the Guardian and the Times talk of ‘workers’ disrupting
production. For the Guardian in 1955: ‘the engine drivers who are threatening
to deprive the public of trains next Saturday are not a group of downtrodden
men: they are doing the oppressing’ (21 May 1955). Commenting on the same
proposed strike, the Times comments that ‘for some groups of workers, readi-
ness to strike has become a habit’ (11 May 1955).
These differences are less notable than are the similarities, however. During
the 1950s and 1960s, regardless of readership, class is a constant feature of the
way in which newspapers talk about the world. The idea of a stratified society
based around social classes is a regular explicit, and implicit, part of the
language used to describe politics. It is also the working class and their needs
and wants that are most frequently discussed, often not as in conflict with the
middle classes, but simply as the group that politics should serve, and to
which political parties should appeal. Clearly this is moderated to some extent
by the type of readers that leader writers are writing for, and the middle class
newspapers are more willing to accept that working class and middle class

105
The New Politics of Class

demands are different, but the basic story remains the same. This is not true
once we enter the 1970s.

Breakdown of Consensus (1974–1992)

By the middle of the 1970s there is a clear recognition within the editorials of
class conflict. Indeed, this can already be seen in 1970. The working and
middle classes are seen as on different sides of a debate about unions, inequal-
ity, and welfare. Just as the demographics of class numbers shift, and as the
nature of competition between the parties shifts, so does media coverage. This
is true of all the newspapers, but is, unsurprisingly, most commonly seen in
the middle class broadsheet papers. Both the Guardian and the Times com-
ment directly on class conflict. Before the first 1974 election the Times says
that it is undeniable that ‘there are now social tensions more acute than existed
ten years ago. There are tensions between the classes’ (12 February 1974) and
similarly in the Guardian it is argued that ‘the country could quickly find itself in
a class war’ (27 February 1974) and that we ‘must face the probability of a
political, social and economic struggle which will make us poorer, more bitter
towards each other, and with the growing inter-class tolerance of the post-war
years sadly eroded’ (5 February 1974). While 1974 is clearly exceptional in some
ways (a snap election following the incumbent Conservative government’s
ongoing dispute with the miners), it is more typical than one might think of
election coverage over the 1970s and 1980s. Both the broadsheets, and to a
lesser extent the Mirror, focus much more on class conflict and how policy
affects the working and middle classes differently. During the second 1974
election campaign, the Guardian lays out this division quite clearly:

Ask most managerial or professional men today and they will say that they are
worse off than a year ago. Taxation is biting deeper, money buys less, salaries and
fees have risen less than proportionately, while mortgages and interest payments
have gone up. Ask most wage earners and they will give the same answer. Wages
have risen, but food and clothes and housing costs and getting to work have all
gone up too. Both groups feel worse off. (23 September 1974)

The editorial then goes on to evaluate which of these groups has lost more (the
former in the Guardian’s opinion, albeit from a much higher base). There is
also a more explicit link made between party politics and opposing working
and middle class interests. The Mirror expands its definition of the ‘the few’ to
encompass a much wider swathe of the middle class. In 1979 the Mirror
says that the Conservative manifesto ‘is thick with predictable prejudice.
Help for private medicine and fee-paying schools. The biggest tax cuts for
the better-off ’ (12 April 1979) and in 1987 similarly accuses Margaret
Thatcher of caring about:

106
 The ‘Papers’

private medicine, private schooling and private privilege. She doesn’t care about
the majority of the British people who will not be voting for her. She doesn’t care
about the National Health Service because she never uses it. She doesn’t care about
a shortage of school books because her children never went to a State school. She
doesn’t care about council tenants because she has never lived in a council house.
(27 May 1987)

This is not the newspaper of the working class railing against the ‘1 per cent’,
but against a much wider section of the middle classes.19 This theme of middle
versus working class interests is quite starkly brought out in the some of the
broadsheet editorials. Again this is most notable in the 1970s, where the
Guardian leads in 1974 with ‘Mr Heath has a party base which does not
encourage him to favour manual workers at the expense of the well to do and
of the middle classes generally’ (13 February 1974). A few months later, during
the October 1974 campaign, the Guardian claims that ‘for the electorate, as
before, the choice is one that often starts from sectional interests, whether as
workers or as property owners’ (18 September 1974). This type of discussion
continues, although at a lower ebb, through the 1980s. For example, the Times
in 1987 talking about privatization of Rolls-Royce argues that:
Part ownership of the company for which one works forges a still closer link
between the fortunes of capital and labour. A class-based party like Labour is
muddled about the desirability of strengthening that link. (2 June 1987)

Even so, by the end of the 1980s the class–party link is increasingly seen as
historical rather than current. The eve of the election editorial in the Times
a week later talks about changes between classes and parties over the last
200 years:
Peel realised that Conservatism’s future lay not with the landed interest but with
the newly enfranchised, early 19th century middle class. Disraeli realised it lay
with the late 19th century, newly enfranchised working class. Mrs Thatcher
realised that it lay with the late 20th century ‘classless’ class of skilled workers
who had long been enfranchised in terms of the vote in elections but who were
trying to enfranchise themselves economically. What is the evidence? . . . the
Labour politicians’ terror—in this campaign—of appearing to be the party of
high taxation, the unions, and nationalization. (11 June 1987)

And by 1992 the Times is talking about the ‘traditional purpose of Labour . . . to
advance the interests of the great unions once concentrated in [the manufac-
turing] sector’ (24 March 1992). This move to emphasizing class as a historical
rather than current division is accompanied by decreasing attention to the
working class and an increasing attention to groups not identified as ‘workers’,
but as the ‘poor’ or the ‘disadvantaged’. Before 1974 ‘the poor’20 were men-
tioned in fewer than 7 per cent of editorials. Between 1974 and 1992 this ran
at nearly triple the rate: 20 per cent of editorials. By 1992 Labour is less of a

107
The New Politics of Class

party of the working class than ‘a party which exists to care for the poor’
(Mirror, 17 March 1992). There are still ideological differences between the
parties that map on to class divisions, but these are slowly changing. As the
Times puts it a week before the 1992 election:

Next week’s general election is widely regarded as devoid of ideological choice. The
Tories, having unceremoniously dropped the author of Thatcherism in 1990,
smartly move towards the centre. The Labour party, smarting from wounds
inflicted on it in three defeats, has performed a similar shift . . . The Labour leader-
ship remains . . . the party of organised labour and of collective action: of the
aspirations of a planned public sector and its multifarious beneficiaries. One of
Mr Kinnock’s achievements has been dramatically to expand the range of those
with an interest in a Labour victory, from blue and white collar workers to the great
professions, to doctors, to clergymen, academics, teachers, scientists, artists.
(4 April 1992)

Already by 1992 the writing is on the wall. The working class is no longer a
significant interest group that newspapers write about and class politics is
slowly changing. By 1997 both of these processes, at least as viewed through
the lens of Britain’s newspapers, are complete.

Middle Class Consensus (1997–2015)

From 1997 onwards the working class as a group almost completely disappears
from newspaper editorials, as do the ideas of class stratification and class
politics. The only editorial in the Mirror that uses the word ‘class’ between
1997 and 2015 refers to Tony Blair’s choice of underwear.21 Moreover, the
middle class is now the default class group against which policies are judged;
the middle class is now representative of society as a whole.
During the 1997 campaign there is a degree of reflection on what was. In
1997 the Guardian suggests that ‘the two sides of industry have declared a
truce in the class war’ (21 April 1997) and that Britain ‘had already ceased to be
the industrial society divided along traditional class lines which moulded the
two party system’ (30 April 1997). This can be seen towards the end of the
period as well. As with many mentions of the working class over this period, it
is seen as a group that is no longer relevant, with retiring Labour MPs referred
to as ‘old hands still in tune with Labour’s working class origins’ (Guardian,
12 April 2010). By 2015 both direct mentions of the working class in the
broadsheets refer to them as a type of ‘problem group’. The Times talks about
the foundation of a think tank as a ‘search for solutions to the challenges facing
the working class’ (2 May 2015) and the Guardian notes that ‘working-class
voters . . . will have been a large part of the third of the electorate who didn’t vote
in 2010’ (16 April 2015). The latter point is a theme we return to in Chapter 8.

108
 The ‘Papers’

Most direct references to class and politics over this period refer to the way
in which class politics has vanished and how the parties have converged. In
2015 the Guardian notes that ‘parties long since stopped representing one half
of the electorate or the other’ (16 April 2015), and there is much coverage of
how Labour is now a middle class party.22 Unsurprisingly, these feature heav-
ily in the paper with a left-leaning middle class readership: the Guardian. In
2001 it leads with this:

Labour is harvesting some stunning endorsements from deep inside what were
once the Conservatives’ sociological heartlands. Tony Blair’s lead over William
Hague among professional and middle-class ‘AB’ voters is running at 42%, the
Observer/ICM poll reported on Sunday . . . Labour has worked long and hard to
earn such levels of support among the groups which once spurned the party’s
appeals with contempt and loathing. (14 May 2001)

This can also be seen in the 2010 campaign. Gillian Duffy, described as a
‘bigoted woman’ by Gordon Brown after she raised the issue of immigration
when he was on the campaign trail in Rochdale, is for the Guardian ‘Labour
family, part of the hereditary working class’ (29 April 2010). Yet this is some-
one whose views Labour now apparently spurns. Again the working class is
only mentioned in order to say that their views are not held by Labour. In a
slightly different way editorials in other newspapers, especially the Mirror
before 2015, emphasize how Labour is now a classless party. Talking about
Alex Ferguson’s support of Labour in 2010, the Mirror argues that ‘Labour
represents working people, whether on low wages or the high income of the
ex-toolmaker turned football manager’ (28 April 2010). Given that Ferguson
earned around £4 million a year in 2010, Labour presumably represented
pretty much everybody bar the Queen. More generally, there is widespread
acceptance that there is a new consensus, and this consensus is on the right.
The Guardian bemoans the fact that no ‘party in the 2001 election advocates
redistribution from the rich to the poor, a rise in the top rate of income tax’ (16
May 2001). Yet it is also more positive that, on the flip side, ‘the business
vote—historically and histrionically anti-Labour—is now split’ (2 May 2005).
And the Times points out in 2005 that:

For all of the furious charges traded by Tony Blair and Michael Howard, in practice
British politics has been witnessing a quiet conceptual convergence . . . There is
also a broader, if not universal, understanding that these openings and opportun-
ities are rooted in individual choices and responsibilities and, generally, not the
result of delegating authority to the government. (3 May 2005)

Class does not completely disappear, of course. For example, the Mirror in
2010 runs an almost identical story about David Cameron’s schooling at Eton
as it did about Anthony Eden, complete with picture of him in Eton uniform

109
The New Politics of Class

in 1955.23 More starkly, class does rear its head again in 2015. In the broad-
sheets this is in the mildest manner. The Guardian talks a little more about
inequality and how the parties will address it, arguing that ‘While most Tories
shrug at that yawning gap between rich and poor, Labour will at least strive to
slow and even reverse the three-decade march towards an obscenely unequal
society’ (1 May 2015). On the opposite side of the partisan fence, the Times
agrees with Cameron that the Conservatives are ‘the party of “of the small
businesses, the techies, the rooftilers, the retailers, the plumbers and the
builders”’ (28 April 2015). Nonetheless 2015 does not mark a watershed
moment for class politics, as seen from the pages of the broadsheets. The
Guardian might ask the question, ‘who is refighting the class war now?’, with
reference to Conservative proposals for union strike ballot legislation, but it asks
it after pointing out the ‘fiscal rectitude’ of Labour and the reinvention of the
Conservatives as the ‘free-spending party of the workers’ (15 April 2015).
The Mirror in 2015 is somewhat different. A leader at the beginning of the
campaign argues that:

The Conservatives have devoted five years to short-changing working people . . .

Working people and Britain can do better than Cameron’s Tories, a leader and
a party who live in a different world to the rest of us. Most Tory candidates are
a different breed—a breed that never cared about working people for five years.
(15 April 2015)

This pattern continues to some extent through the month before the election,
with exhortations that the ‘economy is working for the very wealthy minority
at the expense of the majority . . . Earnings and wealth need to be shared more
fairly. A privileged few should not be receiving vast sums they could never
feasibly spend’ (29 April 2015). This is discussion of class conflict, albeit of the
few versus the many, and this is spelt out in partisan terms. The Conservatives
are ‘not interested in workers struggling to earn higher wages’ (2 May 2015)
and Labour will set ‘about creating an economy and politics which work for us
all instead of that privileged minority’ (6 May 2015). The day before the
election, the Mirror states that ‘Labour is a party of the people. It understands
our lives and comes up with answers’ (6 May 2015).
Part of this change in rhetoric is probably due to a strongly Labour partisan
newspaper reacting to an incumbent Conservative government, but there is
perhaps another message here. By 2015 the mainly working class readership of
the Mirror are likely to feel that the incumbent government is further away
from them than any government has ever been. As the lodestone of British
politics moved to the right, it is the working class that are left behind and the
Mirror is perhaps representing this. At the same time, the Mirror is still men-
tioning class less than it was in the 1980s, and much less than before that. It
might be representing working class dissatisfaction with politics, but it is

110
 The ‘Papers’

doing so through the partisan lens that is has always used. Overall, and even
with the slight exception of 2015, class is simply much less visible after 1992
than it was previously. Few editorials talk about class, few refer to the working
class, and few refer to conflict between class groups.

Conclusions

What does this all add up to? There are three important points to make. First,
there have been dramatic changes in how newspapers portray class in society
over time. The working class was the class that editorials were concerned with.
Now it is the middle class. This reflects the change in society from being one
numerically dominated by the working class to one with an increasingly large
educated middle class. As those numbers changed, so did the nature of class
conflict and how it was portrayed by the press. Conflict between the classes
was not really mentioned in the 1950s, but by the 1970s and 1980s it formed a
core part of many editorials. As the working class shrank further, so did
editorials about conflict between the working and middle class. Perhaps
most noticeably, the very idea of class stratification, which was part and parcel
of journalism in the 1950s, now rarely makes an appearance. These changes
are about quantity of coverage, but also about how stories are framed. Gone is
the rhetoric about ‘all classes’; gone is the talk of the ‘working man’ and how
particular policies will benefit him; gone is the background acceptance of class
differences.
Second, the way in which class is linked to party, and policy, has also
changed. While direct links between policy, party, and class were never that
great, they are now almost completely absent. When editorials mention class
and politics today, it is mainly to note how the links between parties and
classes are no more. The cues that newspapers gave to voters in terms of class
politics are no longer as clear as they once were. The associations of Labour
with the working class and the Conservatives with the middle class that were
made before 1997 in every newspaper largely stop after this point.
Third, while what newspapers say is clearly important, so is the nature of
their readership. The British newspaper market remains highly segregated by
social class, and by partisanship. Few middle class Conservatives read the
Mirror and few working class Labour voters read the Times. These patterns of
partisan and class association have remained unchanged, but the numbers of
people reading a newspaper has plummeted. When 80 per cent of people read
a paper in the 1950s and 1960s, class and party identities were constantly
reinforced by newspapers. Less than a third of people now read a newspaper,
so these reinforcement pressures consequently affect far fewer people. It is not

111
The New Politics of Class

just that the material within newspapers has changed but that newspapers
have less of an impact on the electorate than they did.24
The timing of these changes is also important. Discussion of class, and class
politics, stops in 1997. Newspaper readership starts falling substantially dur-
ing the middle of the 1990s. As we see in Chapter 6, the first of these changes
coincides with changes within the political parties. Parties stop talking about
class, and stop having class-based policies, which means newspapers no
longer talk about class either. This is magnified by the coincidence of a
declining newspaper market, largely due to new technology, which no longer
reinforces class and partisan identities as it used to. We know that media
messages that are consistent with party messages are more likely to shift the
attitudes of partisans (Carey and Burton 2004). The disappearance of both
media and party messages regarding class and party seems highly likely to
disrupt voter views of class-based parties.

Notes

1. The precursor to the Sun, the Daily Herald, was effectively owned by the Trade
Union Congress and was the most leftist mass newspaper after the war. After being
re-launched as the Sun in 1964 it was taken over by Rupert Murdoch in 1969 and
remained broadly on the left until the 1979 election when it endorsed the Conser-
vatives. The impact of its switch back to Labour in 1997 has been a source of much
debate. Whether the existing readership changed partisanship due to the editorial
change, whether Conservative partisans stopped reading and Labour partisans
started reading the Sun, or whether the editorial line simply switched after the
readers switched, is very difficult to test. Editorial endorsements by the Sun have
been argued to have had little effect (Curtice and Semetko 1994; Curtice 1997;
Norris et al. 1999) and relatively large effects (Ladd and Lenz 2009; Newton and
Brynin 2001) in shifting existing reader opinions. Interestingly, Rupert Murdoch
himself said at the Leveson inquiry in April 2012 that claims that it was the ‘Sun
wot won it’ in 1992 by influencing readers were ‘wrong in fact—we don’t have that
sort of power’ (Daily Telegraph, 25 April 2012).
2. This shows the proportion of people supporting the Conservatives of those who
support one of the three major parties. These figures as a proportion of all people
would look very similar, but with a small decline in Conservative support over time
for all readers. This is due to the increasing number of people supporting smaller
parties and the growth in the number of people who do not support any party
(which has increased from under 10 per cent in the 1960s to over 20 per cent in the
2010s). Non-partisans are, perhaps unsurprisingly, over-represented in the groups
of people who do not read a daily newspaper: a group which has grown
substantially.
3. There may be a further effect of reinforcement at election time, which is to increase
turnout. Brynin and Newton (2003, p.72) argue that ‘those whose partisan views

112
 The ‘Papers’

are reinforced by their paper are more likely to vote’. These are relatively small
effects, but show how reinforcing and cross-pressuring messages from newspapers
can alter behaviour.
4. In principle, experimental work can overcome these problems (see, for example,
Ansolabehere and Iyengar 1995; Iyengar and Kinder 1987; Norris et al. 1999). The
problem here is external validity. While we might find persuasion and agenda-
setting when people read newspaper articles in a psychology lab, can we assume
that this happens outside the lab? Some of the best work in this area has used natural
experiments. These typically involve looking at the effects of an outside event, like a
printing strike or a change of newspaper partisan affiliation, on people’s reactions in
the real world. A good example, in the British context, is Ladd and Lenz (2009) who
use panel data that follows the same people over time to argue that changing
newspaper endorsements in Britain in the 1990s affected party choices.
5. These are multinomial logit regression models run separately for each decade of
survey data. As well as education and class, also included in these models are sex,
age group, and region. The probabilities reported here are thus for a woman living
in the South East of England who is in her fifties.
6. The question asked in the BSA in every year since 1983 is, ‘do you normally read
any daily morning newspaper at least three times a week?’. This is slightly different
to the question on the BES surveys which tend to ask a variant of ‘do you read a
morning newspaper regularly?’.
7. There is a separate question about whether this change is primarily generational, or
a decline that has affected everyone. Curtice and Mair (2008, p.166) suggest that it
is a combination of the two: ‘newspapers have failed to keep older readers who were
once loyal to them. But at the same time they have apparently found it more
difficult to recruit and retain younger readers.’
8. In some ways the Telegraph might be thought to be a more obvious representative
of the old middle class, but in fact a higher proportion of the Times readership is
drawn from this group in both the 1960s and 2010s.
9. We also include a small number of election ‘guides’ typically printed on the day of
the election that contain explicit endorsements of candidates.
10. Most notably the Times was not printed from 1 December 1978 until 12 November
1979 due to industrial action, meaning that we do not have any coverage of the
1979 election from the Times. Equally the Mirror did not consistently contain
editorials before the 1980s, meaning that slightly more than 15 per cent of our
Mirror sample does not have an editorial.
11. We include here some synonyms for class terms. For the working class we include
words that concentrate on occupational class status such as ‘workers’, ‘labour’, or
‘wage-earning class’. We do not include mentions of ‘the poor’, ‘low earners’, ‘the
needy’, and ‘slum dwellers’: terms which focus on people with low incomes, but
not a class group. For the middle class, we include ‘the bourgeoisie’, ‘the privileged’,
‘the better off ’, and the ‘well-to-do’. While these terms conflate income and
occupation to some extent, the most common usage by far is ‘middle class’.
12. While many of the editorials that mention specific classes also mention the class
structure or class conflict, many do not. Three quarters of articles that talk about

113
The New Politics of Class

the class structure or mention social class do not specifically mention the ‘working
class’ or ‘middle class’, and over half do not mention any synonyms for working or
middle class such as ‘workers’ or ‘labour’.
13. The upper class group includes references to the upper class, the rich, landowners
and ‘capital’. About half the references are to the middle class and half to the upper
class.
14. The second of these means editorials that, in the main, discuss fairness and social
justice or directly issues of equality. We also include mentions of egalitarianism
and inequality.
15. We see a similar pattern over time for discussion of trade unions. Mentions of
unions peak in the 1970s (in February 1974 nearly two thirds of editorials mention
trade unions) and then fall to very low levels from 1992 onwards. This decline is
particularly notable for the Mirror, the working class newspaper, which does not
mention unions once in an election campaign after 1987. This is unsurprising
given that unions from the 1990s onwards became increasingly dominated by
middle class public sector employees as discussed in Chapter 4.
16. Some editorials also identify parties as being against particular groups. This is most
prominent in the Mirror portraying the Conservatives as against the working class.
There are three of these editorials in the post-war consensus period, two in the
1974–1992 period, and another three after 1992. All but one are in the Mirror.
17. This is not to say that the Guardian endorses Labour in every election. In fact, the
Guardian endorsed the Liberals in 1945 and 1950, and the Liberals/Conservatives
in 1951 and 1955. Nonetheless, the editorial stance, in line with its readership, is
systematically less supportive of the Conservatives than the Times.
18. In the same year the Times also describes ‘unfair pressure on one group: large
families with independent school fees’ (4 May 1955). Needless to say, this is a
group in which many of its readers are likely to find themselves.
19. According to the British Election Study in 1987 over a third of parents had used, or
were thinking of using, a private school for their children and/or held private
medical insurance. Equally, by 1987 only 23 per cent of people rented a house
from a local authority or housing association.
20. Here we include mentions of ‘the poor’, ‘the needy’, ‘slum dwellers’, the ‘less well
off ’, and ‘low earners’. The vast majority of references are to the first, and the
pattern over time is the same if we just look at mentions of the poor.
21. His Calvin Klein underpants get the Mirror seal of approval as they ‘ooze class and
statesmanship. And modernity. And success. If anything personified cool Britannia,
it’s Tony Blair’s pants’ (6 June 2001).
22. There is occasional discussion of how this has meant that the parties have some-
times traded their historic positions as well. For example, the Guardian in 2015
notes that ‘David Cameron launched the Conservatives’ general election manifesto
today in the same spirit of light-fingered transvestism. Twenty-four hours after Ed
Miliband had wrapped himself in Labour’s newfound fiscal rectitude, Mr Cameron
came down the election catwalk sporting a Tory new look as the free-spending
party of the workers’ (15 April 2015).

114
 The ‘Papers’

23. In general, and in all periods, the Mirror focuses much more on the class back-
ground of politicians than the Times and the Guardian. This might be seen as a
response to the bigger differences between politicians and the Mirror readership
than between the broadsheets and MPs, although equally this could be seen as
simply due to a classic tabloid focus on ‘personality’ rather than ‘politics’.
24. One offsetting factor here might be the degree to which parties are distinct from
one another. For example, Newton and Brynin (2001, p.282) argue that ‘voters are
more likely to follow their newspaper’s lead when there is less to choose between
the parties’.

115
6

The Parties

The media contribute their own perspective on politics and class, but they also
provide a channel for communication between parties and voters. In this
chapter we examine exactly what parties offer voters in terms of policy,
group appeals, and descriptive representation. Our focus is on the two main
political parties that have held genuine power throughout the post-war era:
Labour and the Conservatives. Most countries have parties of the left and
right, but Britain is arguably unusual in the political focus on class by these
two parties. Is that now simply history? Our answer is yes, and the impact of
these political changes on the relationship between voters and parties is
fundamental to the message of this book.
In this chapter we focus on different key sources of political messages. First,
we look at the manifestos that parties produce before each election. This
involves coding the two main party manifestos for every election campaign
since 1945. Manifestos are substantive: they elaborate the core policy and
programmatic statements to which governing parties are often, if not always,
held. Very few people actually read them, but they are extensively distributed
and commented on by the media. The manifestos give us a way to examine
both the positioning of parties and their appeals to particular groups. We
particularly focus on the degree to which parties present distinctive left or
right wing policy platforms in their manifestos. As noted by various scholars:
‘the basic logic of party competition in Britain remains similar to that which
held in the 1950s: in policy terms at least, it is . . . a predominantly left-right
dimension’ (Webb 2004, p.39; see also among many others, Laver and
Budge 1992; Clark 2012). Equally, as we saw in Chapter 4, classes differ in
the degree to which they express preferences for such left- versus right-wing
policies. Parties are therefore representing the differing preferences of middle
and working class people by offering distinctive policies. If the parties do not
differ in the left–right choices they offer, this link between classes and parties
will not be present. We argue that over time the parties have converged and,
 The Parties

more specifically, the Labour party has become less left-wing and more like the
Conservatives.
But political representation is not just about policies; it is also about the way
that parties see their political role as representatives of social groups. As well as
policy differentiation, therefore, we are also interested in references to types of
voters. Our expectation is that over time parties will refer less to classes and
more often to other social groups. We look at this with the manifesto data, but
we also examine the party leaders’ speeches. Leader’s speeches are key plat-
forms for party identity building, both positively about the leader’s own party
and negatively with respect to labelling the opposition. This form of commu-
nication is essentially about rhetoric rather than policy proposals, and is likely
to be particularly useful for looking at which social groups are appealed to.
We code party leaders’ speeches at the annual party conferences for every year
since 1945 and see how references to class, and classes, have changed, and
whether other groups have taken their place.
Finally, we look at how the social composition of MPs within each party has
changed over time, and most importantly whether voters have paid attention
to anything that the parties have done. We show that voters have noticed
these changes, both in terms of policy, and also in their views of the parties as
class parties. Ultimately, our argument is that the signals, in terms of policy,
rhetoric and personnel, from politicians to voters have changed and that
voters have recognized these changes.

Party Policy

The post-war story of British politics and policy is thought to be well known.
Labour’s landslide electoral victory in May 1945 produced a raft of policies on
health care, welfare, and public ownership that laid the fundamentals of what
became known as the post-war consensus. The acceptance of many of these
policies by the Conservatives, combined with an initial commitment in its
1947 Industrial Charter not to reverse them, ensured their continuation as the
fundamentals of British public policy until the 1970s. The end of the post-war
‘boom’ in the wake of the 1973 oil crisis signalled the departure from this
consensus. Ensuing strikes, three-day weeks, years of high inflation, and the
intervention of the IMF laid bare underlying differences between the parties
and ultimately led to Thatcherism. The emergence of ‘Blairism’ moving Labour
to the right in response has likewise been well documented: the dropping of the
traditional version of Clause IV of Labour’s constitution in 1995 and the simple
fact that Labour’s manifestos have not contained the word ‘socialism’ since 1992.

117
The New Politics of Class

Of course, this oversimplifies the movements of the two parties. Although

most interpretations of political developments have been carried out by
contemporary historians with rich descriptive studies of parties and their
strategies, we turn here to quantitative measures of where parties stand on
economic left–right issues. We analyse the party manifestos using coded
policy content collected by the Manifesto Project on Political Representation
(MARPOR) (Budge et al. 2001; Klingemann et al. 2006; Volkens et al. 2015).
Sentences in the manifestos are grouped into fifty-six themes and policy areas,
as shown in Appendix Table A6.1. Sentences are then divided by the total
number of sentences in the manifesto to standardize for their varying lengths.
Finally, the manifestos are coded into left-versus right-wing positions.
Although MARPOR has developed a simple additive measure which sums and
subtracts percentages of sentences referring to policies (RiLe), we use a more
sophisticated economic left–right scale developed by Prosser (2014). This uses
some of the same building blocks as the RiLe, but is more focused on economic

a) Manifesto data
70
Conservatives
60

50

40
Labour
30

20
1945 1960 1975 1990 2005

b) Expert survey data

10
Conservatives
8

4
Labour
2

0
1945 1960 1975 1990 2005

Figure 6.1. The left–right position of parties

Note: The figures here show the positions of the two main parties on left–right policy scales; higher scores indicate
more right wing positions. The left-hand graph shows positions using manifesto data (two period moving average)
and a procedure presented in Prosser (2014). The right-hand graph shows expert ratings of parties on 0–10 scales.
Source: Manifesto Project on Political Representation 1945–2015; Castles and Mair 1984; Laver and Hunt 1992; Ray
1999; Steenbergen and Marks 2007; Hooghe et al. 2010.

118
 The Parties

left–right positions and avoids some of the methodological weaknesses associ-

ated with RiLe. More detailed information on how the manifestos were coded
using this procedure is presented in the appendix to this chapter.
The left-hand graph in Figure 6.1 shows the extent to which Labour and the
Conservative manifestos emphasize left- or right-wing policy positions from
1945 until the present day. The absolute positions of parties on a left–right
spectrum cannot be ascertained using these coding procedures, but their
relative positions can be. We can see the size of the policy difference between
the parties and how they move relative to one another. During the immediate
post-war era there is a marked gap between the parties. Indeed it takes until the
late 1960s for the Conservatives to converge with Labour. The policy gap then
almost immediately emerges again during the 1970s. This perhaps fits better
with more revisionist accounts of the post-war consensus that claim that any
consensus was more myth than reality (Pimlott 1992). In fact, some argue that
clear continuities can be observed in major areas such as economic policy,
union relations, and welfare for the first fifteen years after the war (Kavanagh
and Morris 1994). We would tend to agree with this. While the two parties did
converge on policy after the war, this was not until the late 1960s and it was
for a relatively limited period of time.
Yet there is another period of convergence. The policy divisions between
the parties also collapsed during the 1990s, but this time did not re-appear.
The most striking shift in Labour’s position took place under Tony Blair
between 1992 and 2001.1 This was explicitly stated in their 1997 manifesto:
‘We aim to put behind us the bitter political struggles of left and right that
have torn our country apart for too many decades. Many of these conflicts
have no relevance whatsoever to the modern world—public versus private,
bosses versus workers, middle class versus working class. It is time for this
country to move on and move forward.’
This dramatic shift is also shown in another source of information on policy
convergence between the two main parties. This is a set of ‘expert surveys’
conducted over the period 1982–2010. These surveys are undertaken within
the community of academics who study British politics. They are asked to score
parties on a wide range of policy scales. These answers can then be used to give
an assessment of party positions on the left–right dimension.2 In the right-hand
graph of Figure 6.1, we present estimates of the Labour and Conservative
Parties’ general left–right positions from the expert surveys that cover the
elections during this period. Again we see a particularly pronounced policy
convergence in British politics during the 1990s. There is a sharp tightening
of the gap between Labour and the Conservatives at the time of the 1997
election. The move by Labour to a more centrist position, combined with a
general re-branding of its image, appears to have been a one-step change. There
has been some evidence of divergence since, but it is far less pronounced.

119
The New Politics of Class

Group Appeals
Party positions are not the same as group appeals. Manifestos and expert
surveys indicate that a more ideologically neutral policy programme has been
adopted by both main parties. But this is an indirect indicator of different
parties’ concern with class groups. A more direct indicator is whether mani-
festos still refer to classes. While manifestos very rarely refer to class politics
directly, they do talk about particular groups. To examine the use of class
appeals, we code the use of terms that can be reasonably thought as represent-
ing the working class. We also consider references to groups—the unemployed,
the poor, and unions—who are likely to be linked in some way with the working
class, as well as alternative, class-neutral groups like ‘families’. We code the
manifestos using computer-based coding procedures (Lexicoder).3 We were
also able to obtain hand-coded analyses of the original full party manifestos
for all elections between 1974 and 2005, the period of most marked change,
from Mads Thau. He maps out which parties best represent which constituents
in terms of the claims made about their interests (Thau 2016).
Figure 6.2 shows direct mentions of the working class in the manifestos
using both methods. The graph on the left is the computer-coded data; the
graph on the right is the hand-coded data. Both show very similar patterns.
Over time, Labour’s manifesto references to workers clearly decline. This

a) Computer-coded b) Hand-coded
15% 20%

Labour Labour
15%
10%

10%

5% Conservatives
5%
Conservatives

0% 0%
1945 1965 1985 2005 1945 1965 1985 2005

Figure 6.2. References to ‘the working class’ in manifestos

Note: The left-hand graph here shows the percentage of group references that refer to the working class in all
Labour and Conservative Party manifestos (three period moving average). These are coded directly from the
manifesto text using a computer-based coding procedure (Lexicoder). The right-hand graph shows the percentage
of group references to the working class coded from the original manifesto documents 1974–2005 by Thau (2016).
Source: Manifesto Project on Political Representation 1945–2015.

120
 The Parties

change is most dramatic during the elections between 1987 and 2001. For the
hand-coded data, the percentage of Labour class appeals referring to the
working class falls from an average of just over 15 per cent in the 1970s and
1980s, to just 3 per cent from 1997 onwards. These sorts of references are
generally less noticeable for the Conservatives, which is perhaps unsurprising,
though they too used to refer to workers rather more than they do nowadays.
Nonetheless, the main point is that Labour used to appeal regularly to the
working class and it has stopped doing so to such a degree that there is now
essentially no difference between the two parties.
Figure 6.3 shows how references to trade unions and the unemployed follow
a pattern that matches the political saliency of the issues. High rates of union
militancy and high rates of unemployment tend to mean that the parties talk
about them more. The declining prominence of unions following the Thatcher
government’s reforms is reflected in the manifestos of both parties from the
1980s onwards. Union appeals peaked at the beginning of the 1980s for Labour,
but fell after then. Although the Conservatives mentioned unions quite fre-
quently in the immediate post-war period, the peak in the early 1980s is also
quite pronounced. References to the unemployed increase for both parties
during the period of high unemployment in the 1980s and early 1990s and
then subside. Moreover, we can see the origins of a different perspective on the
unemployed from the 1997 Labour manifesto which asserted that: ‘we will get

a) Trade unions b) The unemployed c) The poor

20% 4% 4%

Conservatives
Conservatives Labour

10% 2% 2%

Conservatives
Labour
Labour
0% 0% 0%
1945 1975 2005 1945 1975 2005 1945 1975 2005

Figure 6.3. References to unions, the unemployed, and the poor in manifestos
Note: The left-hand graph here shows the percentage of group references that refer to unions in all Labour and
Conservative Party manifestos (three period moving average). The middle graph shows the percentage that refer to
the unemployed (three period moving average). The right-hand graph shows the percentage that refer to the poor
(three period moving average). These are coded directly from the manifesto text using a computer-based coding
procedure (Lexicoder).
Source: Manifesto Project on Political Representation 1945–2015.

121
The New Politics of Class

30%

Conservatives

20%

Labour
10%

0%
1945 1965 1985 2005

Figure 6.4. References to families in manifestos

Note: The figure here shows the percentage of group references that refer to families in all Labour and Conservative
Party manifestos (three period moving average). These are coded directly from the manifesto text using a
computer-based coding procedure (Lexicoder).
Source: Manifesto Project on Political Representation 1945–2015.

the unemployed from welfare to work’ and we will ‘stop the growth of an
“underclass” in Britain’. The idea of the unemployed as an ‘underclass’ contrasts
with previous manifestos which tended to blame the economy for unemploy-
ment.4 References to the ‘poor’ display volatility in the earlier years but then
reach a stable, but low, level from which there is no decline.
But if class appeals have withered, who or what have taken their place? If we
want to see where parties nowadays focus their references and their appeals, we
need to turn to ‘class-neutral’ notions of family and parenthood. As Figure 6.4
demonstrates, references to families increase dramatically for both parties.
References to families are not only of a much larger magnitude generally,
they nowadays constitute approximately 30 per cent of all references for both
parties. Even at their peak in the late 1950s, no more than 13 per cent of
appeals by Labour, and far fewer by the Conservatives, concerned the working
class in any form. Moreover, this is a dramatic and swift change, taking place
almost entirely after 1997. Families are the inclusive reference group for the
political parties today. The rapid rise of the family as the key political focal
point can be illustrated by examining every mention in party manifestos of
the now fashionable reference to ‘working families’. We can see how it starts
with the Conservatives who refer to ‘working families’ in their 1992 mani-
festo, followed by similar references in 1997 (‘we are shifting power and
wealth back to working families’) and 2001 (‘Labour have increased taxes on
hard-working families’), before it was picked up by Labour in 2005 with ‘a plan
to improve the lives of hard-working families’. From then on both parties

122
 The Parties

deployed this usefully class-neutral and socially positive construct, especially

in 2015. Compare: ‘We want a better deal—and low bills—for hard-working
families’ (Conservatives) versus ‘we will reform the energy market so that it
delivers fairer prices and a better deal for working families’ (Labour).
In short, we find a striking decline in appeals to workers and a striking rise in
references to general notions of family. Much of this transition in emphasis
took place from the 1990s onwards.

Party Rhetoric

But what of rhetoric? Leaders’ speeches are perhaps a better way of assessing
group appeals. To examine leaders’ speeches we use the same coding proced-
ure as we used for the manifestos. As we shall see, they tell pretty much the
same story, with one interesting exception that fits well with their function as
a partisan, in-group versus out-group, rallying-call. This exception is the
tendency for speeches to refer occasionally to class envy, division, and conflict
explicitly. Always, however, this is for rhetorical purposes and even then is
used very sparingly. Direct mentions of social class are so rare (fewer than sixty
mentions in both parties’ leaders’ speeches over almost seventy years) that
plotting quantitative figures makes little sense. We can get a clearer idea of this
use of class by reading the speeches. When we do, we find some surprises. The
leaders who most clearly assert the language of class are Blair for Labour and
Churchill for the Conservatives.
For Churchill, this is true in opposition: ‘The driving force of Socialism is
class hatred and envy’ (1949); and government: ‘our opponents have another
theme on which they greatly count—I mean class warfare and the exploit-
ation of jealousy and envy’ (1953). Labour’s promotion of class warfare and
class envy was a theme echoed more sporadically by later Conservative leaders
such as Macmillan: ‘Our aim is to harmonize different and conflicting inter-
ests, not to set them against each other with the strident accents of the class
war’ (1960). And also Thatcher: ‘Class warfare is immoral, a poisonous relic of
the past’ (1978) and ‘Labour relish class division. They depend on it. It’s the
root of all Socialism’ (1984). John Major briefly reiterated the idea: ‘It’s a
matter of breaking down the false and futile divisions, based on class and
envy, that have been around for generations. Labour fosters those divisions’
(1991). However, the use of class as rhetoric by Conservative leaders seemed to
have faded from then on. New Labour’s leadership had effectively removed
the power of class war rhetoric.5
References to class by Labour leaders are similarly rare. During the 1950s,
class is not present at all, as speeches while in opposition focused primarily on
foreign policy. Wilson did use the class warfare theme negatively to label the

123
The New Politics of Class

opposition: ‘Let us leave to them their vision of Britain, the Britain they seek to
restore, a class-ridden Britain, a Britain of privilege, of social privilege, of
educational privilege, a Britain based on the right of that privileged class to
lord it over all the rest of Britain’s citizens whom we represent’ (1966). But it is
actually Blair who used the rhetoric of class most clearly. Like a Churchill in
reverse, he used references to class divisions as a way of explicitly or implicitly
denigrating the Conservatives: ‘We are proud of our history, but its weight
hangs heavy upon us. Why? Because for far too long it has left us defining
ourselves as a nation, not by what unites us, but by what divides us: a class
system, unequal and antiquated’ (1994). His aim in contrast was to ‘create a
model twenty-first-century nation, based not on privilege, class, or back-
ground, but on the equal worth of all. The class war is over. But the struggle
for true equality has only just begun . . . it is us, the new radicals, the Labour
party modernized, that must undertake this historic mission. To liberate
Britain from the old class divisions, old structures, old prejudices’ (1999).
We might say that Conservative class war rhetoric was pronounced in the
post-war era, returned intermittently in the conflictual 1970s and 1980s, and
then faded once Blair had rendered Labour harder to criticize on that front.
Labour’s rhetorical use of class is like a mirror image of the Conservatives: class
division is bad (them), openness and classlessness is good (us). Ironically, this
class rhetoric was most forcefully invoked just as Labour decisively shifted
away from the working class. It was then dropped again: Ed Miliband made no
reference to class in any of his speeches.

15%

Labour

10%

5%

Conservatives
0%
1945 1965 1985 2005

Figure 6.5. References to ‘the working class’ in speeches by party leaders

Note: The figure here shows the percentage of group references that refer to the working class in leaders’ party
conference speeches (six period moving average). These are coded directly from the full transcripts using both
manual and computer-based coding procedures (Lexicoder).
Source: British Political Speech Archive 1945–2014.

124
 The Parties

These direct mentions of class politics aside, there are substantial

similarities between the content of speeches and manifestos in terms of group
references. Figure 6.5 shows references to workers as a proportion of all
group references. As with the manifestos, we code the party leaders’ speeches
using computer based coding procedures. See the Appendix for details.
The figure is far denser than the same figures for the manifestos because of
the far larger number of leaders’ speeches. Apart from the added detail, how-
ever, this is a very similar picture to that seen with the manifestos. Conservative
leaders rarely mention the working class, although the spirit of ‘big working
class’ Britain can occasionally be seen in the immediate post war period. For
example, in his 1950 speech Churchill encouraged constituencies to select
‘active Trade Unionists and others representing the views of our brothers and
sisters in the working classes of the nation’. Labour leaders emphasize the
working class much more often, but from the late 1980s onwards the working
class as a group largely disappears from view. Kinnock started to de-emphasize
the working class and Blair finished that process. In fact, references to the
working class by Blair declined to levels similar to those found in the party
speeches of Conservative leaders.
If we look at references to unions, the unemployed, and the poor in
Figure 6.6, we see the same pattern observed in the manifestos: both union
and unemployment references spike sharply upwards during the 1970s and the
1980s respectively, but then fall back to insignificance. These changes clearly
reflect the levels and urgency of unemployment (and union conflict) as an

a) Trade unions b) The unemployed c) The poor

Conservatives
Labour
20% 4% 4%
Conservatives

Labour
10% 2% 2%

Labour Conservatives
0% 0% 0%
1945 1975 2005 1945 1975 2005 1945 1975 2005

Figure 6.6. References to unions, the unemployed, and the poor in speeches by party
leaders
Note: The left-hand graph here shows the percentage of group references that refer to trade unions in leaders’ party
conference speeches (six period moving average). The middle graph shows the percentage of all speech units that
refer to the unemployed (six period moving average). The right-hand graph shows the percentage of all speech
units that refer to the poor (six period moving average).
Source: British Political Speech Archive 1945–2014.

125
The New Politics of Class

30%

Labour
20%
Conservatives

10%

0%
1945 1965 1985 2005

Figure 6.7. References to families in speeches by party leaders

Note: The figure here shows the percentage of group references that refer to families in leaders’ party conference
speeches (six period moving average).
Source: British Political Speech Archive 1945–2014.

issue. Unlike those groups, and just as in the manifestos, references to the poor
did not decline in the 1990s. They persisted without being a major feature of
contemporary speeches (a few percent of group references at most). As with the
manifestos, however, the really marked trend is in a much greater emphasis on
families. As Figure 6.7 shows, both parties’ leaders increasingly refer to families,
parents, and children.
In summary, the working class no longer features very noticeably in the
vocabulary of our parties and politicians. It appears that politicians have
moved from appealing to the working class, when formulating and discussing
policy or programmes, to the catch-all politics of the Blair era and beyond.
Modern reference groups are mainly neutral, such as the now familiar ‘hard-
working families’. Of course, this is a way to avoid supporting or opposing any
particular social group.

Party Personnel

Signals can be sent by the things politicians and parties say and do, but they can
also be sent by the background of politicians themselves. The social similarity
between a social group and their representatives has become known as descrip-
tive representation (Pitkin 1967). In principle a man can share the concerns and
values of a woman, a middle class person those of a working class person, and a
white person those of a non-white person, but there is likely to be a connection
between socially similar representatives and their voters with respect to shared

126
 The Parties

values and experiences. Political scientists have argued that ‘demographic facts
provide a low-information shortcut to estimating a candidate’s policy
preferences . . . characteristics such as a candidate’s race, ethnicity, religion, gen-
der, and local ties are important cues because the voter observes the relationship
between these traits and real-life behavior as part of his daily experience’
(Popkin 1991, p.63). So, whatever leaders proclaim publicly, voters might
think it likely that their social character will dispose them to ‘look after their
own’. In short, social dissimilarity from politicians is expected to reduce the
expected benefit of voting for them (Cutler 2002).
Consistent with these arguments, there is a body of empirical research,
mainly in the USA, which finds that people prefer candidates or leaders who
are the same race or gender.6 Oliver Heath also argues that in Britain lower
numbers of working class Labour MPs have depressed Labour voting among
working class voters, reducing class voting (Heath 2015) and increasing
abstention among the working class (Heath 2016). When Campbell and
Cowley (2014) conducted a survey experiment on how people reacted to
hypothetical candidates with differing levels of wealth, they found that
working class people were indeed more negative about wealthy candidates,
particularly with respect to a candidate’s perceived ‘approachability’.7
At the highest echelons of politics we can easily see that descriptive repre-
sentation by class is not the case. Over half of the 2015 Conservative cabinet
was privately educated.8 For Labour, politicians from working class back-
grounds traditionally played prominent roles: think of Ernest Bevin and
Aneurin Bevan in the post-war years. This has changed over time, however.
Whereas half of the 1945 Labour cabinet had previously held working class
jobs, when Labour entered office in 1997 there was just one cabinet minister
who previously had a working class occupation (John Prescott).
If the social make up of parties matters, then it is clearly important to examine
what has happened to descriptive representation by class over the long-term.
Fortunately, we can analyse how the social composition of the Labour and
Conservative parties has changed over the last half century or so. Figure 6.8
shows the percentage of Labour and Conservative MPs who held working class
jobs before being elected. We have data on this back to the 1959 election.9
By far the most noticeable feature of the figure is the very gradual nature of
the decline of working class Labour MPs between 1959 and 1992—a fall, yes,
but of barely more than 10 per cent over more than thirty years. The number
then fell this much again in the space of just one electoral cycle between 1992
and 1997. Labour’s 1997 landslide brought a raft of new MPs into Parliament
and these MPs were middle class. The media attention at the time may have
been on ‘Blair’s babes’ in recognition of the number of women now in the
Parliamentary Labour Party, but the dramatic increase was really in the pro-
portion of people who had held middle class jobs.

127
The New Politics of Class

40% Labour

20%

Conservatives
0%
1945 1965 1985 2005

Figure 6.8. MPs who previously had working class jobs

Note: The figure here shows the proportion of MPs who previously had a working class job before entering
Parliament. These are jobs that fall within our foremen and working class categories.
Source: Datacube collated by the EurElite network (Best and Cotta 2000; Cotta and Best 2007). Updated with figures taken
from the British General Election book series published after each election.

Trade unions were historically an important route of access into Labour

politics by the working class. To examine more detailed information about
MPs we can use evidence from the Parliamentary Candidates UK Dataset.10
This shows that while around 10 per cent of Labour MPs were previously
union officials from the 1950s to the 1980s, by 2015 only 1 per cent of Labour
MPs had worked for a trade union. The opposite story is the case for university
education. Among Labour MPs in 1959 only 41 per cent had a university
education, compared with around 80 per cent by 2010. As Figure 6.9 shows,
the gap between Labour and Conservative MPs in university education has now
been completely closed. In fact, it had closed by the 1990s. Since then both
parties have increased their proportion of university-educated MPs in parallel.
These changes should not be taken to indicate that politics has been taken
over by a socially exclusive elite, however. If we look at more fine-grained
information on where MPs went to school and university as shown in
Figure 6.10, we see that the two main parties have converged. Conservative
MPs have become noticeably less likely to have attended a private school, and
Labour MPs have changed little in this respect. Similarly, Conservative MPs
have become less likely to have attended Oxford or Cambridge, while Labour
MPs have become more likely to have done so. Interestingly, therefore, as
Labour MPs have become more professionalized, the social composition of the
Conservatives has become less exclusive and more similar to that of the
Labour party. In terms of educational exclusivity in particular, the two parties
have become more alike.

128
 The Parties

100%

80% Conservatives

60%

40%
Labour

20%

0%
1945 1965 1985 2005

Figure 6.9. MPs who went to university

Note: The figure here shows the proportion of MPs who have a university first degree or higher degree.
Source: Parliamentary Candidates UK Dataset.

a) Attended Oxford or Cambridge b) Educated at a private school

100% 100%

Conservatives
80% 80%

60% Conservatives 60%

40% 40%

Labour Labour
20% 20%

0% 0%
1945 1965 1985 2005 1945 1965 1985 2005

Figure 6.10. MPs who attended Oxbridge or a private school

Note: The left-hand graph here shows the proportion of MPs who went to the universities of Oxford or Cambridge.
The right-hand graph shows the proportion of MPs who were educated at a private school.
Source: Parliamentary Candidates UK Dataset.

129
The New Politics of Class

Former Conservative Chancellor, and Home Secretary, Kenneth Clarke

recently described his memories of the Commons when he was first elected
in 1970 as ‘old landed gentry with rolling estates’ on one side and ‘retired trade
union regional secretaries and 30-odd miners’ on the other (Independent, 28
July 2014). This does not fit the reality today. We now see a preponderance
of MPs who reach politics via university and jobs in the professions and
management. The influx of women MPs has reduced the discrepancy between
parties and their voters descriptively, but there has been a growing discrep-
ancy between Labour and its working class supporters. This trend in part
reflects the shrinking of the traditional working class as a recruitment base,
but it is also because party recruitment procedures are very different now than
in the mid-twentieth century (Norris and Lovenduski 1995). The contempor-
ary trend is for politicians to enter politics through working as political interns
or members of lobby groups. In 2015 18 per cent of MPs had moved into
parliament from other directly political jobs and another 12 per cent had
come from a different elected office (van Heerde-Hudson and Campbell
2015). This means that an increasing number of MPs spend their entire
adult lives in politics or on its fringes.

Has Anyone Noticed?

The evidence of party change is pretty clear-cut, but the perception of

such signals by ordinary voters cannot be assumed. After all, politics is for
most people ‘a side-show in the circus of life’ (Key 1966). Did voters notice
any of the changes to the parties over the last fifty years? To assess changes
in voters’ perceptions of party platforms on left–right issues, we examine
their perceptions of party polarization. We do this using a question in the
BES since 1964, and the BSA since 1997, that asks people: ‘Considering every-
thing the Conservative and Labour parties stand for, would you say there is
a great deal of difference between the parties, some difference, or not
much difference?’11
Although there is no explicit reference to economic left–right policy in this
question, it seems reasonable to assume that voters are largely reacting to this
dominant axis of political competition. As can be seen from Figure 6.11, the
proportions of people who perceive a ‘great deal’ of difference between the
parties in each election and the difference between the Labour and Conserva-
tive parties in left–right manifesto positions are connected. In particular, we
see the impact of the brief convergence in the late 1960s/early 1970s and the
sharp, and lasting, convergence that occurred between the 1987 and 1997
elections when Labour moved to its much more centrist position and
rebranded itself as New Labour.

130
 The Parties

100%

80%

60%

40%

20%

0%
1960 1970 1980 1990 2000 2010

Figure 6.11. Perceived differences between the parties

Note: The figure here shows the percentage of people who believe there is a ‘great deal of difference’ between the
Conservative and Labour parties at each election from 1964 to 2015.
Source: British Election Studies 1964–2015; British Social Attitudes Surveys 1997–2015.

If the parties are perceived to have converged in recent years, have the
voters’ perceptions of the degree to which they represent the working and
middle classes changed accordingly? Both our analysis of the changes in group
references and the decline in the number of working class Labour MPs might
suggest that Labour is now more likely to be seen by the electorate as repre-
senting the interests of the middle class. We can test this using some questions
on class representation by the main political parties that were introduced in
the 1987 BES. These ask people how closely they think that the two main
parties look after the interests of middle class and working class people. People
can answer from ‘not at all closely’ to ‘very closely’. We use these questions to
create a measure of whether people think the parties look after working class
people better than middle class people or vice versa. To construct this scale we
scored the response categories as 1 (not at all closely) to 4 (very closely) and
subtracted responses to the middle class question from those to the working
class question. This gives a scale that runs from
3 to 3 for each party. As very
few people have scores of +3 or
3, we combine
2 and
3 and +2 and +3
responses. A positive score means that people think that a party looks after the
working class better than the middle class. Negative scores mean that people
think a party looks after the middle class better than the working class.
Figure 6.12 shows how these perceptions have changed since 1987. We can
see that by 1997 Labour’s connection with the interests of the working class
had dropped dramatically. The policy of reaching out across class boundaries
had worked in the eyes of the electorate. Indeed it had worked so well that by

131
The New Politics of Class

More Labour
working
class 0.5

0
1985 1995 2005 2015

More –0.5
middle
Conservatives
class

–1

Figure 6.12. Perceptions of the extent to which parties look after the interests of classes
Note: The figure here shows people’s perceptions of which classes the parties represent. A positive score means that
people believe a party looks after the working class better than the middle class; a negative score means that people think
a party looks after the middle class better than the working class.
Source: British Election Studies 1987–2015.

2001 the middle class was actually thought to be better represented by Labour
than was the working class.12 The changes in perceptions of the party after the
Blair years blur this picture a little, but there has been no return to anywhere
near the former level of perceived association between party and class. Per-
ceptions of the Conservatives are much more stable over time. They are
consistently seen as more likely to represent the interests of the middle class
than the working class.
We can also take an even longer-term perspective on perceptions of working
class representation by the parties. Although we do not have an identical
question from the 1960s to the ones used above, in the 1964 BES people
were asked to choose whether Labour ‘is a working class or middle class
party’. This question was repeated in the 2015 BES. In 1964 90 per cent said
that the Conservatives were a middle class party and 85 per cent of people said
that Labour was a working class party. In 2015 88 per cent still said that the
Conservatives were a middle class party, but only 38 per cent said that Labour
was a working class party. In fact, in 2015 more people thought that Labour
was middle class (48 per cent) than working class.13 Like the USA, Britain’s
party system is now dominated not by two class parties, but by two middle
class parties.
It seems likely that the perception of the parties as class parties is shaped by
several factors. One of these is the degree to which group rhetoric features in
speeches, but another is what the party looks like in terms of elites. If MPs
appear indistinguishable from one another in terms of class background, then

132
 The Parties

voters are likely to make a much weaker connection between parties and class
groups. As mentioned earlier, Heath (2015) makes exactly this argument. Thus
some of the class-party perceptions that we look at here may be driven by a
number of ‘objective’ factors about the parties. The specific role of descriptive
representation can be looked at more closely, however. The 1966 BES survey
asked respondents about their local MPs and specifically whether they ‘would
say that he/she is upper class, middle class, or working class?’. We repeated
this question after the 2015 election. Of course, both surveys cover only a
fraction of the constituencies and so we cannot say anything about general
perceptions of MPs. What we can do is link people’s responses to characteris-
tics of their actual MP, as well as linking those MP’s characteristics to their
general view of the parties.
What would we expect? There are two possibilities. The first is that people
today make a weaker link between their MP’s occupational background and
their class, and they also make a weaker link between their MP’s class and
their general perceptions of the party. The second is that these processes
have remained the same; it is simply that the background of MPs has
changed. It appears that the second of these fits better with the evidence.
It is the parties that have changed, not the way in which people form
political opinions.
Table 6.1 shows predicted probabilities from a multinomial multilevel logit
regression model that predicts the class to which people think their MP
belongs for the 1966 and 2015 surveys. Respondents are also asked, before-
hand, to identify the party of their MP and we just look at people who can
correctly identify their MP’s party.14 We use both the party of the MP and
two key social characteristics of the MP: their educational background and
their occupational social class measured by their previous job.15 Both of

Table 6.1. The proportion of people who think their MP is working class by party and social
background of MP

1966 2015

Conservative MP Elite background 0% 4%

Working class background 1% 24%
Labour MP Elite background 5% 19%
Working class background 44% 63%

Note: The numbers here are predicted probabilities from multilevel multinomial logit models that predict whether people
think their MP is working class (as opposed to upper or middle class). These models include measures of the local MP’s
party affiliation and educational/occupational characteristics and are only run for people who know the party of their
MP. Two types of constituency are displayed, split by party. The first are constituencies with an ‘elite MP’: elite MPs
previously had managerial jobs and attended Oxford or Cambridge. The second are constituencies with a ‘working class
MP’: working class MPs previously had manual jobs and did not attend university. The models that we use here are
multilevel and thus account for the fact that people are clustered by constituency. In 1966 we have data on 68
constituencies with 618 individual voter responses spread across those constituencies. In 2015 we have data on 261
constituencies and 1,387 individual voters.
Source: British Election Studies 1966 and 2015.

133
The New Politics of Class

these factors are strong predictors of how people perceive the class of their
MP. Table 6.1 shows the predicted probability of someone placing their MP
into different classes for two types of MP by party. The first MP type is
someone who was formerly a manager and went to Oxbridge (an MP from
the ‘elite’). The second MP type is someone who previously had a working
class job and had not been to university (a working class MP).
In 1966 44 per cent of people with a Labour MP who previously had a
working class job and had not been to university thought that he (and all
the Labour constituencies included in the 1966 survey had male MPs) was
working class compared to only 5 per cent of people with a Labour MP from
a managerial background with a degree from Oxbridge. As the figure shows,
that difference is almost exactly the same in 2015. The MP’s occupational
and educational background is clearly telling people something about that
MP’s social class, and this has not really changed over time. Indeed, if
anything, people appear to make stronger connections in 2015 between
the characteristics of their MP and their social class. The alleged increased
concentration on candidates in British elections and the increased levels of
constituency work that MPs undertake (Norton 1994) may have strength-
ened the ability of voters to make these judgements. Nonetheless, the bigger
picture is one of stasis.
We can also link the characteristics of MPs more directly to the perceptions
of parties. Part of the reason that people think that Labour is less representa-
tive of the working class may be due to the changing nature of descriptive
representation. If so, then we should find that people with working class
Labour MPs are more likely to view Labour as a working class party. This also
appears to be the case. Figure 6.13 shows the results of models run with
the 1966 and 2015 data predicting whether people think Labour is very or
somewhat working class for people who were in a constituency with a Labour
MP.16 Again we use the MP’s former occupation and educational background
as predictors of these perceptions.
Figure 6.13 shows the differences between people who had a Labour MP
who was formerly a manager and went to Oxbridge (an MP from the elite)
versus a MP who previously had a working class job and had not been to
university (a working class MP). These differences are both large and constant.
In 1966 nearly 70 per cent of people with a working class Labour MP saw
Labour as a working class party compared to fewer than 40 per cent of people
with an elite Labour MP. By 2015 both these figures had dropped, due partly of
course to the changes that we have already discussed in this chapter, but the
gap remains. Nearly 40 per cent of people with a working class Labour MP in
2015 still thought that Labour was a working class party, compared with only
15 per cent of those with an elite Labour MP.

134
 The Parties

1966 2015

60% 60%

40% Elite MP 40%

20% 20% Elite MP

0% 0%
WC MP WC MP

–20% –20%
Difference
Difference
–40% –40%

Figure 6.13. The proportion of people in Labour constituencies who agree that Labour
is somewhat/very working class given the social characteristics of their MP
Note: These figures show predicted probabilities from multilevel logit models that predict whether people agree
that Labour is a somewhat or very working class party. The models include measures of the local MP’s educational
and occupational characteristics and are only run for people who live in a Labour constituency. The darker bar in
both graphs shows people in constituencies with a ‘working class MP’: working class MPs previously had manual
jobs and did not attend university. The lighter bar in both graphs shows people in constituencies with an ‘elite
MP’: elite MPs previously had managerial jobs and attended Oxford or Cambridge. The models that we use here are
multilevel and account for the fact that people are clustered by constituency. In 1966 we have data on 43
constituencies with 453 individual voter responses spread across those constituencies. In 2015 we have data on
161 constituencies and 762 individual voters.
Source: British Election Studies 1966 and 2015.

This helps us to understand how perceptions of Labour have changed. Sixty-

five per cent of people, in all constituencies, thought that Labour was a very or
somewhat working class party in 1966 (a further 11 per cent said it was slightly
working class). By 2015 just 23 per cent said that it was very or fairly working
class (with a further 15 per cent saying it was slightly working class). At least
part of the explanation for that change is that the direct cues that sitting
Labour MPs give to constituents have altered. Using the figures above, if half
of Labour MPs had remained working class as was the case fifty years ago, that
decline would have shrunk by about a third (for at least those people in Labour
constituencies and plausibly beyond). Descriptive representation does
not explain everything as Heath (2015) alleges, but it is an important part of
the story.

Conclusions

In Chapter 5 we argued that media discussion of class appeared to fall into

three broad periods. There was a period of consensus around a dominant

135
The New Politics of Class

working class, a breakdown of that consensus over the 1970s and 1980s and
then a new consensus from the 1990s onwards in which the middle class was
dominant. In this chapter we see some echoes of that pattern of change in at
least policy. While we should not exaggerate the degree of policy consensus
that existed for the first twenty-five years after the war, there were points at
which the major parties held relatively similar economic positions. This was
driven by parties trying to appeal to the ordinary working man or woman.
They had to do so if they were going to win elections. As noted in Chapter 1,
the emergence of the welfare state, the extension of the public sector and the
adoption of a Keynesian approach to minimizing unemployment were all
implemented in the name of the working class. These policy differences widened
during the 1970s and 1980s, and then converged in the 1990s onwards with
the emergence of a new consensus.
Policy differentiation along left–right lines is only one form of interest
representation, however. Parties can signal interest representation through
direct references or appeals to relevant groups, via both policy and rhetoric.
We have seen that references to the working class were particularly pro-
nounced by Labour in the post-war era, but started to fall dramatically in
both manifestos and leaders’ speeches from the late 1980s onwards. Political
appeals to the working class have now effectively disappeared from the lexi-
con of party politics. Just as discussion of class and class politics by the
newspapers more or less stops by 1997, so do appeals to the working class by
the political parties.
This picture is repeated when we look at party elites. Since the 1990s
recruitment has been dominated on both sides by middle class people who
are highly educated. The route from a working class job into politics, often
through trade union activism, has effectively disappeared, and almost identi-
cal proportions of MPs in both parties now have a university background.
We have seen the two parties growing more alike: in policy, in class appeal,
and in their patterns of recruitment by education and social class. Import-
antly, these different types of changes coincided most strongly during the
1990s. The earlier, and rather brief, period of convergence in the 1960s only
applied to policy positions, not to the class-related appeals, rhetoric, and the
personnel of the parties. Growing cross-party uniformity across a broad range
of indicators only arrived in the early 1990s. Significantly, the electorate has in
turn noticed these changes.
As a result of these combined transformations, there is now little connec-
tion between policy, party, and class in mainstream British politics. The
disappearance of class from both media and party messages leaves few class-
based cues for voters. In Chapters 7 and 8 we explore the consequences of this
change for how people choose to vote.

136
 The Parties

Notes

1. There are of course numerous commentaries examining the Blair effect and New
Labour’s transformation which also qualitatively indicate this change (King 1997;
Seyd 1997; Seldon 2007; Fielding 2003).
2. Rehm and Reilly (2010) provide details of most of the expert surveys used here (but
also see Castles and Mair 1984; Laver and Hunt 1992; Steenbergen and Marks 2007;
Hooghe et al. 2010; Bakker et al. 2015). Unfortunately, this approach for measuring
party positions did not start until 1982 so cannot be used to measure positions
before the 1979 election.
3. For example, for workers the coding dictionary is composed of three terms:
‘worker’, ‘miner’, and ‘working class’. The Lexicoder coding system also includes
preceding words, so ‘ordinary workers’ and ‘factory workers’ are both coded as
‘workers’. For ‘the poor’ the stem is just ‘poor’ (so ‘poor’, ‘the poorest’, ‘poor
people’, etc.). See the Appendix for more details of the coding system.
4. In 1966, for example, Labour proclaimed that ‘The level of economic activity in the
community must be sufficient to provide jobs for all’. The idea of ‘jobs for all’ was a
dominant theme of many Labour manifestos before the 1990s.
5. In 1999 William Hague was moved to claim: ‘“The class war is over”, says Tony Blair. Tell
that to thousands of vindictive, mean-spirited, class-obsessed Labour party activists’ But
notice that his reference is not to the Parliamentary Labour Party, but Labour activists.
6. On the gender of candidates see: Tolleson Rinehart (1992); Huddy and Terkildsen
(1993); Bendyna and Lake (1994); Huddy (1994); Cook (1994); Plutzer and Zipp
(1996); Dolan (1998); Campbell et al. (2010); Childs and Webb (2010). On the race
of candidates see: Tate (1993); Terkildsen (1993); Sigelman et al. (1995).
7. We do not address here the question of whether parliamentarians’ own class back-
grounds matter for policy. The most significant research on this has been done by
Carnes (2012) in the US. He shows that, all things being equal, legislators from
working class occupational backgrounds have more left-wing economic preferences.
8. Although the incoming 2015 cabinet was still noticeably less socially exclusive
than it was in the 1950s. Ten of Eden’s fourteen ministers were related to aristo-
cratic families, thirty-five out of eighty-five of Macmillan’s government were
relatives following his marriage to a Duke’s daughter, and half of Douglas-Home’s
cabinet were Etonians (Nordlinger 1967: 41).
9. These data are taken from those collected by the datacube project (Best and Cotta
2000; Cotta and Best 2007) with additional information presented in the British
General Election books published after each election. Similar figures up until 2010
are presented in Heath (2015).
10. For more information see http://www.parliamentarycandidates.org.
11. This question changes very slightly over time. Before 1974 the first option is a
‘good’ rather than a ‘great’ deal of difference, and before 1979 the wording is
‘Considering everything the parties stand for, would you say there is a great deal
of difference between them, some difference, or not much difference?’.
12. An interesting contrast is with the degree to which people think that Labour
represents the interests of ‘black and Asian people’. The proportion of the

137
The New Politics of Class

population who believed this increased to 90 per cent during the first decade of this
century before falling back to just under 80 per cent in 2015; still an extremely high
level. There appears to be a trade-off in the extent to which the party represents the
working class versus how closely it represents ethnic minorities: when one goes up
the other goes down. This may relate to the party’s endorsement of diversity and
affirmative action policies at the perceived expense of groups such as the working
class (Barry 2001, pp.11–12) with, according to some, Labour having ‘learned to
love identity and ignore inequality’ (Michaels 2006).
13. Gavin (1996) shows that people’s answers to open-ended questions about why
they voted for Labour were just as focused on class in 1987 as they were in the
1960s, suggesting that this change in perceptions did not happen before the
1987–2015 data shown in Figure 6.12.
14. The proportions of people who are able to do this are not completely comparable
between the 1966 and 2015 surveys because of question-wording effects regarding
‘don’t knows’, question-ordering effects (whether people were asked the name of
their MP first), and to some extent timing effects. In the 1966 survey, which took
place directly after the election, 90 per cent could accurately identify their MP’s party.
For 2015, for which the fieldwork continued for almost six months after the election,
these rates are lower: 60 per cent of people in England and Wales accurately named
their MP’s party. The 2015 SNP landslide reduced this accuracy rate to 22 per cent in
Scotland, with many people not realizing that their previous Labour MP was no
longer in place. However, because the question on MPs’ social class in the 1960s was
preceded by a filter question (‘Have you heard anything about your MP?’) we actually
end with similar rates of response to class questions for both surveys. For example, in
1966 90 per cent of people knew the party of their MP, but only 50 per cent of them
had ‘heard anything’ about their MP and therefore 45 per cent of people assessed
their MP’s social class (with 8 per cent saying ‘don’t know’). In 2015 56 per cent of
people correctly identified their MP’s party, but all of them gave a response to the
question about their MP’s social class (with 32 per cent saying ‘don’t know’). This
means that overall we end up with about 40 per cent of respondents in both 1966
and 2015 giving an upper, middle, or working class designation to their MP.
15. Education is categorized into three groups: no degree, non-Oxbridge degree, and
Oxbridge degree. This divides the 1966 MPs roughly into three. The occupational
class groupings are not exactly the same as the one we have used before for voters,
but are similar, apart from more differentiation within the middle class and less
differentiation within the working class. This gives us five groups: managers;
employers including farmers; professionals; intermediate ancillary and supervisory
workers and any junior non-manual workers; working class which includes personal
service workers and foremen/technicians. MPs previously in the armed forces are
coded as managers. The models that we use here are multilevel and thus account for
the fact that people are clustered by constituency. In 1966 we have data on 68
constituencies (with 616 individual voter responses spread across those constituen-
cies). In 2015 we have data on 261 constituencies and 1,387 individual voters.
16. This question is the same as the seven category question discussed earlier which
asks people, ‘how middle class or working class are the main political parties?’ and
allows them to answer from ‘very middle class’ to ‘very working class’.

138
 The Parties

Appendix to Chapter 6

Further Information on the Coding of Party Manifestos and

Leaders’ Speeches

Manifestos
The text of the manifestos came from http://www.politicsresources.net/area/uk/
man.htm and the manifestos were also cross-checked manually. Numerous mani-
festo components are included in the manifesto left–right scale. These are shown in
Table A6.1.
Responses for each of these categories are then coded to create a uni-dimensional
left–right scale. The scale was derived by an inductive process that begins from a ‘naïve’
starting scale (based on earlier scales). The scale is then constructed using Lowe et al.’s
(2011) logit scaling technique, using the formula:

R þ 0:5
θ L ¼ log
L þ 0:5

where R is the total number of quasi-sentences in the manifesto components on the

‘right’ of the scale and L is the total number of quasi-sentences in the manifesto

Table A6.1. Manifesto components included in the left–right scale

Left Right

105 Military: Negative 109 Internationalism: Negative

106 Peace 401 Free Enterprise: Positive
107 Internationalism: Positive 407 Protectionism: Negative
202 Democracy 414 Economic Orthodoxy: Positive
301 Decentralization 505 Welfare State Limitation: Positive
303 Governmental and Administrative Efficiency 507 Education Limitation: Positive
403 Market Regulation: Positive 601 National Way of Life: Positive
408 Economic Goals 603 Traditional Morality: Positive
411 Technology and Infrastructure 608 Multiculturalism: Negative
412 Controlled Economy: Positive 702 Labour Groups: Negative
413 Nationalization: Positive
416 Anti-Growth Economy
501 Environmental Protection
502 Culture
503 Social Justice
504 Welfare State Expansion: Positive
506 Education Expansion: Positive
602 National Way of Life: Negative
604 Traditional Morality: Negative
701 Labour Groups: Positive
705 Underprivileged Minority Groups
706 Non-economic Demographic Groups

139
The New Politics of Class

components on the ‘left’. The logit scaling method combines the advantages of both
additive and ratio-scaling methods for manifesto data, while avoiding the problem of
polarization found in ratio scales. It also has an additional benefit of a diminishing
impact of repeated emphasis and it therefore mirrors natural language usage. For ease of
interpretation the scale is rescaled as follows:1
100
Rescaled dimension ¼ ðScale position
Scale mean þ 7Þ 
14
The robustness of this interpretation was checked using the hand-coded policy estima-
tions of these party manifestos provided by Ian Budge, Judith Bara, and colleagues in
the Comparative Manifesto Project (Budge et al. 2001; Volkens et al. 2015). These
authors also provide an estimate of party positions on a left–right scale—what they
refer to as RiLe.2 As can be seen from Figure A6.1, the pattern identified by the RiLe scale
mirrors those identified by our own coding procedures.
RiLe converges more quickly in the 1950s and there is a greater spike of polarization
in 1974. It does not converge completely in the 1990s but the parties get very close in
1997 and remain so. On the whole it confirms the validity of the procedure used here.
For more information on this scale, see Prosser (2014). Prosser’s coding is preferred to
RiLe, not just on the basis of face validity but because Prosser’s scales make fewer
assumptions about what categories ‘should’ go on the left or right. To give one
example that changes sides, RiLe has ‘Freedom and Human Rights’ on the right-
hand side (presumably because classical liberal parties will talk about individual
freedoms) but left parties actually talk much more about things that get coded in
that category (which is unsurprising). In the RiLe scale talking about human rights
makes a party more right-wing; in Prosser’s scale it makes a party more left-wing.
Estimates of the reliability of Prosser’s left–right scale, using the method that the
authors of the RiLe scale themselves propose, find that it is more reliable than RiLe
(McDonald and Budge 2014).

Leaders’ Speech Sources

The group references in the leaders’ speeches are coded using the same dictionary
developed for the manifestos. The speeches are taken from the ‘British Political Speech
Archive’ (http://www.britishpoliticalspeech.org/speech-archive.htm) collated by Alan
Finlayson and Judi Atkins (2015). Finlayson and Atkins kindly sent us the Churchill
speeches, which are not posted on the website. Although there are Labour speeches
from 1945 to 1951, there are no Labour leader’s speeches from 1952 to 1964 in the
archive. This was because there was no ‘leader speech’ at those Labour party

1
The logit scale does not have a natural midpoint or endpoint and so the mean of each scale is
subtracted from the scale (which gives a midpoint and mean of 50 once rescaled). In practice, the
endpoint of each scale approaches 7, suggesting the rescaling formula used here.
2
There have been recent debates about the reliability and validity of the CMP measures of party
positions and the RiLe left–right scale (Laver and Garry 2000; Armstrong and Bakker 2006; Benoit
and Laver 2007; Bakker and Hobolt 2013).

140
 The Parties

40
Conservatives
30

20

10

0
1945 1965 1985 2005
–10

–20
Labour
–30

–40

Figure A6.1. The comparative manifesto project estimate of left–right party positions
(RiLe)
Note: The figures here show the positions of the two main parties on the RiLe left–right policy scale (three period
moving average); higher scores indicate more right-wing positions.
Source: Manifesto Project on Political Representation 1945–2015.

conferences. However, between 1952 and 1964 the leader almost always spoke in
various debates and was unsurprisingly given a prominent place. For those years we
obtained the Labour party records held in the Bodleian Library and scanned each of the
speeches. These were then turned into OCR files. This yielded several extra speeches
(1954, 1958, 1963, and 1964).

Coding Procedure for Group Appeals in Manifestos and Leaders’

Speeches
Coding up the group references was a multi-step process. The first stage was conducted
on a subset of the full corpus, using the speeches from the party conference before each
general election.

1) Each speech transcript was read and any reference to a group was coded as a group
reference.
2) A group classification scheme was developed inductively.
3) Using the manual group coding, a dictionary of words identifying groups was
constructed.

The second stage was conducted on the full corpus.

1) Using the dictionary constructed in stage 4, the full corpus was coded using
Lexicoder 2.0 (http://www.lexicoder.com/).

141
The New Politics of Class

2) The fact that some group speeches were coded both manually and by Lexicoder
means we obtained an estimate of how reliable the auto-coding is. For the thirty-
four speeches that are double-coded, the average correlation between the coding
for all categories is an exceptionally high 0.91. But it is actually substantially
higher for the ones we are interested in (e.g. workers = 0.98, families = 0.99).
The average is brought down by some groups that we are less interested in
(e.g. farmers = 0.66 and the elderly = 0.65).

142
Part III
Consequences
7

Class Politics Is Dead

The day before the 1964 election, the then leader of the Labour party
Harold Wilson went to see Hughie Greene, the then Director General of
the BBC. On his mind, unsurprisingly, was the upcoming general election.
More surprisingly, also on his mind was the sitcom ‘Steptoe and Son’.
Greene said that Wilson ‘had been very much upset because the BBC had
planned the beginning of a series of repeats of a very popular light enter-
tainment programme, “Steptoe and Son”, on the evening of polling day.
He thought that would keep away particularly Labour supporters from the
polls’. Working class people voted Labour and working class people
watched ‘Steptoe and Son’. Indeed Wilson’s comments on this are almost
a parody; he claimed that having ‘Steptoe and Son’ on at 9 p.m. meant
that people would not get to the polling station. That for ‘a lot of our
people—my people, working in Liverpool, long journey out, perhaps then
a high tea and so on, it was getting late, especially if they wanted to have a
pint first’.1
If anything sums up post-war assumptions about class and electoral behav-
iour, it is this anecdote. Class was at the centre of British politics in 1964.
In this chapter, and in Chapter 8, we explore the consequences of the
political changes described in Chapters 5 and 6 for class and voting. In
this chapter we show that class is no longer at the centre of politics, and
that class has become a marginal force within mainstream politics. We
concentrate on not just the way the link between class and party choice
has changed but also the reasons for this change. We show that there is little
evidence of class divisions in party support narrowing before the 1990s.
During the 1990s, however, there was a striking dealignment between
classes and parties. The reasons for this are mixed, but clearly relate to the
changes we have described in Chapters 5 and 6: changes in party policy and
rhetoric, changes to politicians, and changes to the media environment.
The decline of mainstream class politics is a consequence of the decline of
mainstream class parties.
The New Politics of Class

The Decline of Class Voting?

There is of course nothing novel about suggesting that class has declined as a
political force. Any article or book on British political history will at some
point refer to this as a simple statement of fact. For example, Clarke et al., in
Political Choice in Britain, conclude that ‘at the end of the twentieth century
class had come to play a very limited role in determining the voting prefer-
ences of the British electorate’ (2004, p.50). In fact, tales of this decline date
back to some of the first books and articles to look at how class shaped vote
choices in Britain. By the time Butler and Stokes published the second
edition of their study of the 1960s elections in 1974, they were already
talking of the ‘weakening of the class alignment’ (Butler and Stokes 1974,
p.208) and others argued that Butler and Stokes’ data, particularly the study
after the 1970 election, showed that class voting was waning (Books and
Reynolds 1975). This was followed in the early 1980s by an important book
authored by the new team in charge of the BES arguing that ‘class voting is
on the decline and has already reached modest levels’ (Sarlvik and Crewe
1983, p.91). Franklin (1985), Robertson (1984), and Rose and McAllister
(1986) made similar points in three other influential books of the period.
It seemed at the time that there was no doubt class voting was, if not dead, at
least dying, that the ‘old class-equals-party model of politics is no more’
(Rose and McAllister 1986, p.1).
This consensus led to a ‘revisionist’ challenge which emphasized stability
rather than change, and it was this perspective that dominated debate during
the late 1980s and 1990s. Anthony Heath and colleagues consistently argued
that the death of class voting had been exaggerated and coined the expression
‘trendless fluctuation’ to describe levels of class voting up to the mid 1980s
(Heath et al. 1985, p.35; see also Weakliem 1989). They pointed to problems
with data and measurement in previous work, arguing that 1964 (the first BES)
showed an unusually high level of class voting and was therefore an unhelpful
baseline against which to measure change and that the common manual/non-
manual distinction was an overly crude way of capturing class. Much of the
initial critical response to Heath et al. (1985, 1991, 1994) was focused on
measurement (Dunleavy 1987; Crewe 1986), but whether the measures were
problematic or not was largely seen as a moot point a decade or two later.
Why? Because by then, a new consensus had emerged around the fact that
class voting by the end of the 1990s and certainly by the 2000s was definitely
lower than it had been at some rather vaguely specified point in the past.2 Yet,
while the decline of class voting in Britain is now widely accepted, and as we
will proceed to show, well evidenced, the pattern of change and the reasons
for that change remain much less clear.3

146
 Class Politics Is Dead

Most accounts of decline are shaped by the idea that classes are more
similar; that there has been a ‘loosening’ of social structures (Butler and
Kavanagh 1984, p.8), an attenuation of ‘bonds linking voters to politically
relevant class groups’ (Dalton 2008, p.157) or a ‘weakening of class stratifica-
tion, especially as shown in distinct class differentiated life-styles’ (Clark and
Lipset 1991, p.408). These arguments reverberate all the way back to the
1950s. After the 1959 Conservative election victory, there were predictions
of Labour’s inevitable decline due to the ‘the growing homogeneity of styles of
life of workers and middle classes, and the continuing prosperity of Britain’
(Alford 1964, p.127) or the fact that the ‘old working class ethos is being
eroded by prosperity and the increasing fluidity of society’ (Abrams and Rose
1960, p.106).
The basic assumption was that classes, and particularly the working class,
had lost their social cohesion and distinctiveness. Of course, this is not what
we have seen in previous chapters. Neither the objective realities of class,
nor the political attitudes that these objective realities produce, have
changed greatly. This does not, therefore, seem like a promising explanation
for change. Rather, we argue here, as we have previously argued (Evans and
Tilley 2012a, 2012b, 2013), that it is political change which produces
class voting change. This perspective emphasizes the ‘top down’ structuring
of cleavages by the actions of political parties, in Przeworski’s words:
‘individual voting behaviour is an effect of the activities of political parties’
(1985, pp.100–1). Robertson (1984, p.225) ends his book by noting that
changes to class voting in Britain are essentially ‘changes inside the clay
round which the mould fits. The clay is what it always has been, and
modifications to the Conservative and Labour parties ought to suffice them
as moulders for a long time to come’. Less poetically, if parties look very
different, then classes vote differently; if parties look very similar, then
classes vote similarly. As we will see, the pattern of change strongly supports
this view.
Before focusing too much on the ‘whys’, we need to establish the ‘what’.
What has changed and when did it change?

Labour and the Conservatives

Assessing whether, and more importantly when, the link between class and
vote choice changed is not quite as straightforward as it might seem. We
need consistent measures of class over long periods of time, consistent
measures of party choice over long periods of time, and long-running survey
series that allow us to track change. Unfortunately, no single survey series
spans the period since the Second World War in Britain. We therefore use

147
The New Politics of Class

data from three different overlapping sources in this chapter: Gallup surveys
from 1945 to 1968, BES surveys from 1964 to 2015, and BSA surveys from
1983 to 2015. We have discussed the BES and BSA data previously; suffice to
say that for the post-election BES surveys we use vote choice at the election,
but since the BSA surveys are yearly, we use a broader measure of party
choice. This uses three questions, as below, to identify which party someone
supports.

‘Generally speaking, do you think of yourself as a supporter of any one political

party?’
[IF NO] ‘Do you think of yourself as a little closer to one political party than to others?’
[IF NO] ‘If there were a general election tomorrow, which political party do you
think you would be most likely to support?’

The Gallup surveys allow us to track class voting back to the 1940s, as we can
use three surveys from 1945 and 1946 as well as a run of nearly annual
surveys from 1955 to 1968. The measure of occupational class is extremely
similar to the one that we have used previously with the BSA and BES data: it
is based on occupation and self-employment status and allows us to effect-
ively distinguish at least four groups: new middle class professional workers,
old middle class managers and employers, the junior middle class of clerical
non-manual workers, and manual workers.4 We use recalled vote choice in
election years and vote intention otherwise (‘How would you vote if there
were a general election tomorrow?’).5 In this chapter we focus on the three
main parties, although all vote percentages are proportions of people ques-
tioned so include non-voters and voters for smaller parties. In Chapter 8 we
look at both non-voters and the emergence of class voting for minor parties,
most notably UKIP since 2010 and the SNP in 2015 in Scotland. Here we are
interested in how class matters for picking one of the three main parties and
how this has changed.
Figures 7.1 and 7.2 show changes in vote share for Labour and the Conser-
vatives by the four main occupational class groups: the old middle class, the
new middle class, the junior middle class, and the working class. The Gallup
and BSA graphs present moving averages as the samples are yearly and rather
small. The BES graph shows the raw percentages. There are three very import-
ant points to make. The first is about the lack of change, particularly for the
Conservative vote. It is easy to assume from some of the literature that
occupational class has everything to tell us about party choices in the past,
and nothing to tell us about people’s party choices today. That is incorrect.
There were always working class Conservatives and middle class Labour sup-
porters: Eden and Macmillan’s election victories in 1955 and 1959 were due to

148
 Class Politics Is Dead

a) Gallup data

60%
WC
40%

JMC
20%
NMC
OMC
0%
1945 1960 1975 1990 2005

b) BES data

60%

40%

20%

0%
1945 1960 1975 1990 2005

c) BSA data

60%

40%

20%

0%
1945 1960 1975 1990 2005

Figure 7.1. Labour support by occupational class

Note: The figures here show the proportion of people who support the Labour party as a percentage of the
electorate. The top graph shows vote intentions from Gallup data (three period moving average); the middle
graph shows vote choices from BES data; the bottom graph shows party support from BSA data (three period
moving average). Four occupational class groups are displayed: old middle class (OMC), new middle class (NMC),
junior middle class (JMC), and working class (WC).
Source: Gallup 1945–1968; British Election Studies 1964–2015; British Social Attitudes Surveys 1983–2015.

more than a third of the working class supporting the Conservatives. Equally
there is still a difference in Conservative vote share between the old middle
class and the working class in 2015: 25 per cent for the BES data and 29 per
cent for the BSA data. This should not surprise us given the results from earlier

149
The New Politics of Class

a) Gallup data
OMC
60% NMC

JMC
40%

WC
20%

0%
1945 1960 1975 1990 2005

b) BES data

60%

40%

20%

0%
1945 1960 1975 1990 2005

c) BSA data

60%

40%

20%

0%
1945 1960 1975 1990 2005

Figure 7.2. Conservative support by occupational class

Note: The figures here show the proportion of people who support the Conservative party as a percentage of the
electorate. The top graph shows vote intentions from Gallup data (three period moving average); the middle graph
shows vote choices from BES data; and the bottom graph shows party support from BSA data (three period moving
average). Four occupational class groups are displayed: old middle class (OMC), new middle class (NMC), junior
middle class (JMC), and working class (WC).
Source: Gallup 1945–1968; British Election Studies 1964–2015; British Social Attitudes Surveys 1983–2015.

chapters. As Chapter 4 showed, people’s policy preferences are driven by class

in the same way today as they always have been. Chapters 5 and 6 showed
that while parties are less obviously talking about class and less obviously
different from one another, they are still somewhat different. Occupational

150
 Class Politics Is Dead

class remains a useful predictor of vote choice in Britain and we should not
forget this.
The second point is about the scale of change. Class may still predict vote
choice, but it is much less important than it was. This is most obvious for the
Labour vote. After the war, there was around a 30 percentage point gap
between the old and new middle class groups and the working class in Labour
support. From 1997 onwards that gap was more like 10 percentage points. The
BES data shows that at the most recent election, that difference has completely
disappeared. There can be no doubt that class voting has declined over the
period that we are examining. This change is apparent for both major parties,
but is much starker for Labour.
The third point is perhaps most important, and this is about the nature of
change. What is fascinating is the way in which differences between the
groups alter at particular points in time. Most notably the main change
occurs in the mid-1990s, although for Labour 2015 also appears to be a
critical election. The three middle class groups all sharply move towards
the working class group in their preferences between 1995 and 2000. This
contrasts with only slight changes in the gap between the three middle
class groups over the entire period.6 Figure 7.3 illustrates just how abrupt
the change is by plotting the difference between the three middle class
groups and the working class in their support for Labour. Here we combine

a) OMC versus WC b) NMC versus WC c) JMC versus WC

1945 1975 2005 1945 1975 2005 1945 1975 2005
0% 0% 0%

1997–2014
–8% 1997–2014
–10% 1997–2014 –10% –10% –10%
–13%

–20% –20% –20%

1945–94
–23%
1945–94
–30% –30% –29% –30%
1945–94
–34%

–40% –40% –40%

Figure 7.3. Differences in Labour support between working class and middle class groups
Note: The figures here show the difference between the proportions of middle and working class people
who support the Labour party as a percentage of the electorate (three period moving average). The top-left
graph compares the old middle class (OMC) to the working class (WC); the top-right graph compares the new
middle class (NMC) to the working class; the bottom graph compares the junior middle class (JMC) to the
working class.
Source: Gallup 1945–1967; British Election Studies 1964–1992; British Social Attitudes Surveys 1983–2014.

151
The New Politics of Class

the Gallup, BES, and BSA data to give a continuous time series over the
1945–2014 period.7
The average difference in Labour support between the old middle class and
the working class before 1994 is 34 per cent. There is fluctuation around this
level, but it is essentially trendless. Similarly there is trendless fluctuation from
1997 to 2014, but at a much lower level of difference: just 13 per cent. There is
an identical story for the junior middle class group: trendless fluctuation in
class voting before 1994 and then trendless fluctuation in class voting,
but at a much lower level, from 1997 to 2014. The pattern is slightly
different for the new middle class group: there seems to be some shift
towards Labour in the 1950s, although we should note the slightly different
measure of ‘professional’ workers that Gallup uses in the 1945–68 period
compared to the BES and BSA surveys that cover later time periods. Overall,
the changes in the 1990s mainly involve the middle class groups. By moving
towards Labour in the 1990s, they close the gap with the working class in
Labour support.
There are also hints from Figures 7.1 and 7.2 that the 2015 election appeared
to be another break point, at least for Labour. Table 7.1 shows people’s
reported votes by occupational class for 2015 from the BES and BSA data.8
This is the first election since the war in which the working class voted for
Labour at a lower rate than some of the middle class groups. The new middle
class and junior middle class are now more likely to vote Labour than the
working class, and the difference between the working class and the old
middle class has narrowed considerably. This is largely driven by a sharp fall
in working class support for Labour. In 2005 almost identical numbers of the
new middle class (31 per cent in the BES and 28 per cent in the BSA) and junior
middle class (23 per cent in the BES and 27 per cent in the BSA) supported
Labour as in 2015. Yet the proportion of the working class supporting Labour

Table 7.1. Impact of occupational class on vote choice in 2015

Labour Conservative Liberal Other No vote

BES data
Old middle class 19% 46% 6% 11% 19%
New middle class 29% 32% 7% 13% 19%
Junior middle class 22% 36% 3% 12% 26%
Working class 20% 20% 4% 21% 35%
BSA data
Old middle class 17% 44% 5% 11% 23%
New middle class 26% 30% 9% 14% 21%
Junior middle class 24% 30% 4% 12% 30%
Working class 25% 17% 3% 13% 42%

Note: The numbers here show the proportion of people who reported voting for each party in the 2015 general election.
Source: British Election Study 2015; British Social Attitudes Survey 2015.

152
 Class Politics Is Dead

has shrunk dramatically. In 2005 33 per cent of the working class BES
sample and 37 per cent of the working class BSA sample reported voting
Labour. In 2015 the equivalent figures in Table 7.1 are 20 per cent and
25 per cent.
In the rest of this chapter we focus on the changes before 2015 that affected
class voting most obviously and most dramatically. As Chapter 8 shows, the
changes to Labour support in 2015 were a mixture of defection to UKIP and
the SNP as well as continued increases in non-voting among the working class
that started in the late 1990s.
Drawing this all together, there is a relatively clear story to be told about
the two main parties. Class voting was very high in the immediate post-war
period. The middle classes were much less likely to vote Labour, and much
more likely to vote Conservative, than the working class. This remained
the case until the mid-1990s when there was an abrupt change which
reduced class voting to much lower, although by no means zero, levels.
Macmillan may have claimed that the ‘class war is obsolete’ after the 1959
election, but it was not until forty years later that things actually
changed with Blair’s declaration that the ‘class war is over’ to the Labour
party conference in 1999. To start with, this change was largely due to the
middle class groups becoming more likely to vote Labour, but after 2001 it
has mainly been driven by the desertion of the working class from Labour.
As Chapter 8 shows, the nail in the coffin of Labour class voting was
hammered in at the 2015 election, which, unprecedentedly, saw lower
proportions of the working class voting Labour than some of the middle
class groups.

The Liberals
Britain over the last seventy years has not been a two-party system. As well
as nationalist and minor parties, which we discuss in Chapter 8, there has
been the constant presence of a third party in various incarnations: the
Liberals from 1945 until 1981, the Alliance of the Liberals and the SDP
from 1981 until 1988, and after 1988 the Liberal Democrats. Whatever the
name, the third party has traditionally been seen as a more centrist, but still
recognizably middle class, alternative to the Conservatives. While Sand-
brook’s characterization of the Liberals in the 1950s as ‘the province of
middle class eccentrics’ (2006, p.94) is perhaps a little harsh, the image of
the Liberals as a haven for the bearded and sandal-shod middle class has
some basis in fact. Academic studies of voting behaviour have generally
found a middle class basis to the Liberal vote since the 1960s (Robertson
1984; Heath et al. 1985, 1991; Russell and Fieldhouse 2005) especially

153
The New Politics of Class

a) Gallup data

30%

20%

10%

0%
1945 1960 1975 1990 2005

b) BES data

30%

20%

10%

0%
1945 1960 1975 1990 2005

c) BSA data

30%
NMC

20% JMC
OMC
10%
WC
0%
1945 1960 1975 1990 2005

Figure 7.4. Liberal support by occupational class

Note: The figures here show the proportion of people who support the Liberal party (and later incarnations) as a
percentage of the electorate. The top graph shows vote intentions from Gallup data (three period moving average);
the middle graph shows vote choices from BES data; the bottom graph shows party support from BSA data (three
period moving average). Four occupational class groups are displayed: old middle class (OMC), new middle class
(NMC), junior middle class (JMC), and working class (WC).
Source: Gallup 1945–1968; British Election Studies 1964–2015; British Social Attitudes Surveys 1983–2015.

for ‘core’ Liberal support (Alt et al. 1977). Figure 7.4, which tracks Liberal
support over time by occupational class, supports this view. The three
middle class groups are always more likely to vote for the Liberals than the
working class. Moreover, rates of Liberal voting among the new middle class

154
 Class Politics Is Dead

group are especially high, almost twice the rate of Liberal voting among the
working class. One might argue that part of class politics in Britain is located
not just in the competition for voters between the Conservatives and
Labour, but also in the appeal of the third party.
The pattern of change in the link between class and party is rather different
for the Liberals, however. We see, if anything, growth in the differences
between the middle class groups and the working class over the 1960s and
1970s, and little evidence of the steep decline in class-based support in the
1990s that we saw for the Conservatives and Labour. The explanation for
the lack of change for the Liberals is intimately tied to the explanation
for the changes in support for the two main parties. Part of this story relates
to the reasons why occupational class matters for vote choice. In Chapter 4 we
discussed how people’s occupation, but also their education, shapes their pol-
itical views. In fact, much of the difference in Liberal support between the
new middle class and working class groups is due to educational differences
(see Franklin et al. 1992, pp.126–7, for an earlier discussion of this). Table 7.2
shows estimates from a multinomial logit regression model that allows us to
look at the separate effects of education and occupation on party support for
each decade from the 1960s onwards. Party choice is a five category variable:
Conservative, Labour, Liberal, other party, and no party. We run these
models using the BES and BSA data separately (the Gallup surveys do not
ask about educational attainment). Similar predictions from these same
models for Conservative and Labour support are in Appendix Tables A7.1
and A7.2.
As Table 7.2 shows, it is both education and occupation that account for
differences in the Liberal vote. People with higher education are about 10
percentage points more likely to support the third party than are people who
left school at fifteen or sixteen.9 This illustrates a much broader point about
the underlying processes that generate party support. Chapter 4 showed that
occupational class is a good predictor of people’s economic attitudes (whether
they want redistribution, public ownership, and the like), but a much weaker
predictor of social liberalism (whether they want the death penalty reintro-
duced and so forth), which is much more a function of education. This
directly relates to people’s party choices. As we will see shortly, the decision
to vote Conservative or Labour is largely about people’s positions on the
economic left–right dimension. The decision to vote Liberal is largely about
people’s positions on the social liberal–conservative dimension. As we have
discussed in previous chapters it is the two main party positions on left–right
economic policy that shifted in the 1990s, just as class voting shifted in the
1990s. For us, it is party change that is a key part of understanding why we see
class voting decline for the two main parties, just as the lack of party change is

155
The New Politics of Class

Table 7.2. Impact of occupational class and education on Liberal support

a) BES data

Vote Liberal 1960s 1970s 1980s 1990s 2000s 2010s

Old middle class 10% 13% 18% 15% 10% 5%

New middle class 11% 23% 27% 18% 14% 4%
Junior middle class 10% 20% 20% 14% 14% 4%
Working class 5% 15% 18% 12% 10% 3%
New middle class – working class 7% 8% 9% 6% 4% 1%
High education (degree) 19% 28% 28% 21% 25% 7%
Medium education (A Level) 13% 21% 21% 15% 16% 5%
Low education (school leaving age) 9% 20% 19% 11% 14% 3%
High education – low education 9% 8% 8% 11% 11% 4%

b) BSA data

Support Liberals 1980s 1990s 2000s 2010s

Old middle class 15% 9% 10% 3%

New middle class 23% 12% 13% 4%
Junior middle class 21% 11% 12% 4%
Working class 18% 8% 10% 3%
New middle class – working class 5% 4% 3% 2%
High education (degree) 32% 15% 22% 8%
Medium education (A Level) 23% 12% 15% 6%
Low education (no qualifications) 19% 10% 11% 3%
High education – low education 14% 5% 11% 5%

Note: The numbers here are predicted probabilities from multinomial logit regression models using pooled data by
decade that predict vote choice for the BES data and party support for the BSA data using occupational class, education,
and year. The predicted occupational class probabilities are for someone in the medium education category and the
predicted education probabilities are for someone in the junior middle class category. Year is set as close as possible to
the middle of each decade (and the same year for both datasets).
Source: British Election Studies 1964–2015; British Social Attitudes Surveys 1986–2015.

key to understanding why class voting, or more accurately education voting,

for the Liberals remains the same.

Explaining Change in Policy, and Therefore Class, Voting

In order to explain the changes that we saw in the 1990s we need to connect
people’s political views with their party choices. There are two important tests.
The first is that when we account for people’s policy views and their changing
impact on party choice, we should eliminate much of the change in class
voting over the 1990s when the parties radically altered their policy. The
second is that people’s policy stances, especially on economic issues, should
have a much weaker effect on their party choices when the parties became
more similar in terms of policy in the 1990s.

156
 Class Politics Is Dead

Figure 7.5 tests the first of these. Here we aim to assess whether holding
constant economic ideology, but allowing its effect to vary over time weak-
ens the relationship between class and vote and reduces any change in that
relationship. Since class differences in ideology are relatively stable, if we
account for ideology and the declining motivation for choosing parties due
to ideology, we should eliminate changes in class voting. We focus here on
the critical period of change: 1987–2010. The two left-hand graphs show
Labour support using both the BES and BSA data between the mid 1980s and
the 2010 election. These simply replicate what we saw in Figure 7.1 at the
beginning of the chapter, and the steep decrease in class voting is very
obvious for both data sets. The two graphs on the right-hand side are
from regression models which hold constant people’s views on the eco-
nomic left–right scale that we used in Chapter 4.10 The lines at the bottom
of the figures show the differences between the middle classes and the
working class, and illustrate quite clearly that much of the rapid change
in class voting over the period that we see in the left-hand graphs is not
present in the right-hand graphs. It is not only that most of the differences
between classes in vote choice are due to differing values but also that the
weaker effect of these values over time has decreased class voting. As we
have found previously (Evans and Tilley 2012b), the changes in class voting
are not eliminated, but they are substantially reduced. This is remarkable
given that there are no variables other than occupational class and left–
right values in the model.
The second test is to look at how well people’s policy stances predict their
vote choice. For our argument to hold, these views on left–right economic
policy should have become less important as the parties become more similar
in the 1990s. Using the same regression models as in Figure 7.5, Figure 7.6
shows the proportion of left-wing and right-wing people who we predict
support Labour over time.11 Both the BSA and BES data tell the same story.
In the 1980s and early 1990s, while very few people on the right voted Labour,
a large majority of people on the left crossed the box next to the Labour
candidate. In 1987 the difference in Labour support between people on the
left and right was enormous. This gap begins to decline from the mid-1990s
and almost disappears during the 2000s. In 1992 the difference between
people on the left and right was nearly 40 percentage points in the BES data;
by 2001 this difference was only 3 percentage points. It is not surprising that
class voting has declined, because the underlying basis of much class voting,
economic policy voting, was in freefall over the 1990s, and by the 2000s had
virtually disappeared.
These figures allow us to put together the different patterns of change from
earlier chapters. Chapter 4 showed that people in different occupational
classes had different positions on redistribution and other economic policy.

157
The New Politics of Class

BES data (1987–2010):

a) Raw numbers b) Controlling for left–right values
WC
40% 40%

JMC
NMC
20% 20%
OMC

0% 0%
1985 1995 2005 1985 1995 2005

–20% Differences between the –20%

the middle classes and
the working class
–40% –40%

BSA data (1986–2010):

a) Raw numbers b) Controlling for left–right values

60% WC 60%

NMC
40% 40%

JMC
20% 20%
OMC

0% 0%
1985 1995 2005 1985 1995 2005

–20% –20%
Differences between the
middle classes and
–40% the working class –40%

Figure 7.5. Labour support by occupational class controlling for left–right values
Note: The figures here show predicted probabilities of Labour support from multinomial logit regression models
that predict vote choice (BES) and party support (BSA, two period moving average) as a percentage of the electorate
for each year separately. The left-hand graphs include occupational class as an independent variable; the right-
hand graphs include a measure of economic left-right values as well as occupational class. Four occupational class
groups are displayed: old middle class (OMC), new middle class (NMC), junior middle class (JMC), and the working
class (WC).
Source: British Election Studies 1987–2010; British Social Attitudes Surveys 1986–2010.

These differences between classes are fairly constant over time, which is
perhaps not surprising given that the objective differences between occupa-
tional groups discussed in Chapter 2 are also fairly constant. Chapter 6
showed a rather different pattern when it came to the political parties.

158
 Class Politics Is Dead

a) BES data 1987–2010 b) BSA data 1986–2010

75% 75%
Left-wing Left-wing

50% 50%

25% 25%
Right-wing Right-wing

0% 0%
1985 1995 2005 1985 1995 2005

–25% –25%
Difference
Difference

–50% –50%

Figure 7.6. Labour support by economic left–right position

Note: The figures here show predicted probabilities of Labour support from multinomial logit regression models
that predict support as a percentage of the electorate for each year separately. The dependent variable in the left-
hand graph is vote choice. The dependent variable in the right-hand graph is party support (two period moving
average). The models include occupational class and a measure of economic left–right values. Left-wing people are
assumed to score one standard deviation below the mean score for the year and right-wing people one standard
deviation above the mean score for the year.
Source: British Election Studies 1964–2010; British Social Attitudes Surveys 1986–2010.

Here there was change, especially in the 1990s when the two main parties
converged in terms of policy. People’s policy views have not changed, but the
policy choices that parties offer have. This has meant that ideology, and the
root of ideology—class, has become a weaker determinant of people’s party
choice. After all, if the parties all offer the same economic policies, why would
someone’s views on economic policy affect their vote?
All this makes the evolution of the Liberal vote an interesting contrast.
Higher education, and to a lesser extent having a new middle class job, are
relatively good predictors of a Liberal vote, as we saw earlier. Moreover, they
are a fairly consistent predictor over time. There has been none of the rapid
change in class voting that we see for the two main parties in the 1990s. There
are two reasons for this. First, divisions among voters by social liberalism have
remained relatively constant over time, but second, the Liberals have almost
always offered a more socially liberal platform of policies than the two main
parties. The expert survey data described in the previous chapter has measures
of social liberalism that go back to 1999 and these show that the Liberals had

159
The New Politics of Class

a) BES data 1987–2010 b) BSA data 1986–2010

40% 40%

Socially
liberal Socially
liberal
20% 20%

Socially
conservative Socially
conservative
0% 0%
1985 1995 2005 1985 1995 2005

Difference Difference

–20% –20%

Figure 7.7. Liberal support by social liberal–conservative position

Note: The figures here show predicted probabilities of Liberal support from multinomial logit regression models
that predict support as a percentage of the electorate for each year separately. The dependent variable in the left-
hand graph is vote choice. The dependent variable in the right-hand graph is party support (two period moving
average). The models include a measure of economic left–right values and a measure of social conservative–liberal
values. Socially conservative people are assumed to score one standard deviation below the mean score for the year
and socially liberal people one standard deviation above the mean score for the year. Left–right values are held at
their mean for each year.
Source: British Election Studies 1964–2010; British Social Attitudes Surveys 1986–2010.

more consistently socially liberal policies than both the Conservatives and
Labour from 1999 to 2014.12 In 2006 the Conservatives score 6 (on a 0–10
scale, where 10 is most socially conservative), Labour 4.7, and the Liberals
just over 2.5.
The second of these factors means that we should expect that social liberal-
ism remains a fairly consistent predictor over time of a Liberal vote. Figure 7.7
shows that this is the case. Since the Liberal party is always more socially
liberal, more socially liberal voters are always more likely to support it.13
Indeed, if anything, social liberalism appears to have become a little more
important in explaining people’s votes for the Liberals.

Beyond Policy Voting

People’s ideology affects to their vote. Left-wing people are more likely than
right-wing people to prefer Labour, just as socially liberal people are more

160
 Class Politics Is Dead

likely than socially conservative people to prefer the Liberals. But this rela-
tionship is altered when the parties’ policy offerings alter. When parties
offer similar policies, then fewer people use their own policy preferences to
choose between the parties. This process can explain much of the steep
decline in class voting that we saw in the 1990s. It cannot explain it entirely,
however. Figure 7.5 certainly showed that holding constant people’s left-
right values reduced the change in class voting, but it did not eliminate
that change.
Why is this? Part of the reason relates back to the discussion in Chapter 6
about the parties themselves. While people choose parties on the grounds of
policy, they also make choices because they feel that one party represents
them, or their group, better than another. This is related to the policies that a
party offers, but is not the same. Two parties might have very similar policies,
but if one offers rhetoric that emphasizes that it is for the working class, has
candidates from recognizably working class backgrounds, and both of those
factors are emphasized by newspapers that everyone reads, then it seems likely
that there will still be class voting. This hypothetical example is, of course,
rather similar to the situation that we describe in Britain in the late 1960s. The
parties briefly offered quite similar policies, but they were still clearly distinct-
ive in terms of their appeal and this was strongly reflected by a mass partisan
press stratified by class. A final part of our story then is how people’s views of
the parties as class parties have changed and ultimately how this has depressed
class voting.
How can we test this part of our explanation? The key issue here is percep-
tions of the parties. Do people think that particular parties defend particular
classes? Fortunately, and as discussed in the previous chapter, there is a
question in the BES surveys since 1987 that asks people exactly this. For the
two main parties survey respondents are asked first how well each party ‘looks
after working class people’ and then how well each party ‘looks after middle
class people’ on a four point scale. We focus here on perceptions of Labour
since that is the party which has changed the most and use the same two
questions discussed in Chapter 6 that measure the class basis of party image.
Again, positive scores mean that people think Labour looks after the working
class better than the middle class; negative scores mean that people think
Labour looks after the middle class better than the working class. If Labour was
seen as a purely working class party it would score 2; if a purely middle class
party it would score 2. As shown in Chapter 6, Labour’s image has changed
dramatically over time. In 1987 the mean score was 0.74, about a standard
deviation above zero. Most people in 1987 saw Labour as a party for the
working class. In 2005 the mean score was 0.09 and in 2010 it was 0.17.
By the 2000s people were almost as likely to see Labour as a middle class
party as they were to see it as a working class party.

161
The New Politics of Class

a) Controlling for left–right values b) + Controlling for perceptions of

Labour as a working class party

WC
40% 40%

JMC NMC
20% 20%

OMC

0% 0%
1985 1995 2005 1985 1995 2005

–20% –20%
Differences between the
the middle classes and
the working class
–40% –40%

Figure 7.8. Labour support by occupational class controlling for left–right values and
class perceptions of Labour
Note: The figures here show predicted probabilities of Labour support from multinomial logit regression models
that predict support as a percentage of the electorate for each year separately. The lines below zero show the
differences between the working class and the three middle class groups. The left-hand graph includes occupa-
tional class and economic left–right values as independent variables; the right-hand graph also includes a
measure of whether people perceive Labour to be a more working class or middle class party. Four occupational
class groups are displayed: old middle class (OMC), new middle class (NMC), junior middle class (JMC), and the
working class (WC). Left–right values are held at their mean for each year and perceptions of Labour are set for
every year at zero (i.e. as though people thought that Labour was equally good for the middle class as the
working class).
Source: British Election Studies 1987–2010.

These changing perceptions affect how people vote. If we add these

perceptions into a model of vote choice then we can see how perceptions
of the parties as class parties are important. Figure 7.8 has two graphs. The
left-hand one is simply a repeat of the graph in Figure 7.5 showing the
occupational class differences in Labour vote choice, but holding constant
people’s left–right ideology. The right-hand graph shows support for Labour
by occupational class from a model that holds not just left–right ideology
constant but also our measure of people’s perceptions of Labour as a work-
ing or middle class party.14 We show our prediction of class voting for
Labour if everyone had scored zero on our class-party perceptions scale in
every year. That is, we are asking the counterfactual question: if Labour had
always been viewed as an equally middle class and working class party,
what levels of class voting would there have been in the 1980s and
1990s? The answer to that question is that levels of class voting would

162
 Class Politics Is Dead

have been dramatically lower. In fact, we would have seen almost no

change in class voting for Labour over the period. What is most striking
about Figure 7.8 is that what remains of class dealignment over the period
up to 2010 is essentially eliminated. The lines are now parallel for all
occupational classes. The new middle class is almost identically likely to
support Labour as the working class at all times when we simply account for
left-right ideology and class perceptions of the parties. The old middle class
is still slightly less likely to vote Labour than the working class, but crucially
this difference is invariant over time. This is not because the relationship
between perceptions of Labour as a class party and vote choice have
changed (these are remarkably constant: people in working class occupa-
tions that see Labour as a working class party are always more likely to
support Labour); rather it is because far fewer people see the parties as
‘looking after’ one class or the other.
Therefore, there are two processes here. First, parties adopted more similar
policies, broadly speaking on the right, and that meant that people’s
own policy differences became less important in shaping their vote choices.
Second, however, as parties became more similar, not just in terms of policy,
but also personnel, rhetoric, and media coverage, they became less identified
with particular classes. As fewer people recognized the parties as class parties,
fewer people voted on that basis. Ultimately, this means that the decline in
class voting in the 1990s in Britain was caused by parties. Parties became more
similar, and therefore both ideology and the direct class appeal of parties
mattered less to voters. This was likely exacerbated by the decline of mass
class-based partisan newspapers. If we account for changes in party policy and
class perceptions of the parties, we account for the class dealignment that
we saw between 1992 and 2005.
What happened in 2015? After all, perceptions of Labour as a middle or
working class party barely change between 2005 and 2015. We will deal with
this in more detail in Chapter 8, but in essence new parties become viable
options. UKIP in England and Wales and the SNP in Scotland attracted sup-
port from the working class. All the change before 2010 was about change to
the main parties; all the change after 2010 is about new parties that offer
policies that appeal differently to different groups.

Conclusions

Class voting has declined in Britain. If you knew someone’s occupation, and
to a lesser extent their education, in the 1950s, you had a very good chance of
guessing how they would vote. Today those clues would be a lot less helpful.
On that we can agree with almost all observers of British politics. Just as was

163
The New Politics of Class

claimed after the 2001 election, the ‘relationship between class and vote has
weakened appreciably over time’ (Clarke et al. 2004, p.33). Where we disagree
is in the pattern of change, when that weakening happened and the reasons
behind it. Using the best possible data over the longest period of time we have
shown that levels of class voting were largely static for half a century from the
1940s until the mid-1990s. There was then a dramatic weakening of the
relationship between class and party over a short space of time. How does
this fit with our findings from the earlier chapters on parties, politicians, and
the media environment, and how well does it fit with the top down model
that we advocated at the start of this chapter?
In the narrowest sense it does not fit. Parties had relatively similar policies in
the 1960s when class voting was high and also had relatively similar policies
in the 2000s and 2010s when class voting was low. This neglects three very
important points. The first is that the narrowing of policy differences in the
1960s was more sudden and more temporary than is often thought. While the
post-war consensus is often talked about, the actual economic policy differ-
ences between parties as measured by the manifesto data in the last chapter
were relatively large for most of the fifty years after the war apart from the late
1960s. Policy differences between the two main parties immediately after the
war were sizable, and the brief convergence at the 1966 and 1970 elections was
reversed by 1974. Yet economic policy differences between the parties have
been consistently small since 1997.
The second point is about perceptions relative to realities. The two main
parties may have adopted relatively similar policies in the 1960s, albeit
briefly, but that does not mean that voters’ views of the parties as better or
worse for particular classes changed. Perceptions of parties are not just about
policy but about the people, the rhetoric, and the mediation of both via the
press. None of these three factors changed until the 1990s. Parties can
change policy, but if that does not affect people’s perceptions of the parties,
both in policy and class image terms, then it will not lead to changes in
class voting.15
The final point is about the persistence of image. David Weakliem (2001)
argues that people’s views of the two US political parties as defenders of
different occupational groups had barely shifted from the 1947 to 1990. This
might seem surprising given the policy shifts of the Democrats, in particular,
over that time period. He argues that, invoking Hout et al.’s (1993) idea of class
as a psychological heuristic, ‘people expect to see a party of the rich and party
of the poor; even if parties try to go beyond the traditional categories of Left
and Right, people will continue to put them in these roles’ (Weakliem 2001,
p.219). The point here is that these views of parties are stubbornly held and it
takes a shock to change them. In Britain in the 1990s there was such a shock.
Labour radically changed its nature in a short space of time and crucially made

164
 Class Politics Is Dead

this very obvious to the electorate. The information that people received
about politics changed both as a result of, and coincidentally to, this. The
press, which was previously structured around class divisions, stopped talking
about class and people also stopped reading the press.
Radical change all at once, in every form, changed people’s views of the
parties. It became clear that both were offering similar policies and Labour
was a party for the middle class as much as it was a party for the working
class. These changes in perceptions help explain the findings at the end of
this chapter for the 1987–2010 period. Class voting has disappeared
because economic left–right policy voting for the two main parties declined
when the parties had recognizably similar policies, and because group-based
voting declined when the parties no longer clearly represented classes. The ‘top
down’ view of party shifts causing voters to change is correct, but with the
proviso that those party shifts need to be very obvious to voters. While Labour
and the Conservatives offered rather similar policies at the 1966 election, this
convergence was short-lived and not accompanied by any other change in
party image. It was thus only in the 1990s, the era of party policy and image
change, that the hold that class voting had on the British electorate was finally
overthrown.
The story does not end there, however. If one consequence of these changes
is that working class voters deserted Labour, where have they gone? In
Chapter 8 we show that the political changes of the 1990s and 2000s have
led to many working class voters exiting the system and simply not voting. We
also show that new viable and visible parties, the SNP and UKIP, have led to
Labour losing even more of the working class vote at the 2015 election.

Notes

1. When Greene asked what he would prefer on television at that time, Wilson
allegedly replied ‘I suggest that you put on Oedipus Rex’ (Cockerell 1988, p.107).
All the quotations in the text come from interviews with Wilson and Greene that
were made available by the BBC in 2015: http://www.bbc.co.uk/historyofthebbc/
elections/invention-3.
2. For example, while Weakliem and Heath (1999) convincingly show that class
voting at the 1992 election is much the same as it was in 1935, Heath’s book
after the 1997 election admits that class voting had finally declined in Britain
(Heath et al. 2001).
3. Some argue that declines in class voting have happened in other developed coun-
tries (Dalton 2008; Clark 2001b; Nieuwbeerta 1996; Pakulski and Waters 1996a,
1996b; Franklin 1992), but others claim that there is no universal decline (Elff
2007, 2009) or even that in some cases, like the US, there have been increases in
class voting over the last twenty years (Bartels 2008).

165
The New Politics of Class

4. Inevitably there are minor differences. Foremen and supervisors cannot be separ-
ated from manual workers, and personal service workers cannot easily be separated
from manual or non-manual workers. Housewives (in some cases even if in part-
time work) are also not generally classified by husband’s profession, so form a
group which is simply excluded from our analysis. The latter problem means that
we are likely to overestimate any class voting in the early period given that people’s
own occupations are generally stronger indicators of vote intention than their
spouse’s occupation.
5. People who said ‘don’t know’ were asked their party ‘inclination’ from 1963 to
1968 and we have treated these answers as vote intentions.
6. The one exception to this is the move of the new middle class away from the
Conservatives, and away from the old middle class, after the 1950s, and even this
may well be an artefact of the slightly different Gallup measure of professional
workers compared to the BES and BSA.
7. We use the Gallup data from 1945 to 1968, the BES data from 1964 to 1992, and the
BSA data from 1983 to 2015. This allows for some overlap between the different
surveys while maintaining consistency of measurement. In years with data from
more than one source, we take the mean of the two data points.
8. This is not the broader measure of party choice for the BSA data, but a separate
question asking people how they actually voted in 2015.
9. The impact of education on Labour and Conservative voting is more muted, and it
has little effect on the occupational class differences (see Tables A7.1 and A7.2 in
the Appendix to this chapter). For example, those with school leaving age educa-
tion or higher education support Labour at relatively similar rates, and those with a
medium level of education are about 5 per cent less likely to support Labour than
those with a degree. The effect of holding constant education does not substan-
tially change the magnitude of class differences in Labour or Conservative support,
or the pattern of change. For example, holding constant education, there is around
a 20 per cent gap between the new middle class and working class in Labour
support in the 1980s, which reduces to around 5 per cent in the 2000s.
10. We use multinomial logit regression models which predict party choice, with
survey year, class, left–right ideology and interactions between survey year and
class and survey year and ideology. Party choice is a five category variable: Conser-
vative, Labour, Liberal, other party, no party. For the BSA data the left–right scale is
identical to that described in Chapter 4. To maintain consistency across the whole
time period, the BES left–right scale is slightly different and uses three questions
about whether there is ‘one law for the rich and one law for the poor’, whether
‘ordinary people get a fair share of the wealth’, and whether people think that ‘there
is no need for strong trade unions to protect employees’ working conditions and
wages’. As could be seen in Chapter 4, the average score on these kinds of scales
remains similar over time, but we hold the values constant at each year’s mean value.
11. Left-wing people are taken to have a score one standard deviation below the mean
value for that year, and right-wing people are taken to have a score one standard
deviation above the mean value for that year. The other variable in these models is
class, which is held constant at the working class category.

166
 Class Politics Is Dead

12. This is the ‘Gal-Tan’ measure (see Bakker et al. 2015). Experts are asked where they
would place the parties given that: ‘ “Libertarian” or “post-materialist” parties favor
expanded personal freedoms, for example, access to abortion, active euthanasia,
same-sex marriage, or greater democratic participation. “Traditional” or “authori-
tarian” parties often reject these ideas; they value order, tradition, and stability,
and believe that the government should be a firm moral authority on social and
cultural issues.’
13. These figures come from multinomial logistic regression models that predict party
choice using left–right values, social conservative–liberal values, and survey year.
The first two variables are all interacted with survey year in the models and are
hence allowed to vary in impact from year to year. We use a slightly different
version of the liberal–conservative scale described in Chapter 4. It uses only four of
the five questions as the item about the morality of homosexuality is not present in
some of the surveys. Socially liberal people are taken to be one standard deviation
above the mean value for that year, and socially conservative people are taken to
have a score one standard deviation below the mean value for that year. Left–right
values are held constant at their mean value for that particular year.
14. This model takes the same form as previously, except with the addition of the scale
score of which class the Labour party is ‘for’. We interact the scale score with
occupational class, since middle class people will prefer a party that represents
middle class people and working class people will prefer a party that represents
working class people. Both the direct and moderated effects of the scale are allowed
to vary by year.
15. It is also possible that voters after the 1970s were more responsive to what parties
said and did. Some argue that party choice became more instrumental and less an
expression of partisan loyalty (Franklin 1984, 1985; Rose and McAllister 1986). If
this did happen, and class-based responsiveness to parties’ ideological signals was
stronger by the 1980s, then this would further exaggerate the impact of the key
changes made by Labour.

167
The New Politics of Class

Appendix to Chapter 7

Table A7.1. Impact of occupational class and education on Labour support

a) BES data

Vote Labour 1960s 1970s 1980s 1990s 2000s 2010s

Old middle class 12% 11% 9% 22% 26% 18%

New middle class 22% 19% 15% 32% 35% 23%
Junior middle class 25% 22% 18% 31% 29% 22%
Working class 51% 39% 36% 46% 41% 25%
High education (degree) 16% 29% 22% 37% 22% 29%
Medium education (A Level) 19% 17% 17% 31% 24% 20%
Low education (school leaving age) 37% 33% 24% 40% 25% 20%

b) BSA data

Support Labour 1980s 1990s 2000s 2010s

Old middle class 11% 27% 30% 20%

New middle class 17% 36% 35% 25%
Junior middle class 18% 34% 33% 25%
Working class 36% 50% 43% 32%
High education (degree) 26% 43% 39% 32%
Medium education (A Level) 17% 34% 35% 27%
Low education (no qualifications) 26% 41% 40% 31%

Note: The numbers here are predicted probabilities from multinomial logit regression models using pooled data by
decade that predict vote choice for the BES data and party support for the BSA data using occupational class, education,
and year. The predicted occupational class probabilities are for someone in the medium education category and the
predicted education probabilities are for someone in the junior middle class category. Year is set as close as possible to
the middle of each decade (and the same year for both datasets).
Source: British Election Studies 1964–2015; British Social Attitudes Surveys 1986–2015.

Table A7.2. Impact of occupational class and education on Conservative support

a) BES data

Vote Conservative 1960s 1970s 1980s 1990s 2000s 2010s

Old middle class 64% 67% 63% 44% 47% 49%

New middle class 53% 48% 47% 30% 33% 40%
Junior middle class 50% 46% 50% 31% 33% 35%
Working class 27% 29% 31% 17% 20% 21%
High education (degree) 39% 30% 36% 19% 21% 27%
Medium education (A Level) 52% 48% 46% 28% 23% 33%
Low education (school leaving age) 36% 35% 42% 22% 20% 31%

168
 Class Politics Is Dead

b) BSA data

Support Conservatives 1980s 1990s 2000s 2010s

Old middle class 66% 46% 43% 51%

New middle class 50% 32% 31% 39%
Junior middle class 49% 33% 29% 36%
Working class 30% 17% 17% 21%
High education (degree) 30% 24% 20% 31%
Medium education (A Level) 51% 36% 27% 35%
Low education (no qualifications) 41% 31% 27% 35%

Note: The numbers here are predicted probabilities from multinomial logit regression models using pooled data by
decade that predict vote choice for the BES data and party support for the BSA data using occupational class, education,
and year. The predicted occupational class probabilities are for someone in the medium education category and the
predicted education probabilities are for someone in the junior middle class category. Year is set as close as possible to
the middle of each decade (and the same year for both datasets).
Source: British Election Studies 1964–2015; British Social Attitudes Surveys 1986–2015.

169
8

Long Live Class Politics

If people do not like the party that they used to vote for, what can they do?
There are really three choices: vote for the opposing party, vote for a new
party, or stop voting.1 This chapter sets out how the decline of class voting
for the two major parties, has led to a revival of class voting in terms of new
parties, but most importantly in terms of abstention. As Labour, and to some
extent the Conservatives, have left people with working class jobs and low
levels of education behind, those same people have become disengaged from
mainstream politics. We show here that the previously small gap between the
classes in terms of turnout sharply increased during the 2000s, resulting in a
voting population that has been systematically shorn of the working class.
Class voting may have disappeared, but class non-voting is stronger than ever.
This is a direct result of disillusionment with political parties that fail to
represent people’s views. The 2015 election also heralded a set of new party
choices that have increased class voting. In Scotland the Scottish National
Party (SNP) is now the most working class party after taking great swathes of
former Labour voters in 2015. In England and Wales, the United Kingdom
Independence Party (UKIP) is the new home of the working class, with
voters deserting the mainstream to join the emergent radical right party.
This chapter thus shows how both non-voting and new parties have signalled
a resurgence of the importance of class in politics.

To Vote Or Not To Vote?

Rates of turnout in Britain until 1997 were relatively static. Some elections saw
small increases, and some saw small decreases, but around 75 per cent of the
electorate regularly turned out to vote. The 78 per cent turnout rate in 1992
was almost the same as the 79 per cent in 1959. Even 1997, which saw the
lowest turnout since the war at 71 per cent, witnessed only slightly fewer
voters than 1970 (with 72 per cent turnout). This changed quite dramatically
 Long Live Class Politics

100%
Reported
turnout (BES)

80%

Official
60%

40% Validated
turnout (BES)

20%

0%
1960 1975 1990 2005

Figure 8.1. Turnout over time

Note: The figure here shows official turnout figures, reported turnout from the BES, and validated turnout from
the BES.
Source: Electoral Commission; British Election Studies 1964–2015.

after 1997. The nadir was the 2001 general election when voter participation
fell to 59 per cent. While turnout has crept back up since then, rates of voting
are still substantially below where they stood in 1992. Figure 8.1 shows these
changes (from 1964 to 2015) and also illustrates an important issue with
measuring turnout. The dotted line shows official turnout rates as recorded
by the Electoral Commission (this is the percentage of people who cast a ballot
as a proportion of all registered voters). The dashed line is the reported rate of
turnout in the British Election Surveys since 1964. This is systematically
higher because a) the BES sample tends to favour the inclusion of voters
rather than non-voters and b) people tend to exaggerate their involvement
in politics. The thick line in between accounts for the second of these, as it
shows validated turnout rates from the BES. These figures come from the BES
as well, but show the actual turnout rates of people who were surveyed using
checks of the official electoral record.2 These validated turnout figures are only
available from 1987 onwards, but are the data that we will use mainly in this
chapter, not least because there is so little change in turnout between the
1960s and the 1980s.
Regardless of the exact measure, Figure 8.1 clearly depicts the sharp decline
in turnout since 1992. The reasons for this dramatic change are often argued
to be similar to the factors that underpinned the changes in class voting that
we discussed in Chapter 7. Heath and Taylor (1999) focus on the lack of
ideological distance between the main parties, and Heath (2007) argues that
public perceptions of the main parties as indistinguishable were the main
driver of lower turnout. The particular circumstances of the 2001 election

171
The New Politics of Class

are also potentially important since this was an election which was widely
seen as a foregone conclusion and sparked little interest among the public
(Johnston and Pattie 2003). Implicitly these kinds of explanations call upon
ideas of the ‘rational voter’: someone who pays attention to their chances of
being the pivotal voter who decides an election and weighs the expected
benefits from one party or another winning against the costs of actually voting
(Downs 1957; Aldrich 1993). This is one way of thinking about the decision to
vote or not. We will largely duck the long-running debate about whether any
voters really think like this, because it is not crucial for us to explain why fewer
people in general vote today than did in 1992.3 What we are interested in is
whether certain types of people are less likely to vote today than they were
twenty years ago. In fact, there are two other important explanations of the
decision to vote that we should consider insofar as they relate to different
types of people voting. One involves mobilization and psychological engage-
ment: do people care about politics and care about the parties? The other
concerns resources.
Mobilization can take many forms, but in essence the argument is that
‘people participate in electoral politics because someone encourages or inspires
them to take part’ (Rosenstone and Hansen 1993, p.161). Related to this is the
expressive power of voting: that turning out to vote is about expressing one’s
belonging to a particular partisan or social group (Brennan and Hamlin 1998;
Hamlin and Jennings 2011). These explanations are rooted in ideas about how
well parties relate to specific groups of voters and represent their interests. Of
course, this all intersects with the process of parties becoming indistinguishable
from one another. If Labour no longer looks like a working class party then it is
less able to ‘inspire’ working class voters. As those same working class voters
leave Labour, it makes sense that they will become non-voters if there are no
more ‘inspiring’ options available. This model suggests that rising levels of class-
based abstention may be due to the changing parties.
Yet some might argue we would expect lower turnout from certain
social classes anyway. The resources model of participation holds that socio‑
economic resources have substantial effects on people’s likelihood to vote.
Implicitly, this model suggests that the costs of voting, and maybe also the
benefits, are different for different types of people. In the US, it has long been
recognized that class, income, and education influence participation (Verba
and Nie 1972; Verba et al. 1978; Leighley and Nagler 1992a, 1992b; Verba et al.
1995; see Leighley and Nagler 2014 for an excellent overview). The argument
runs that people in higher social classes, with more education and income, are
more likely to vote because it is less costly for them to vote (especially in terms
of acquiring political information). This is an explanation that lends itself to
more static and constant class differences in non-voting, as in the US case
(Leighley and Nagler 1992a, 2014).

172
 Long Live Class Politics

a) Social class b) Education

Low
40% WC 40%

Medium
JMC
NMC
20% OMC 20%

High

0% 0%
1985 1995 2005 2015 1985 1995 2005 2015

Differences between Differences between

the working class and the low educated
the middle classes
–20% –20% and the rest

Figure 8.2. Non-voting rates by occupational class and education

Note: The figures here show predicted probabilities of non-voting from logit regression models that predict
validated non-voting for each year separately. The left-hand graph disaggregates the results by social class; the
right-hand graph disaggregates the results by education. As well as occupational class and education, these models
include controls for trade union membership, gender, age, region, religion, and race. Four occupational class
groups are displayed: old middle class (OMC), new middle class (NMC), junior middle class (JMC), and working
class (WC). Three educational groups are displayed: people with degree-level education (high), people with A Level
equivalent education (medium), and people who left school at the minimum school leaving age for their cohort
(low). The predicted probabilities are for a white man with no religion in his forties, who has middling educational
attainment/a junior middle class job, who lives in the south east of England, and is not a trade union member.
Source: British Election Studies 1987–2015.

Is it mobilization or resources that matters? In Britain, like the US, class has
sometimes been shown to be associated with rates of turnout (Swaddle and
Heath 1989; Pattie and Johnston 2001; Heath and Taylor 1999; Whiteley et al.
2001; Parry et al. 1992; Heath 2016). However, the differences in Britain have
been generally rather small and before the 2000s often zero (Crewe et al. 1977a;
Heath et al. 1991; Denver 1995; Pattie and Johnston 1998). Figure 8.2 shows the
differences by occupational class and education in rates of non-voting. High
numbers thus indicate that people are less likely to vote. As in Chapter 7, the
lines below the x-axis are the differences between the three middle class groups
and the working class, and the differences between those with a high or medium
education compared to those with a low level of education. The numbers here
are derived from regression models using the BES data with class, education, age,
sex, region, religion, race, and trade union membership as independent variables
predicting validated voter turnout.4 As we look at validated votes only, we are
limited to the 1987–2015 period. Nonetheless, the pattern is quite clear. From
1987 to 1992 there are only small differences between class and educational

173
The New Politics of Class

50% Proletariat

30%

Intelligentsia
10%

1985 1995 2005 2015

–10%
Difference

–30%

Figure 8.3. Non-voting rates by combined social group

Note: The figure here shows predicted probabilities of non-voting from logit regression models that predict non-
voting for each year separately. As well as occupational class and education, these models include controls for trade
union membership, gender, age, region, religion, and race. The predicted probabilities are for a white man with no
religion in his forties, who lives in the south east of England and is not a trade union member. The two groups are
low education and a working class occupation (proletariat) and high education and a new middle class occupation
(intelligentsia).
Source: British Election Studies 1987–2015.

groups in rates of non-voting. The maximum difference between occupational

or educational groups in these two elections is barely five percentage points. This
reflects a long-standing pattern that is found in the 1960s and 1970s as well.
Using non-validated measures of voting, the differences between the three
middle class occupational groups and the working class in 1964 are between 2
and 6 per cent; in October 1974 they are between 1 and 5 per cent.
There are differences between the classes before the 1990s, but there is little
real support for the resource model of turnout, because these differences are so
small. More importantly, while resources have not changed, non-voting rates
between the classes have dramatically altered. By the 1997 election we see the
start of bigger gaps between the classes and by 2010 these have become large
divides. To return to our class archetypes from Chapters 4 and 5, we predict that
in 2015 someone in the bourgeoisie (old middle class job and medium educa-
tion) has a 16 per cent chance of not voting and someone in the intelligentsia
(new middle class job and high education) has a 13 per cent chance of not
voting. Over 85 per cent of the middle class archetypes vote. Compare that to
the proletariat (people with a working class job and low education), of whom we
predict 52 per cent do not vote. Figure 8.3 displays the changes in non-voting
over time for the intelligentsia and proletariat. From a position in the 1980s
of almost equality between groups in turnout rates, we now see large

174
 Long Live Class Politics

Table 8.1. How 1997 Labour voters voted in 2001 by occupational class

Labour Conservative or Liberal Didn’t vote

Old middle class 70% 15% 13%

New middle class 70% 16% 10%
Junior middle class 66% 14% 15%
Working class 67% 8% 23%

Note: This table shows the percentages of people who voted Labour in 1997 and how those same people voted in 2001.
The total unweighted number of 1997 Labour voters in the sample is 947. Not shown are the percentage of people who
voted for minor parties (3 per cent of the sample) or people in the personal service, own-account, foreman occupational
class categories, or could not be coded (183 respondents).
Source: British Election Panel Study 1997–2001.

inequalities. Heath shows a similar pattern up to 2010, and argues that by

2010 ‘class is more important as a participatory cleavage than it is as an
electoral cleavage’ (Heath 2016, p.9). This change occurred precisely when
parties and the media environment radically changed their nature and there-
fore suggests that it is not resources that matter, but party appeals.5 As the set
of parties available becomes less appealing to certain groups, those groups vote
with their feet and stop participating. If it is not possible for me to express my
identity or preferences because no parties represent that identity or prefer-
ence, then why would I vote?
Using panel surveys which interview the same people repeatedly, we can
show exactly this process at work in the 2000s. We take people who voted
Labour in 1997 and then look at how those same people voted in 2001.
Table 8.1 shows the proportion of 1997 Labour voters, by their occupational
social class in 1997, who a) continue to support Labour in 2001, b) switch to
another mainstream party, and c) stop voting. As the table shows, the propor-
tion of voters who Labour retains in the different occupational class groups is
fairly similar (around two thirds). What is quite different is the proportion of
people who become non-voters over those four years. Only 10 per cent of the
new middle class 1997 Labour voters do not vote in 2001. For the equivalent
working class group, the figure is 23 per cent. Middle class voters may have
tired of Labour between 1997 and 2001 at the same rate as the working class,
but they had other party options that were appealing. Fifteen per cent of the
three middle class groups switched to the Conservatives or Liberals. Only
8 per cent of the working class group did this.6
We also have panel data for the 2005–2010 period. Unlike the 1997–2001
election cycle which saw a large decrease in turnout, 2005–2010 saw a (small)
increase in turnout. Nonetheless, the same pattern of the working classes
deserting Labour to become non-voters can be seen. Table 8.2 shows that
nearly 10 per cent of people in working class jobs7 who voted Labour in

175
The New Politics of Class

Table 8.2. How 2005 Labour voters voted in 2010 by occupational class

Labour Conservative or Liberal Didn’t vote

Old middle class 60% 30% 3%

New middle class 64% 34% 1%
Junior middle class 58% 27% 5%
Working class 62% 22% 9%

Note: This table shows the percentages of people who voted Labour in 2005 and how those same people voted in 2010.
The total unweighted number of 2005 Labour voters in the sample is 666. Not shown are the percentage of people who
voted for minor parties (6 per cent of the sample) or people who could not be placed in the four occupational class
categories shown (166 respondents).
Source: British Election Panel Study 2005–2010.

2005 became non-voters in 2010, but fewer than 2 per cent of the old and new
middle class groups who voted Labour in 2005 did not vote at the
next election.
What does this tell us? Chapter 7 showed how the decline of class voting is
partly a product of middle class people being more likely to support Labour.
What we reveal here is that the disappearance of the class differential for
Labour is also partly a product of working class people ceasing to vote
altogether. This has produced the kind of class non-voting that has long
been prevalent in the US, but until 20 years ago was barely noticeable in
Britain. Just as the changes to the Labour party in the 1990s and 2000s
attracted middle class voters, they put off working class voters. This is most
obvious when we look at people who stopped voting and their attitudes
towards Labour. The 1997–2001 BES panel study asks people whether they
think Labour ‘looks after working class people’ in 2001. Figure 8.4 shows
whether people who voted Labour in 1997 agree with this statement. We
separate people by occupational class group and their eventual 2001 vote
choice. Essentially everyone who stuck with Labour, regardless of their class,
thought Labour looked after working class people. On average 93 per cent of
the middle class groups and 95 per cent of the working class agreed. This figure
is somewhat lower for middle class people who stopped voting in 2001 (82 per
cent on average), but substantially lower for working class people who left
Labour for the ranks of the non-voting in 2001. Nearly 40 per cent of working
class voters who voted Labour in 1997, but stayed at home in 2001, thought
Labour did not look after working class people.
There is no matching effect for middle class people in perceptions of
whether Labour looks after the middle classes. Labour loyalists consistently
think that Labour is a party that helps the middle class (94 per cent of people
in middle class jobs and 89 per cent of people in working class jobs agree with
this) and the figures among those who switch from Labour to abstention are
very similar (91 per cent for the middle class occupational groups and 81 per
cent for the working class occupational group). As the changes that we

176
 Long Live Class Politics

a) Voted Labour in 2001 b) Did not vote in 2001

Average = 93% Average = 82%
100% 95% 100%

75% 75%
63%

50% 50%

25% 25%

0% 0%
OMC NMC JMC WC OMC NMC JMC WC

Figure 8.4. The proportion of people who voted Labour in 1997 and think that Labour
‘looks after working class people’ well in 2001 by occupational class and 2001 vote
choice
Note: The figures here show the percentage of 1997 Labour voters who agreed that the Labour Party ‘looks after
working class people’ in 2001. The left-hand graph gives results for those who voted Labour in 2001; the right-
hand graph gives results for those who did not vote. The total unweighted number of 1997 Labour voters in the
sample is 947. Not shown are people who voted for parties other than Labour (16 per cent of the sample) or people
who could not be placed in the four occupational class categories shown (183 respondents). Four occupational
class groups are displayed: old middle class (OMC), new middle class (NMC), junior middle class (JMC), and
working class (WC).
Source: British Election Panel Study 1997–2001.

discussed in Chapters 5 and 6 occurred, working class Labour voters stopped

seeing Labour as a suitable repository for their votes. Given that there were no
more appetising choices on the table in terms of viable parties that either
offered policies that were appealing to the working class or talked about
representing the working class, the only option was to exit the system. This
is what happened. The winning ‘party’ among working class voters today
(whether measured by education or occupation) is none of the above.
If this was being written in 2010, we would leave it there, but the 2015
election saw further changes to vote choices. Why? Because there were now
viable alternative options to the three main parties. The next part of this
chapter therefore turns to how class voting continues to evolve with the
emergence of new class-based parties.

Fruitcakes, Loons, and Nationalists

The 2015 election saw a small increase in turnout (from 65 per cent in 2010 to
66 per cent in 2015). The big story of the 2015 election, however, was the rise
of non-mainstream parties. In Scotland, the SNP took 56 of the 59 seats
available with over half of Scottish voters casting their vote for the

177
The New Politics of Class

nationalists. In England and Wales, UKIP increased their share of the vote by
10 percentage points, putting them in third place on nearly 15 per cent of the
vote. Both of these changes have something to tell us about class voting in
Britain and particularly the continued decline of Labour as a working class
party. While the 1997–2010 period is characterized by working class voters
leaving Labour to join the ranks of non-voters, the 2010–2015 electoral
cycle is one in which working class voters that who stuck with one of the
major parties departed to support the SNP in Scotland and UKIP in England
and Wales.
The rise, fall, and rise again of the SNP makes for a fascinating political tale.
Members of the party were ‘generally regarded as cranks’ (Kellas 1968, p.202)
before the 1960s, with little electoral support and little organization. Mitchell
et al. (2012, p.22) suggest that the party had only ‘200 members in the late
1950s’ and until the 1964 general election the SNP only contested a handful of
seats. Yet a series of local election successes and a by-election win in 1967 were
a precursor to unprecedented electoral success in the two 1974 elections.
It almost became the biggest party, by votes, in Scotland at the October
1974 election with over 30 per cent of Scottish votes cast for the nationalists,
which was less than 6 per cent behind Labour. While many different reasons
were given at the time for this success, some to do with the mainstream parties
and their record in government, some to do with better SNP organization, and
some to do with North Sea oil, much of that success evaporated fairly quickly
(see Lynch (2002) for a full account). In fact, from 1979 to 2010 the SNP rarely
got more than 20 per cent of the votes at Westminster elections. Nonetheless,
from the 1970s onwards Scotland clearly had a four party system, with the
Conservatives, the Liberals, and the SNP vying for second place behind a
dominant Scottish Labour party. Who then was voting for the SNP?
One might be forgiven for thinking that the SNP has always attracted
working class voters. In fact, the history of Scottish nationalism is not a
history of working class activism. While Scottish nationalists might talk of
working class Scotland, historically their votes have actually been drawn from
all classes. Most literature emphasizes the way in which social cleavages8 such
as class fail to shape SNP support. Writing in the 1990s, Newman argues that
the ‘SNP was not dependent on one particular class for the votes it received’
(Newman 1992, p.14) and that this was not surprising given that the ‘SNP
social and economic programme did not highlight class differences . . . the
villain of SNP propaganda was not the upper class or the lower class but the
centralized British state’ (p.15). This is also what we find, at least until 2010.
Table 8.3 shows SNP vote share by occupational class from when the SNP
first emerged as a political force in the 1970s. The number of Scottish respond-
ents in the 1970s and 1980s surveys is small, so we combine the two 1974 and
1979 elections (at which the SNP did rather well, receiving 30 per cent of the

178
 Long Live Class Politics

Table 8.3. SNP vote share by occupational class and education

1970s 1980s 1992 1997 2001 2005 2010 2015

Old middle class 16% 11% 15% 18% 14% 22% 13% 32%
New middle class 28% 12% 23% 15% 13% 11% 16% 37%
Junior middle class 19% 7% 19% 16% 18% 13% 8% 34%
Average (middle class groups) 21% 10% 19% 17% 15% 15% 12% 34%
Working class 21% 7% 21% 16% 16% 16% 17% 42%
Middle class average - working 0% 3% 2% 1% 1% 0% 4% 8%
class

Note: The numbers here are the proportions of people who voted for the SNP as a percentage of the Scottish electorate.
The 1970s numbers are the average for the two 1974 elections and the 1979 election, the 1980s numbers are the
average for the 1983 and 1987 elections.
Source: British Election Studies 1974–2015; British Social Attitudes Surveys 1997–2015.

Scottish share of the vote in the October 1974 election), and the 1983 and
1987 election surveys (at which the SNP did rather badly, receiving less than
10 per cent of the vote).9 What the table shows is that differences by occupa-
tional class in SNP support are essentially zero before 2010, but are of modest
importance in 2010 and 2015, with the working class being more likely to vote
for the SNP than the middle class groups. These differences would be more
striking if we looked at shares of the Scottish vote rather than shares of the
Scottish electorate as the proportion of working class people not voting is
much higher. In fact, nearly 60 per cent of working class voters voted for the
SNP in 2015, compared to fewer than 45 per cent of voters in the three middle
class groups.
Holding constant other factors does not affect this change; if anything it
makes it more obvious. Figure 8.5 shows the predicted probability of SNP
support by occupational class from logit models that hold constant education,
age, religion, and other important factors.10 The lines at the bottom show the
differences between the three middle class groups and the working class. All
show the increasing differences between the working class and the other
groups after 2005. Ultimately, the SNP has gained votes among all classes in
Scotland over the last ten years, but it disproportionately gained votes among
the working class. This matches perceptions of the party. Unfortunately, we
cannot track perceptions over time, but the 2015 BES asked whether people
thought of the SNP as a working class or middle class party. Whereas 67 per cent
of Scots thought it represented the working class, only 34 per cent thought it
represented the middle class. Although SNP support has not historically been
class-based, Scottish nationalists have traditionally ‘emphasize[d] their concern
with the working class community’ (Brand et al. 1994, p.629).
At the same time, we should not exaggerate these changes. A revived post-
referendum SNP meant that a new option was available to Scottish voters in
2015 and to some extent that option was more appealing to the working class

179
The New Politics of Class

60%
WC
OMC

40%

JMC
20%
NMC

0%
2000 2005 2010 2015

Differences between the

middle classes and
the working class
–20%

Figure 8.5. SNP support by occupational class

Note: The figure here shows predicted probabilities of SNP support in Scotland from multinomial logit regression
models that predict vote choice as a percentage of the electorate for each year separately. As well as occupational
class, these models include controls for education, trade union membership, gender, age, religion, and race. The
predicted probabilities are for a white man with no religion in his forties, who has middling educational attain-
ment and is not a trade union member. Four occupational class groups are displayed: old middle class (OMC), new
middle class (NMC), junior middle class (JMC), and working class (WC).
Source: British Election Studies 2001–2015; British Social Attitudes Surveys 2001–2015.

than the middle classes. But the SNP relied on mostly middle class voters to
make its electoral breakthrough. The perception that it is ‘for the working
class’ is actually weakest among working class people (52 per cent of working
class people see it as a working class party compared to nearly 80 per cent of
the old and new middle classes). Nor was this breakthrough based on under-
lying attitudinal differences between the classes. There is no greater enthusi-
asm for Scottish independence among the working class than among the
middle class. The 2015 BES shows that of people who voted in the referendum
in 2014, 44 per cent of people with working class jobs claimed to have voted
for an independent Scotland: exactly the same percentage as the overall
Scottish sample.
Probably more important were perceptions of Labour. In 2015 only 34 per
cent of working class Scots saw Labour as a party for the working class (and of
those nearly two thirds said it was only ‘slightly working class’). The rising tide
of Scottish nationalism did affect working class voters more than middle class
voters, but this was more due to changing perceptions of Labour than percep-
tions of the SNP. As Brand et al. (1994) pointed out in the 1990s, long before
the SNP breakthrough of the twenty-first century, the SNP and Labour had

180
 Long Live Class Politics

similar policies and offered similar rhetoric. What has changed is the position
in which the two parties find themselves. Twenty years ago Brand et al. (1994,
p.629) said that ‘the important cards are in Labour’s hands’ because it had
‘much more media exposure than the Nationalists’, ‘a tradition of support
among Scottish voters’, and that the SNP had not built ‘secure bases of local
support’, partially because it did not have a very happy record in local gov-
ernment. All those factors have changed. The Scottish Parliament has meant
that the SNP has a strong base of support in Scotland and that ‘traditions of
support’ have been tested. Equally, the independence referendum meant that
the SNP built up a formidable campaigning organization and dominated
media coverage. It is this shock to the system that generated realignment of
class voting in Scotland. Nevertheless, we should not overstate this. The lines
of party conflict have been redrawn to some extent, but there has certainly not
been a complete redefinition of class politics in Scotland.
What of England and Wales? The story of UKIP south of the border is not
the same as the SNP north of the border. The insurgent party here is newer and
differs from the SNP in that its class appeal is more pronounced: UKIP has
distinctive policy stances that more clearly and directly appeal to working
class voters. This is not just about rejecting Labour, but also embracing UKIP.
In the run-up to the 2015 election it looked remarkably like David Cameron’s
dismissal of UKIP as ‘fruitcakes, loonies and closet racists’ a decade earlier
(Ford and Goodwin 2014, p.71) would come back to haunt him. In the end,
the British electoral system translated UKIP’s four million votes into one seat.
Moreover, while many of those four million voters were Conservative defect-
ors, many were also Labour and Liberal defectors.11 In fact, there was a
substantial renewal of class voting in England and Wales in 2015 because it
was working class voters who were predominantly drawn to UKIP. This is not a
new point, nor is it specific to Britain. The links between class and radical right
parties are well known both historically in Britain and comparatively on
the continent.
Support for radical right parties, commonly defined as parties that oppose
immigration and multi-culturalism, is low among new middle class and
highly educated voters in almost all European countries (Arzheimer 2009;
Oesch 2008; Ivarsflaten 2005, 2008; Lucassen and Lubbers 2012). Indeed,
the academic consensus today is that core radical right support is located
within the less educated working class and those voters are primarily mobil-
ized by the issue of immigration (Mudde 2007; Rydgren 2008). This has
traditionally been the case in Britain as well, although until recently parties
on the extreme right have been relatively unpopular with everybody in
Britain. This is partially because of the electoral system and partially because
of these parties’ association with fascism and some rather nefarious characters.
The first past the post electoral system has not allowed parties to flourish and

181
The New Politics of Class

establish an electoral base (Norris 2005), and the lack of a radical right party
with, what Ivarsflaten (2006) calls, a ‘reputational shield’ has meant that
extremist parties on the right have been easy to tar as ‘fascists’ or ‘Nazis’.
While the National Front experienced some success in the mid 1970s, primar-
ily driven by younger working class voters in inner cities (Harrop et al. 1980),
this did not last.12 More recently, the British National Party managed to
mobilize a small percentage of voters over the immigration issue at European
and local elections in the 2000s, with the BNP regularly getting over 10 per
cent of the vote in local elections over the decade. The class basis of BNP
support, like National Front support, was again working class and less edu-
cated men, albeit older rather than younger men (Ford and Goodwin 2010;
Cutts et al. 2011). Ultimately the lack of national representation and links
with violence and overt racism now appear to have put paid to the BNP as a
longer-term project, however.
UKIP is of course a very different beast to the BNP and the National Front.
Its policies around immigration are not racist, and it has another dominant
policy issue of opposition to EU integration. Like other parties in Europe that
started as anti-EU parties, before adopting an anti-immigration platform,
UKIP has been largely protected from vote shedding accusations of racism.
In the parlance of the literature it is a ‘populist’, not a ‘neofascist’ party like
the BNP (Golder 2003; see Ford et al. (2012) about UKIP specifically). A recent
influential book by Ford and Goodwin (2014) suggests that UKIP has thus
become a party similar to many successful populist radical right parties on
the continent. Part of that similarity is the mobilization of working class and
less educated voters. Although Ford and Goodwin show links between occu-
pational class, and more importantly education, and UKIP support, almost
all their data comes from a period before UKIP’s sustained success outside of
European elections. Here we confirm to a large extent the predictions that
they made three years ago about the role of social characteristics in predict-
ing UKIP support.
Figure 8.6 shows the basic picture using yearly BSA data on party support by
occupational class and education. The rise of UKIP is obvious. Also obvious is
the growing occupational and educational differential in its support. These
effects are not due to the slightly older nature of the UKIP electorate, nor its
regional basis. Table 8.4 shows the effects of occupational class and education
on UKIP party choice from models that hold constant the usual long list of
other social characteristics. We split the data into two time periods: dormancy
(2004–2013) and success (2013–2015). The first column shows pooled data
from 2004–2012 for the BSA and the 2005–2010 elections for the BES. The
second column shows pooled data from 2013–2015 for the BSA and the 2015
election for the BES.13 UKIP support is generally very low in the dormant

182
 Long Live Class Politics

a) Occupational class b) Education

WC
Low
10% 10%

Medium
OMC

5% 5%
NMC
High

0% 0%
2002 2012 2002 2012

Figure 8.6. UKIP support by occupational class and education in England and Wales
Note: The figures here show the proportion of people who support UKIP as a percentage of the English and Welsh
electorate. The left-hand graph disaggregates results by social class; the right-hand graph disaggregates results by
education. Three occupational class groups are displayed: old middle class (OMC), new middle class (NMC), and
the working class (WC). Three educational groups are displayed: people with degree-level education (high), people
with A Level equivalent education (medium), and people who left school at the minimum school leaving age for
their cohort (low).
Source: British Social Attitudes Surveys 2004–2015.

pre-2013 period, but it is interesting that there are very few consistent differ-
ences between classes and educational groups.
After 2012 both the BSA and BES data show that occupational class and
education have a clear role to play in explaining UKIP support. The three
middle class occupational groups and people with a degree are systematically
less likely to support UKIP. The pattern for occupational class is a facsimile of
Labour support of yesteryear. Is this class voting then? It is, but for different
reasons than we saw previously. Labour class voting was based on left-wing
policies appealing to left-wing (and therefore working class) voters. UKIP also
offer policies that are particularly appealing to those in working class occupa-
tions. The difference is that these are not left-wing economic positions, but
rather anti-immigration and anti-EU integration policies. While the policy
area might be different, voter attitudes are nonetheless strongly shaped by
class and education, particularly in the last decade. As discussed in Chapter 4,
both economic and cultural factors drive people’s attitudes towards the EU
and immigration. Middle class, degree-educated people gain economically
from EU integration and mass immigration, and people in working class
jobs or without higher education lose (Anderson and Reichert 1995; Gabel
and Palmer 1995; Hooghe and Marks 2004; McLaren 2015). Equally the

183
The New Politics of Class

Table 8.4. UKIP support by occupational class in England and Wales

Support UKIP (BSA data) 2004–2012 2013–2015

Old middle class 2% 9%

New middle class 2% 10%
Junior middle class 3% 11%
Working class 2% 15%
New middle class – working class 0% 5%
High education (degree) 2% 6%
Medium education (A Level) 3% 11%
Low education (school leaving age) 3% 14%
High education – low education 1% 8%

Vote UKIP (BES data) 2005 and 2010 elections 2015 election

Old middle class 1% 12%

New middle class 2% 14%
Junior middle class 2% 13%
Working class 2% 23%
New middle class – working class 0% 9%
High education (degree) 2% 3%
Medium education (A Level) 0% 13%
Low education (school leaving age) 1% 12%
High education – low education 0% 9%

Note: The numbers here are predicted probabilities of UKIP support in England and Wales from logit regression models.
As well as occupational class and education, these models include controls for housing, trade union membership,
gender, age, region, religion, and race. The models using BSA data also include controls for employment sector and
agricultural employment. The predicted class probabilities are for a white man with no religion in his forties, who is a
homeowner, has middling educational attainment, lives in the south east of England, is not a trade union member, and
works in the private sector. The predicted education probabilities are for the same type of person in the junior middle
class category.
Source: British Election Studies 2005–2015; British Social Attitudes Surveys 2004–2015.

cultural threat that integration and immigration pose to people is greatest

among people who have the strongest group, cultural, and national identities.
These people are typically working class and do not have degrees (McLaren
2002, 2015; McLaren and Johnson 2007; Garry and Tilley 2009; Sides and
Citrin 2007).
This helps to explain the pattern of support by education. Degree-level
education tends to make people different in their attitudes towards immigration
and the EU, and it is this split that really matters for UKIP support. Whereas
Labour traditionally drew support from highly educated people (because it was
a little more socially liberal than the Conservatives) and also people with low
levels of education (because of its left-wing economic policies), UKIP draws
support from the 75 per cent of people without a degree relatively evenly.
It is perhaps not surprising then that if we hold attitudes towards immigration
and EU integration constant, most of the differences between the educational
and occupational groups disappear. Figure 8.7 shows predicted probabilities of
UKIP support from models that do (on the right) and do not (on the left) control

184
 Long Live Class Politics

a) BSA data without attitudinal controls b) BSA data with attitudinal controls
Occupation Education Occupation Education
20% 20%
WC Low
WC
NMC NMC Low
10% 10%
High
High

0% 0%

–10% –10%

c) BES data without attitudinal controls d) BES data with attitudinal controls
Occupation Education Occupation Education
20% 20%
WC
WC
NMC Low
NMC
10% 10%
Low
High
High

0% 0%

–10% –10%

Figure 8.7. UKIP support by occupational class and education in England and Wales
controlling for attitudes to immigration and the EU
Note: The figures here show predicted probabilities of UKIP support in England and Wales from logit regression
models. The top two graphs are based on BSA data; the bottom two graphs are based on BES data. As well as
occupational class and education, these models include controls for housing, trade union membership, gender,
age, region, religion, race, and agricultural employment. The models using BSA data also include controls for
public sector employment and year. Both figures on the right also include three attitudinal variables: attitudes
towards immigration, attitudes towards the EU, and left-right values. The predicted class probabilities are for a
white man with no religion in his forties, who is a homeowner, has middling educational attainment, lives in the
south east of England, is not a trade union member, and works in the private sector. The predicted education
probabilities are for the same type of person in the junior middle class category. All attitudes are held at their
mean. Two occupational class groups are displayed: new middle class (NMC) and the working class (WC). Two
educational groups are displayed: people with degree-level education (high) and people who left school at the
minimum school leaving age for their cohort (low).
Source: British Election Study 2015; British Social Attitudes Surveys 2013–2015.

for these attitudes. For both the pooled 2013–2015 BSA data and the 2015 BES
data, holding constant EU and immigration attitudes removes most of the
differences between groups, despite having to use rather crude measures of
both attitudes.14 This makes sense. UKIP attracts voters who are most opposed

185
The New Politics of Class

to immigration and EU integration. For example, at the 2015 election 79 per cent
of UKIP voters felt very strongly that immigration should be reduced compared
to 38 per cent of the whole electorate (and only a quarter of Labour voters).
Equally, 53 per cent of working class people very strongly agreed that there
were too many immigrants compared to 26 per cent of the new middle class
group. Put together these two facts—policy differences between classes and a
party offering a distinctive set of policies which appeals to one class more than
another—and you get a recipe for UKIP class voting.
Is this a new class politics? In 2013 Betz and Meret claimed that the ‘the
working class profile is valid for virtually all right-wing populist parties’
(p.108), and it is clear that UKIP is no exception to that rule. But the basis of
this class politics is rather different to the past. Rather than being based on
economic issues of redistribution and public ownership, it is based on issues
that, while linked to economic factors, are about cultural identities as well. It is
also not a mobilizing form of class politics; the rise of UKIP does not appear to
have ameliorated the class and education differences in turnout that we saw
earlier in this chapter.

Conclusions

‘Saturday Night and Sunday Morning’ was a landmark novel of the 1950s,
arguably one of the first authentic accounts of working class life. In the
penultimate chapter, Arthur says of politicians: ‘They shout at you from
soapboxes: “Vote for me, and this and that”, but it amounts to the same in
the end whatever you vote for’ (Sillitoe 1958, p.202). In 1958 Arthur was a
rebel, an exception. Most working class people voted, and most voted Labour.
In 2016 Arthur’s view would be shared by most others, doing the actual, or
modern equivalent of, sweating over a lathe. The new party of the working
class is no party at all. This is not due to differing amounts of ‘resources’ that
different people have. Resource models of turnout suggest education and
wealth allow people to engage more with politics in various ways, resulting
in higher rates of turnout for middle class educated people. This is not what we
see in Britain. Working class people were almost as likely to vote in the 1960s,
1970s, and 1980s as the middle classes. There was no gap in turnout to
explain. The gap only began to emerge in the 2000s. This is not because
working class voters have suddenly realized they lack ‘resources’; it is because
of their lack of political choices. With no party committed to representing
working class views, working class people choose not to vote.15 Writing about
the 1997 election Heath et al. (2001, p.155) argue that there were ‘hints’ of
working class disillusionment with Labour already and speculate that Labour’s

186
 Long Live Class Politics

changing nature would lead to ‘a gradual rise in class non-voting’. That

prediction has proved very accurate.
We started this chapter by presenting the options to voters who do not like
their former party: vote for the opposition, vote for a new party, or do not
vote. For most of the twenty-first century, the second of those options was not
a real possibility and working class voters who no longer felt attached to the
mainstream parties therefore chose option three: stay at home. At the last
election, real choice opened up, and the calculus changed for working class
voters. In Scotland, a revitalized SNP offered a real alternative to Labour which
attracted all voters, but disproportionately appealed to working class voters.
In England and Wales, UKIP became the party with, by a narrow margin, the
highest percentage of the working class vote of any party. While there are
similarities, there are also differences between the Scottish and English experi-
ences. In Scotland, the SNP is offering policies that are either very similar to
Labour or, like the issue of Scottish independence itself, do not strongly divide
people by class. In that sense, increased class voting in Scotland is due mainly
to the rejection of Labour as a working class party and the availability of a
substitute that happened to be relatively popular. In England and Wales,
voting for UKIP was not just a rejection of Labour, not least because much of
their support came from former Conservative and Liberal voters, but was also
an endorsement of UKIP policies that appeal directly to working class voters.
Is this a lasting change? We are writing this shortly before the 2016 EU
referendum, and it seems likely that the results of this may determine the
future electoral success of both the SNP and UKIP. It may have less effect on
class differences in their support. It seems difficult to imagine that Scottish
Labour will regain the title of the working class party and this means that
working class voters will likely stay with the SNP or stop voting, depending on
whether the SNP can sustain its universal appeal. The class basis of UKIP
support is stronger because there are real policy issues that divide the classes
and that divide UKIP from the other parties. Given the unpopularity of the EU
and mass immigration among working class voters, plus the seeming inability
of unwillingness of any other major party to articulate that unpopularity, the
continuation of class divisions in UKIP support seems likely. Whether UKIP as
a party can maintain its existence after the EU referendum, regardless of the
result, while handicapped by the electoral system, is much more difficult to
predict. What is more certain is that there will remain significant numbers of
voters, many working class, providing a potential pool of support for any
niche party advocating less immigration, less multi-culturalism, and greater
national sovereignty.
Nonetheless, what is crucial in that last sentence is the word ‘voters’,
because while UKIP may well continue to attract working class people to
defect from other parties, the biggest change in class voting patterns is the

187
The New Politics of Class

new bloc of working class non-voters. And for all the excitement that SNP and
UKIP success may have generated, it is this shift that is the most important. It
means that not only are working class people less likely to be voters due to
mainstream party changes but also that vote seeking parties have even less
reason to put forward policies that might appeal to those non-voters. To echo
Oliver Heath (2016, p.22) writing about turnout changes up to the 2010
election: ‘the working class have not become incorporated within the political
system, they have become marginalised from it.’

Notes

1. Or to quote Przeworski and Sprague (1986, p.61): ‘the workers who would other-
wise have voted for a Socialist party have three avenues open to them: they can
vote for bourgeois parties; they can abstain from voting altogether, and in some
countries, they can vote for other parties that appeal to them as workers.’
2. For example, in 1987, 3,104 people reported that they voted. After checking the
electoral records, the BES team found that 158 of those 3,104 people did not
actually vote. There were also a small number of people who reported not voting
when they did. Of the 514 people who reported that they did not vote, 29 actually
made it to the polling station according to the official records. It may be particularly
important to use validated votes when looking at turnout differences between
classes as there is evidence that middle class and highly educated people are more
likely to over-report their turnout (Bernstein et al. 2001; Silver et al. 1986; although
see Swaddle and Heath (1989) who find the opposite effect).
3. Although we can at least rule out institutional changes, and their consequent effect
on costs and benefits, as being responsible for decreased voting rates. It has been
argued that the lowering of the voting age to 18 before the 1970 election accounted
for the low levels of turnout at that election (Heath and Taylor 1999; Franklin
2004). However, there have been no large institutional changes since the 1970s
that could be expected to affect turnout. Indeed, if anything, the moves towards
increased postal voting should have increased turnout rates.
4. We run a model separately for each election, thereby allowing all the independent
variables to vary over time in their impact. Education is measured by qualification
(seven categories), age is six age-groups, region consists of the eleven standard
regions of Britain, religion is eight categories measuring religion and Christian
denomination, race is five categories. These are as reported in Appendix 4.1A for
the 1974–2015 time period. The most important of these control variables is age.
Although there is much debate in the literature as to whether it is ageing that
causes increased turnout (Glenn and Grimes 1968; Wolfinger and Rosenstone
1980; Plutzer 2002), or whether it is older generations of voters who are more likely
to vote (Clarke et al. 2004; Franklin 2004; Bhatti et al. 2012), we make no claims
one way or the other. By including age-groups and modelling turnout for each
survey separately, we are explicitly making no assumptions about whether it is

188
 Long Live Class Politics

really age or generation that matters. For the figures, we show predicted probabil-
ities for a white man in his forties, living in the southeast of England, with no
religion who is not a member of a trade union. The occupational class figure holds
education constant at O Levels (or equivalent) and the education figure holds
occupational class constant at the junior middle class category.
5. As Heath (2007, p.499) also points out, the resources model presents ‘something of
a paradox in relation to levels of turnout. Since the middle class are more likely to
vote than the working class, and there are more middle class people in Britain today
than there were previously, it would seem logical all other things being equal that
turnout should be higher’. Equally, over time the electorate has also become more
educated, yet we have seen decreasing political participation.
6. There is a similar pattern by educational attainment. Fewer than 10 per cent of
1997 Labour voters with a degree switch to non-voting in 2001, but 20 per cent of
1997 Labour voters with no qualifications become non-voters in 2001.
7. Unfortunately, the 2005–2010 panel has a rather limited measure of occupation
and so these categories are broad approximations of the occupational class categor-
ies that we use elsewhere in this chapter and the rest of the book, which are based
on the SEG. The 2005–2010 panel asks people to place themselves in a category
(with eight options), which inevitably means that there is much less measurement
precision. As one of the respondents themselves notes in an open-ended response
to the occupational class question: ‘your general options are frankly reductive.’
8. While most authors concentrate on class, as we do here, there were substantial
political divisions within Scotland along religious lines as well: Catholics sup-
ported Labour and Presbyterians supported the Conservatives (Seawright and
Curtice 1995; Seawright 2000; Tilley 2015). Some have argued that SNP support
is somewhat weaker amongst Catholics (McLean 1970; Bennie et al. 1997), but this
is disputed (Johns et al. 2010; Mitchell et al. 2012).
9. We combine the BES and BSA surveys for each election from 1997 onwards. Doing
this means that we have over 500 respondents for each year/period: 536 respond-
ents for the 1970s, 725 for the 1980s, 911 for 1992, 1,152 for 1997, 998 for 2001,
1,363 for 2005, 697 for 2010, and 531 for 2015.
10. We run a model separately for each election, thereby allowing all the independent
variables to vary over time in their impact. Education is measured by qualification
(seven categories), age is six age-groups, religion is eight categories measuring
religion and Christian denomination, race is two categories. For the figures, we
show predicted probabilities for a white man in his forties with no religion who is
not a member of a trade union and has education equivalent to O Levels.
11. Interestingly, it appears that many of the Liberal and Conservative defectors to
UKIP in 2015 were actually former Labour, and often working class, voters who had
left Labour in 2005 to vote for other parties in 2010 (Evans and Mellon 2016a).
12. The National Front may have been partially derailed by the Conservative party
adopting a somewhat tougher line on immigration: what Meguid (2005) would call
an ‘accommodative strategy’ by a mainstream party. In January 1978 Thatcher gave
a famous interview in which she sympathized with people who were ‘afraid that
this country might be rather swamped by people with a different culture’

189
The New Politics of Class

(Cockerell 1988, p.239) and, as Sandbrook (2012, p.593) points out, Keith Joseph
shortly afterwards sounded ‘rather like a diluted Enoch Powell’ when he said that
‘there is a limit to the number of people from different cultures that this country
can ignore’ in a by-election speech.
13. We run a model separately for the two periods, thereby allowing all the independ-
ent variables (sex, age, religion, trade union membership, race, and housing tenure)
to vary in their impact. We include sector of employment and agricultural employ-
ment for the models using BSA data. Education is measured by the five-category
variable that combines qualification and school leaving age, age is six age-groups,
region consists of the ten standard regions of Britain outside Scotland, religion is
eight categories measuring religion and Christian denomination, and race is two
categories. For the figures, we show predicted probabilities for a white man with no
religion in his forties, who is a homeowner, has middling educational attainment,
lives in the south east of England, is not a trade union member, and works in the
private sector. The predicted education probabilities are for the same type of person
in the junior middle class category.
14. For the BSA data we use a question that asks people what they think ‘Britain’s long-
term policy should be’ with options of ‘leave the EU, stay in the EU and try to reduce
the EU’s powers, to leave things as they are, to stay in the EU and try to increase the
EU’s powers, or to work for the formation of a single European government’ and two
0–10 scale questions which we combine which ask whether ‘it is generally bad or
good for Britain’s economy that migrants come to Britain from other countries’ and
whether ‘Britain’s cultural life is generally undermined or enriched by migrants
coming to live here from other countries’. For the BES data, we use a 1–5 scale
question that asks whether people ‘approve or disapprove of Britain’s membership
in the EU’ and a question that asks whether people think ‘too many immigrants have
been let into this country’. We also include the general measure of economic left–
right values in the models. In both cases it is not a strong predictor of UKIP support,
but UKIP supporters are, on average, slightly more economically left-wing.
15. This mirrors patterns found elsewhere. For example, Hill and Leighley (1996) find
that US states that have a more left-wing Democratic party are better able to
mobilize working class voters, and Weakliem and Heath (1999) show that class
non-voting in the US increases sharply after 1960 as the Democrats move away
from the New Deal working class coalition of voters.

190
9

Conclusion

In the 1950 general election ‘the party contest was not so much “wooing the
middle class” as more intense competition for the working class’ (Bonham
1954, p.35). The presence of a large working class shaped British party politics
for much of the post-war period. As we have seen, this has fundamentally
changed. Why has this happened? Class has not disappeared: objective
inequalities among classes, class identities, and ideological divisions between
classes are unchanged. Britain remains a class-divided society. It is the very
fact that class divisions have remained so pronounced that has produced
such important changes in the political parties. As the middle classes have
expanded, the policies and images that parties present to voters have changed
to accommodate this. Voters have noticed and reacted accordingly. In the
preceding chapters we have mapped these processes in detail.
In Part I of the book, we showed that differences over time among groups
based on occupational class and education are very static. In Chapter 2 we
focused on evidence of continuing class-based inequalities in resources,
opportunities, and risks. Whether it is income, unemployment, or health,
there are still systematic differences among class groups. Indeed, in some
respects these differences may have actually increased. Equally, perceptions
of income inequality remain largely unchanged, although awareness of
inequality is systematically less pronounced than the reality.
In Chapter 3 we documented the resilience of class identities and people’s
continued awareness of class. Changing class sizes and increased upward
social mobility do not appear to have dramatically weakened class identity,
although some people in middle class occupations are more likely to see
themselves as working class today. Conceptions of broad class divisions are
also persistent, as are perceptions of class barriers. The more day-to-day mani-
festations of class appear to show little sign of erosion as well. Most people
appear aware of class and our data appear to indicate that everyday manifest-
ations of class constraints, in terms of friendships for example, are still
present.
The New Politics of Class

This continued distinctiveness is also evident in people’s social and political

attitudes. In Chapter 4 we showed that ideological differences among social
classes have remained constant across the entire half century or so that we
examine. Persistent class divisions in resources, risks, opportunities, and edu-
cational attainment have fostered continuing differences in policy preferences
by class. These differences have combined with the changing sizes of classes to
produce a different average voter today than when Bonham was writing about
the 1950 election. As levels of education and the numbers of people with
middle class jobs have increased, the average voter’s views have become more
economically right-wing and more socially liberal. The working class and less
educated are now further away from the average voter. The electorate is also
more fractured than it was in 1950. Immediately after the war most people had
left school at fourteen and most people had a working class job. Today there is
no equivalent monolithic group that numerically dominates society.
The political elite have responded to these changes in class size by decoup-
ling class and politics. This can be seen in a number of ways. In Chapter 5 we
showed that while class, and especially the working class, was central to media
coverage of politics sixty or seventy years ago, today class is rarely mentioned.
It is particularly rare to find newspaper editorials that directly link parties and
class. This is not surprising given the evidence of party change in Chapter 6.
Here we saw three big changes to the main parties, especially Labour.
First, the policy stances of the parties converged during the 1990s towards
a more right-wing set of policy positions. The two main parties now cover a far
less extensive ideological range than was the case over most of the twentieth
century. The second change was not about ideology, but other types of signals
from parties to voters. We looked at how both manifestos and speeches
referred to different groups. Labour’s shift to being a party of the middle
class under Tony Blair was as much about the party’s social image as policy
change. While Labour, and to a lesser extent the Conservatives, regularly
referred to the working class in both speeches and policy documents in
the past, today there is little recognition of class. Both parties have settled
on more neutral terminology (their beloved ‘hard-working families’) rather
than making explicit appeals to class groups. The third change concerns descri-
ptive representation. More than ever politicians, and most notably Labour
politicians, are drawn from a similar pool of highly educated, upper middle
class people. All three of these changes have combined to affect voter percep-
tions of the parties. Chapter 6 showed that people today see the main parties
as both offering similar policies and representing similar types of (middle
class) people.
Chapters 7 and 8 showed the consequences of these changes. In Chapter 7
we demonstrated how the connection between class and party choice was

192
 Conclusion

static for most of the post-war period, until it dramatically weakened in the
mid 1990s. This timing indicates that the decline of class voting resulted
from the increasing shift to the centre by Labour and thus party convergence
on both policy and image. Because of party change, left–right ideological
positions, largely based on class, no longer predict vote choices with anywhere
near the same strength as before. Equally, as perceptions of the parties as class
parties disappeared, people were no longer able to match themselves to a party
on the basis of group identity. As Chapter 7 showed, the combination of these
two factors essentially accounts for all the change in class voting since the war.
There is a further consequence of Labour’s shift to the political centre and
their recasting as a party of the middle class. In Chapter 8 we explored how the
evolution of Labour has allowed new parties to flourish, but also depressed
voter turnout among certain groups. Both UKIP and, to a lesser extent, the
SNP in Scotland are examples of class-based parties. UKIP draws much of its
support from working class people. In that sense, class voting is alive and well.
Nonetheless, the major change is not really this re-emergence of class voting,
but the emergence of class non-voting. Chapter 8 showed that differences by
class in turnout rates through most of the post-war period were negligible.
This changed in the 2000s because of a lack of political choice. As Labour’s
policies moved to the right and its image became more middle class, working
class people have increasingly chosen not to vote.
What does this tell us about electoral politics? Ultimately our findings
illustrate the critical role of top down processes in accounting for major
electoral changes. Class differences in voting and non-voting depend upon
the choices offered to voters, as well as the presence of class differences. The
political choice thesis argues that when party signals to voters are indistinct,
the differences between the party choices of social groups will decline
(Przeworski 1985; Przeworski and Sprague 1986; Evans et al. 1999). As dis-
cussed in Chapter 1, many scholars have previously assumed a deterministic
‘bottom up’ explanation of the strength of the social bases of voting, in which
the transition to a post-industrial society has been accompanied by a blurring
of the class structure. This argument holds that class is not a source of political
preferences because it is no longer a source of identity and interests. Our
detailed analysis of the British case demonstrates that this is simply wrong.
It is mainly parties that shape class voting. The political choice approach also
explains increasing class differentials in turnout. They also account for
positions the preferences of classes are more likely to find expression than
when the parties cluster around similar positions. As Labour has changed,
working class policy preferences have been left out of the choices presented by
parties. A similar omission has occurred in terms of party signals about what
sorts of people they stand for. As the working class has become increasingly

193
The New Politics of Class

divorced from Labour, working class people have chosen not to vote. The
question that remains is whether this situation is likely to persist.

What Next?
Class Politics Within the Mainstream
The above discussion should not imply that Labour is unique among social
democratic parties. As Przeworski and Sprague (1986) observed thirty years
ago, the standard Social Democratic dilemma in post-industrial societies is
how to retain voters from the shrinking working class while also attracting
parts of the expanding educated middle class. As the latter group also forms
the recruitment pool for candidates, the tensions inherent in that trade-off for
Labour have presented the British media with wry amusement in recent years.
As we are writing this, Labour’s shadow Europe minister is in the news after
complaining that some voters in Derbyshire raised immigration as an issue:
‘The very first person I come to is a horrible racist. I’m never coming back to
wherever this is’ (Guardian, 19 May 2016).
Yet parties can change. This particular path has been chosen; it has not
been an inevitable consequence of social and economic development. Indeed,
the Labour Party changed its policy in many areas quite substantially after the
election of Jeremy Corbyn as leader in 2015. If Labour moves to the left then
might that reignite mainstream class politics? It might, but Labour under
Corbyn is not just more left-wing but also more socially liberal, especially on
key issues like immigration. This suggests that Labour will find it difficult to
regain its old ‘core’ voters who tend to be more socially conservative than the
average person. Discussion of the strategic difficulty of this combination of
policies has recently surfaced in academic journals (Bale 2016; Jackson 2016;
Richards 2016), but these tensions are also obvious when we look at current
survey evidence on how Corbyn himself is perceived.
Table 9.1 contains data from Waves 5 and 7 of the BES panel survey. This
shows how much people in different occupational classes like different party
leaders and whether they think Jeremy Corbyn would be a suitable prime
minister. Unfortunately, we are not able to separate the middle class into our
three standard groups using this data as it is based on the seven NS-SEC
categories. Nonetheless, we can still look at the difference between middle
class (broadly equivalent to our new and old middle class groups combined)
and working class people. Preferences for Corbyn as PM in 2016 are uniformly
lower than preferences for David Cameron, but, more interestingly, there are
few differences by occupational class in Corbyn’s appeal. Equally, although
working class people like him a little more than middle class people, this is no
different to evaluations of Ed Miliband in 2015. Although the evidence so far

194
 Conclusion

Table 9.1. Judgements of party leaders

Best PM? Middle class Working class All

David Cameron 41% 29% 36%

Jeremy Corbyn 19% 23% 22%
Neither 40% 48% 41%

Leader likeability (0–10) Middle class Working class All

David Cameron (2016) 3.9 3.1 3.7

Jeremy Corbyn (2016) 3.7 4.0 4.0
Ed Miliband (2015) 3.7 4.2 4.0

Note: The table shows responses to questions on ‘whether David Cameron or Jeremy Corbyn would make a better Prime
Minister’ and ‘how much they liked’ each leader on a 0–10 scale ranging from ‘strongly dislike’ to ‘strongly like’. The
distinctions between the middle class groups used elsewhere in the book are not yet available for these data so the classes
are constructed from NS-SEC categories: middle class (NS-SEC 1 and 2) and working class (NS-SEC 5, 6, and 7).
Source: British Election Study Panel Wave 5 (2015) and Wave 7 (2016)

is limited, there is little reason to believe that Jeremy Corbyn and what he
represents will affect working class non-voters, or change the pattern of class
voting.1
How long the current Labour leader will even stay in place is not certain
given that a large proportion of Labour MPs want to replace him. Perhaps we
should therefore ask the question: if Labour wants to maximize its votes, what
should it do? It is worth looking at the social composition of party support and
how this has changed. Figure 9.1 shows the make-up of the Labour and
Conservative electorate since 1964 in terms of occupational class and educa-
tion. The groupings for both are as we have discussed them in previous
chapters with two amendments. For occupation we show an ‘other’ category
which includes personal service workers, foremen, and own account workers,
and for education we distinguish between a lowest level (minimum school
leaving age) and a low level (above minimum school leaving age, but
below A Level).2
There have been dramatic changes. Labour used to receive less than a
quarter of its votes from the middle class and those with any education
beyond minimum school leaving age. Today two thirds of its votes are from
those groups. Some of this change is simply demographic; there are more
people with degrees, but the comparison with Conservative support is
instructive. The class mix for the Conservatives is quite similar over time,
and although the average educational level of Conservative voters has clearly
increased, changes to the educational composition of the Conservative vote
are much smaller than they are for Labour.
This illustrates how the logic of electoral competition has forced Labour to
focus on different voters. The new middle class and the highly educated now
form large Labour constituencies. Nonetheless, neither of these groups is

195
The New Politics of Class

a) Labour by occupational class b) Labour by educational group

100% 100%

WC
75% 75% Lowest

Other Low

50% 50%

25% 25%
JMC NMC
Medium
High
OMC
0% 0%
1964 2015 1964 2015

c) Conservative by occupational class d) Conservative by educational group

100% 100%

WC
Lowest
75% 75%
Other
Low
50% JMC 50%

NMC Medium
25% 25%

OMC High
0% 0%
1964 2015 1964 2015

Figure 9.1. Social make-up of Labour and Conservative voters

Note: The figures here show the proportion of voters for Labour and the Conservatives who come from different
occupational classes and educational groups. Five occupational class groups are displayed: old middle class (OMC),
new middle class (NMC), junior middle class (JMC), other, and the working class (WC). Four educational groups
are displayed: people with degree-level education (high), people with A Level equivalent or some higher education
(medium), people who left school above the minimum leaving age, but below eighteen (low), and people who left
school at the minimum school leaving age for their cohort (lowest).
Source: British Election Study 1964–2015; British Social Attitudes Surveys 2005 and 2010.

electorally ‘dominant’ in the way that the working class was in the 1960s. The
new middle class makes up only a quarter of the population today: this is
nothing like the situation fifty years ago when the working class made up over
half the population. Similarly, people with degrees are still only a quarter of

196
 Conclusion

the population, but in the 1960s nearly 70 per cent of the population had no
qualifications whatsoever. We have not seen a replacement of one big class
with another big class, but a change to a more heterogeneous mixture of class
and educational groups. This is inevitably a more difficult strategic situation
for the parties, as there are divisions within the middle classes, to some extent
due to educational differences.
Thus, the parties need to bundle together more complex coalitions, which
means balancing the preferences of these socially divided classes and educa-
tion groups. From a top down perspective, it is possible for parties to change
the distributions of their class coalitions by concentrating policy appeals and
mobilization, but this might not be worthwhile if groups are small. Moreover,
while the social composition of Labour voters has changed because of both
demographic change and because of Labour’s altered policies and image, it is
also due to turnout change. Parties need to consider the proportion of actual
voters within social groups. We saw in Chapter 8 that our stylized proletariat
group is now twice as likely as the intelligentsia to not vote. It is thus difficult
to see Labour wanting to compromise its appeal to the latter who vote in large
numbers, for the former who are more unlikely to turn out.
It might no longer make sense for the main party of the left to make
policy appeals to the working class, but this does not necessarily imply the
end of working class-oriented political appeals. Parties could try and reinvig-
orate the political significance of class identity. As we saw in Chapter 3, with
only gentle prompting, 60 per cent of the population think of themselves as
working class. Could the widespread extent of working class identity provide
an opening for a political strategy of the left? Probably not, because, as
Chapter 4 then showed, the middle classes continue to hold different policy
preferences to the working class. To try and bind such heterogeneous groups
together on the basis of identity is fraught with difficulty, not least when all
parties have for many years downplayed class identity politics. Just as
importantly, it is not clear that the proportion of people with working
class identities will remain high. As Chapter 3 showed, class identity is
often a residue of working class origins. Britain experienced a large amount
of upward mobility from the working class to the rapidly growing middle
classes in the second half of the twentieth century. Since the post-industrial
occupational transition has more or less run its course, the proportion of first
generation middle class people is already likely to be in decline. As this
number falls, so should the number of people with working class identities.
They will be replaced by a second generation middle class with a greater
degree of consistency between class origins and destinations, and a more
consistently middle class social and political outlook (De Graaf et al. 1995;
Tolsma and Wolbers 2010).

197
The New Politics of Class

Class Politics Outside the Mainstream

It seems unlikely that mainstream parties will revive class based appeals. This does
not necessarily spell the death of class politics, however. One obvious possibility is
that UKIP will become a distinctively working class party. In 2015 UKIP drew only
40 per cent of its support from the three middle class occupational groups. The
same figure for the three main parties was between 64 and 72 per cent. This echoes
the experience of other countries in which left-wing parties have moved to
centrist positions and lost their former working class supporters to radical right
parties (Achterberg and Houtman 2006; Kitschelt and McGann 2005; Rennwald
and Evans 2014).3 Indeed, it is now a standard finding that radical right parties are
disproportionately supported by the working class (Betz 1994; Kitschelt 1995;
Houtman 2003; Ivarsflaten 2005; de Lange 2007; Arzheimer 2008). In that sense,
continued working class support for UKIP seems highly likely.
However, there are future obstacles to UKIP providing an effective source of
working class representation. Once the party steps away from positions on the
EU and immigration, it is less clear that it is in tune with its core voters. UKIP
have no clear view on redistributive policies, for example, yet staying within
the laager of anti-immigration rhetoric impairs the party’s ability to break
through more generally in electoral terms.4 Radical right party success else-
where has often relied on the avoidance of strong stances on economic left–
right policies. But this is more viable in multi-dimensional PR systems where a
party can win votes on a single issue. It is more difficult to pull off this trick in
Britain’s majoritarian system, although the continued success of the Front
National in France suggests that it is not impossible.
Moreover, the effects of the referendum are unclear. The outcome may
invigorate UKIP. Yet that invigoration may reduce its class appeal if it adopts
a raft of consistent economic policies on the right. Equally, an implosion of
UKIP would probably reduce class voting, given that there is no obvious con-
tender in the wings waiting to take over its supporters. Indeed, it seems highly
likely that its voters, after having once abandoned mainstream politics, would
now abandon politics altogether and start to abstain.

Class Non-Politics
If UKIP voters were to stop voting, that would intensify class non-voting. But
can we make any predictions about the future of class differences in turnout
rates more generally? At one level, since parties seek to win elections, they care
less about groups that do not vote. This potentially leads to a spiral of exclu-
sion, in which parties do not represent certain types of people, those people do
not vote, and therefore parties become even less likely to represent those non-
voting groups. That spiral depends on whether those who have stopped

198
 Conclusion

voting can be persuaded to return. The prospects are not promising. Leighley
and Nagler (2014) find no increase in working class electoral participation in
the US over more than forty years. This is despite the ideological polarization
of the parties (Poole and Rosenthal 2011) over the last half of that period, a
change which should have increased expressive voting. Once the habit of
participation is lost, it appears to be difficult to recover.
This is not that surprising, since voting is often thought of as a habitual act.
A substantial body of recent research shows that the act of voting itself
increases the likelihood of voting in future elections (Green and Shachar
2000; Gerber et al. 2003; Spahn and Hindman 2014; Denny and Doyle 2009;
Plutzer 2002; see Aldrich et al. 2011 for a good overview). This can be due to
many factors: lower costs (knowing the location of the polling station),
increased benefits (a greater sense of regular voting as part of one’s self-
image), greater expressive voting (by strengthening a sense of partisanship),
and so on. Whatever the exact mechanism, the basic idea is that voting is a
habit. Once people are habitual voters they tend to remain voters unless there
are individual shocks within someone’s life (perhaps moving to a new area) or
shocks to the political system that destabilize partisanship (as we describe in
Chapter 8).
If voting is a habit, then it seems reasonable that non-voting may be
habitual as well. Gerber et al. (2003, p.540) explicitly say that habit works
both ways: ‘When people abstain from voting, their subsequent proclivity for
voting declines; when they vote, they become more likely to vote again.’ If
that is the case, then it means that shocks that produce decreases in turnout
among certain groups will continue to be felt many years later. Class non-
voting in Britain will thus continue. Related to this, many have argued that
the political context experienced during people’s formative years affects the
likelihood of entering into the voting habit (Franklin 2004; Gorecki 2013;
Plutzer 2002). Most research concentrates on electoral competitiveness, but
the wider political context of parties appealing to groups of voters would seem
just as important. If this is the case, then it means that working class people
who entered the electorate since the late 1990s are rather unlikely to ever start
voting.5 This might mean that class differentials in turnout will not only
remain but actually increase as newer generations replace older generations
in the electorate.

Conclusions

Modern Britain exemplifies the consequences of a top down influence on

cleavage dissolution and formation. When Tony Blair declared that the ‘class
war is over’ in 1999, he did so as an instigator of that ceasefire. Following the

199
The New Politics of Class

changes that Labour underwent in the 1990s, class voting decreased sharply.
By 2015 the social composition of Labour’s voters was not terribly dissimilar to
the Conservatives. Why did Labour change? The reason is a very simple one:
the shifting shape of the class structure. The transition from a big working
class to a big middle class was the engine of political change. The decisions
made by parties then constrained voters’ choices.
For the working class, these constraints are not merely about which party to
choose but also about no longer having any party to choose. Top down party
change has resulted in increasing non-participation in the democratic process
by working class and low educated voters. As we have discussed in this
chapter, it seems rather unlikely that these processes will be reversed. There
is little incentive for mainstream parties to start appealing to the working
class, and even if they did, voters who have stopped voting are unlikely to
start again easily. Anderson and Davidson (1943) described electoral politics in
the 1940s as ‘the democratic class struggle’. While that description might have
been true of Britain for much of the twentieth century, it is no longer. Politics
in Britain today involves middle class parties fighting it out for middle class
voters. The working class are bystanders, no longer represented within the
political mainstream.

Notes

1. It also seems unlikely that Scottish Labour will quickly recover its appeal to working
class voters in Scotland. The transformative effect of the 2014 Referendum has
restructured Scottish politics around the independence question (Fieldhouse et al.
2017). This suggests that the Scottish working class will either maintain their recent
support for the SNP, or eventually follow the English working class by increasingly
abstaining.
2. As in Figure 1.2, we include both people with A Level equivalent education and
those with some higher education below a degree in the medium category to
maintain continuity over time.
3. For example, Spies’s (2013) study of thirteen West European societies from 1980 to
2002 found that, in countries where the economic dimension of party competition
decreased, support for radical right parties was considerably higher among the
working class.
4. This has been exacerbated by the lack of political professionalism of the party’s
recently developed organization (Goodwin and Milazzo 2015).
5. This seems to be supported by the BES data. In 2015 the turnout differential between
the new middle class and working class was 14 per cent for people under forty but
only 3 per cent for people over sixty. Equally, the turnout differential between those
with degrees and those with no qualifications was 30 per cent for those under forty,
yet only 8 per cent for those over sixty.

200
10

Postscript: Brexit as an expression

of the ‘democratic class struggle’

Not long after we finished writing this book a remarkable political event
occurred in the UK. The outcome of the referendum on the 23rd June 2016
transformed the landscape of British politics and Britain’s relations with the
European Union in one fell swoop. It also provides an illustration of our argu-
ments about the suppression of political choice and its impact on class politics.
There has been much post-referendum discussion of Brexit and the social
divisions it has exposed. Birch (2016, p.107) refers to ‘the new cleavage that
has been revealed by the Brexit vote.’ Ford (2016), along with Goodwin and
Heath (2016), similarly argues that the result was driven by so-called ‘left-
behind’ voters: older, white, economically disadvantaged people who had
turned against a political class they regarded as privileged and out-of-touch.
Others have focused on divisions produced by globalisation (Curtice 2016),
while Bernstein (2016) and Dorling (2016) explicitly attribute the referendum
result to recent government austerity measures. Yet it would be wrong to
believe that these divisions are in any sense new, even if they have not been
so effectively expressed for many years. As we have seen in this book, Britain is
no more or less divided by class in 2016 than it was decades earlier. The
difference is that the referendum, with a clear choice between two competing
visions, allowed voters to give voice to these divisions. In particular, for the
first time since the political transformation of the Labour Party in the 1990s,
the preferences of the working class could be expressed unambiguously at the
ballot box.
It should not surprise us, therefore, to find a pronounced effect of class
and education upon people’s vote choice in the referendum. This can be seen
using data at the council level and at the individual voter level. If we take every
council area that reported in the referendum, there is an obvious connection
between the social make-up of the area and the percentage of people
who voted to remain or leave. Table 10.1 shows how the educational and
The New Politics of Class

Table 10.1. Proportion of people voting Leave in the 2016 EU referendum

% working class (N) % vote Leave % with degrees (N) % vote Leave

<20% (62) 42% <20% (62) 66%

20–25% (100) 52% 20–25% (90) 59%
25–30% (107) 57% 25–30% (95) 54%
30%+ (77) 63% 30%+ (99) 44%

Note: This table shows the proportions of people who voted to leave the EU by type of council area in England and Wales,
excluding the City of London. The proportion of working class people in a council area is derived from the 2011 census
and refers to the proportion of people in routine and semi-routine jobs. The proportion of people with degrees is also
derived from the 2011 census.
Source: BBC.

occupational character of council areas in England and Wales matches the

proportion of people who voted to leave the EU. In the 77 council areas with
more than 30 per cent of people in working class jobs, the Leave vote was
more than 20 per cent higher than in the 62 council areas with fewer than
20 per cent of the population in working class jobs. Equally, there is more than
a 20 per cent gap in the leave vote between the 62 areas with fewer than 20 per
cent graduates and the 99 areas with more than 30 per cent graduates.1 Most
of the geographical variation in the Leave vote can be explained by the social
make-up of different areas.
These effects of class and education can also be clearly seen at the individual
level. We can look at vote choices using data from the British Election Study
Referendum Survey carried out directly after the referendum. The differences
in vote choice by occupational class and education are enormous. While
72 per cent of people with no qualifications voted to leave, only 35 per cent
of people with a degree did. Equally, while 63 per cent of people in working
class jobs voted to leave, only 44 per cent of those in new middle class jobs did.
These differences persist even when we account for age and regional differ-
ences. The first column of Table 10.2 shows vote choices from a model that
separates the effects of education and occupational class, and also holds
constant other important control variables such as age and region.2 The effects
of education and occupation remain large. There is an almost 10 per cent gap
between the new middle class and the working class, and a 30 per cent gap
between people with high and low education.
These differences in referendum vote choice are similar to the divisions in
attitudes towards the EU and immigration that we saw in Chapter 4. This is
unsurprising since there is bound to be a direct link between these policy
preferences and vote choice in the referendum. The referendum gave people
an opportunity to choose between two distinct outcomes rather than the
constrained options offered by the main political parties in recent elections.
All of the three main parties3 were committed to remaining in the EU and were
unwilling or unable to reduce immigration. The second column of the table

202
 Postscript: Brexit as an expression of the ‘democratic class struggle’

Table 10.2. Proportion of people voting Leave in the 2016 EU referendum

% vote Leave Without attitudinal controls With attitudinal controls

Old middle class 52% 52%

New middle class 47% 50%
Junior middle class 50% 48%
Working class 55% 51%
New middle class – working class 8% 2%
High education (degree) 36% 44%
Medium-high education (A Level) 50% 45%
Medium-low education (O Level) 60% 53%
Low education (no qualifications) 66% 58%
High education – low education 30% 14%

Note: The numbers here are predicted probabilities from logit regression models to predict a Leave vote. As well as
occupational class and education, all models include controls for housing, gender, age, region, religion, and race. The
results in the right-hand column are from a model that also includes four attitudinal variables: attitudes towards
immigration on a 1–10 scale, attitudes towards the EU on a 1–10 scale, left–right values and liberal–conservative values.
The predicted class probabilities are for a white Anglican woman in her 40s who is a homeowner, has middling
educational attainment, and lives in the south east of England. The predicted education probabilities are for the same
type of person in the junior middle class category. All attitudes are held at their mean.
Source: British Election Study 2016 post-referendum survey.

shows class and educational differences in vote choice once we account for
people’s attitudes towards the pooling of sovereignty across the EU and atti-
tudes towards immigration.4 Immigration and the reclamation of powers from
EU institutions were the two most frequently cited concerns among those
who voted Leave (Prosser et al. 2016), and, as Chapter 4 showed, both these
attitudes are strongly linked to class and education. Even though the attitu-
dinal measures we have available are fairly poor (being simply 1–10 scales
upon which people place themselves), once we account for them, we see
much smaller differences in referendum vote between social groups. Class
and education shape attitudes towards immigration and the pooling of sov-
ereignty, and those attitudes were crucial for the decision to vote Remain or
Leave. How does this link to the argument that runs through this book? In
essence, it shows that when people in different social classes and educational
groups are given a choice, they express their different preferences by choosing
different outcomes. Removing the top down suppression of social divisions in
preferences by parties allows those class divisions to be expressed.
One question that might remain, so to speak, is why did Leave win if the
working class is no longer demographically dominant? After all, it is the
change to an increasingly highly educated and middle class electorate that
we have argued renders a working class focused politics electorally non-viable.
There are two answers to this question. The first is to make clear that while
divisions on an issue might be class based, the level of general opposition or
support is also important. Working class voters are clearly less favourable to
immigration than middle class voters, but that does not mean that the middle

203
The New Politics of Class

classes want more immigration. In fact, few people in the British electorate are
happy with current rates of immigration. The binary choice of a referendum
divides a population into two, regardless of the extremity of people’s issue
positions.
The second answer requires reflection on the extent to which changes in
educational qualifications have actually altered the electoral landscape. As
Chapter 4 showed, it is people’s education, and not so much their occupa-
tional class, which determines their views on social liberalism and EU
integration. There is a particularly large effect of higher education. However,
the fact that the number of people with degrees has increased does not mean
that most people have degrees. As Table 10.1 showed, in only 99 out of 346
councils in England and Wales are more than 30 per cent of the population
graduates. Taking the UK electorate as a whole, only slightly over a quarter of
people have a degree. As we discussed in Chapter 9, policies that only appeal to
the university educated, as opposed to the broader middle class, are liable to
remain relatively unpopular. EU membership is no exception to that rule.

Reversing the spiral of exclusion?

Chapter 8 focused on the decline in working class representation by the main

political parties and the consequent rise in working class non-voting. The
question that the referendum poses is, what happens when people’s prefer-
ences are represented on a level playing field? While turnout was higher at the
referendum than it was at the general election a year earlier for all types of
people, it seems clear that working class participation increased the most.
Again, we can look at both aggregate and individual level data. Table 10.3
shows how turnout changed in the 50 most and least working class areas and
the 50 areas with most graduates and 50 areas with fewest graduates.
The differences in levels of turnout between 2015 and 2016 by type of area
are not huge, but their direction is interesting. Areas with a large number of
middle class graduates saw a smaller increase in turnout than did areas with
few middle class graduates. This evidence of a renewal of participation in
poorer socio-economic areas is new. As Chapter 8 showed, increases in par-
ticipation after the nadir of 2001 have been far more pronounced among the
middle class and the highly educated than among the working class and less
educated. This has partially explained the growing class cleavage in participa-
tion. The disproportionate increases in participation between 2015 and 2016
in the working class and less highly educated areas of the country thus mark a
partial reversal of this trend.5
We can see something similar when we look at survey data as well.
Unfortunately the BES post-referendum survey is not suitable for examining

204
 Postscript: Brexit as an expression of the ‘democratic class struggle’

Table 10.3. Proportion of people voting

2015 election 2016 EU referendum Difference

50 most working class council areas 61.7% 70.1% +8.4%

50 least working class council areas 68.9% 75.9% +7.0%
50 council areas with fewest graduates 62.0% 70.3% +8.4%
50 council areas with most graduates 68.7% 75.3% +6.6%

Note: This table shows the proportions of people who voted in the 2015 general election and the EU referendum by type
of council area in England and Wales, excluding the City of London. The proportion of working class people in a council
area is derived from the 2011 census and refers to the proportion of people in routine and semi-routine jobs. The
proportion of people with degrees is also derived from the 2011 census.
Source : BBC.

turnout as it systematically oversamples voters, but we can use a wave of the

BSA (the NatCen Panel Pre-Referendum Survey) that happened just before the
referendum and that asked people how likely it was that they would vote. If we
run similar statistical models to those in Chapter 8 that hold constant age and
so forth, then we predict that 87 per cent of the intelligentsia group (highly
educated professionals) were likely to vote, compared to 67 per cent of the
proletariat (people in working class jobs with low levels of education).6 There is
clearly still a difference, but it is about half the difference (20 per cent com-
pared to 39 per cent) that we saw at the 2015 general election in Chapter 8.
There is good evidence that there was some narrowing of participation
differences by class in 2016 compared to the last few general elections.
When people are given a choice, and a choice that is not strongly linked
with allegiances to political parties, they can re-engage to some extent. The
growing habit of non-participation described in earlier chapters can be coun-
teracted, but only under very specific conditions. The referendum produced
high turnout and strong class voting because direct democracy did not enable
parties to restrict the choices available. In that sense, the Brexit vote clearly
confounds sociological, bottom up accounts of political change that assume
that the influence of social structure on politics is in terminal decline.
Whether the political class will allow voters such expression in the foreseeable
future remains to be seen. Yet without such direct democracy, it is difficult to
see how the pattern of working class disengagement from party politics shown
in Chapter 8 can be reversed. Theresa May’s avowal of one-nation Conserva-
tism, complete with references to the working class in her leadership accept-
ance speech and her first party conference speech as leader, is unlikely to be
pursued with sufficient effectiveness to overcome entrenched habits of work-
ing class non-voting. Indeed, much like UKIP’s appeal to working class voters
at the 2015 election, it is more likely to attract soft Labour partisans as the
Labour party (and UKIP) stumble through extended leadership crises.
The ongoing convulsions within the Labour Party are between the liberal
left and the Blairite right, neither of which provide much likelihood of

205
The New Politics of Class

representing the preferences of the working class, for whom both choices are
of mixed appeal. Jeremy Corbyn’s resounding victory in the 2016 Labour
leadership contest appears to have consolidated an economically left-wing,
but also, and more obviously, socially liberal platform. Criticisms of his
London-centred cabinet for its lack of connection with the provincial working
class have been expressed in the party and the media. Likewise, the party’s
support for Remain, low-key though it was, will have done little to endear it to
its residual working class base.7 At the same time, the option of choosing UKIP
is endangered by its potential redundancy—as a party with a one issue goal
that may now have been achieved—and the (current) loss of its high profile,
charismatic leader. As with Labour, it too has been beset by internal struggles
that threaten to tear it apart, so much so that it is not clear whether UKIP will
continue to be an effective electoral party in future elections. These problems
are likely to weaken its attraction for disillusioned working class Labour voters,
leaving them no viable choice at the ballot box. Thus, despite many working
class voters finding themselves on the winning side of the referendum vote,
the spiral of working class exclusion from broader electoral politics is likely
to continue.

Notes

1. Although both occupational class and education appear to be important when

holding the other constant, educational levels seem to matter more. This can be
seen when we model the Leave vote share for England and Wales using an OLS
regression. The independent variables were the proportion of people in working
class jobs, the proportion of people with degrees, the proportion of people over 65,
and the proportion of non-whites. This model predicts that a 10 per cent increase in
the proportion of people in working class jobs increases the Leave vote share by 3
per cent and a 10 per cent increase in the proportion of people with degrees
decreases the Leave vote share by 13 per cent. The adjusted R-square of this
model is 0.83 (0.80 with just class and education in the model), showing just
how important these characteristics are in explaining the geographical differences
in the vote.
2. The model holds constant a number of variables. These are: gender, age (six age
groups), region (the eleven standard regions of Britain), religion (eight categories
measuring religion and Christian denomination), housing tenure (five categories),
and race (five categories). Education is measured by a seven category variable based
on qualification and occupational class is based on the full NS-SEC measure.
3. Of course, UKIP arose in part to represent these anti-EU and anti-immigration
preferences and, as Chapter 8 showed, UKIP voting is strongly predicted by voter
attitudes towards those issues. Yet party politics is by its nature concerned with
multiple policy areas; a referendum offers a simple binary choice.

206
 Postscript: Brexit as an expression of the ‘democratic class struggle’

4. We also include measures of economic left–right values and liberal–conservative

values. These are batteries of questions that are almost identical to those discussed
in Chapter 4. Neither had a very strong effect on vote choice.
5. Again if we model these changes using an OLS regression model predicting turnout
change between 2015 and 2016, educational differences by area appear to be most
important. A simple model that includes independent variables measuring the
proportion of people in working class jobs and the proportion of people with a
degree gives a non-statistically significant effect of occupational class on turnout
change. Nonetheless education is clearly important: the model predicts that a 10 per
cent increase in the proportion of people with degrees would decrease turnout
change by 4 points.
6. The turnout question is a 1–10 scale that measures how likely someone is to vote. As
is commonly done, we take people who said 9 or 10 as likely voters and the rest as
likely non-voters. Taking just 10 as likely voters decreases the gap between the
intelligentsia and proletariat further to just 13 per cent. The models include age,
housing tenure, region, and sex as control variables, along with education and
occupational class. Education is measured by highest qualification and occupa-
tional class is based on the NS-SEC schema. The predicted probabilities cited in
the text are for a woman in her 40s who lives in the South East and owns her own
house. For more details of the data see Cabrera-Alvarez et al. (2016).
7. This seems unlikely to have the extreme consequences that followed from being on
the unionist side in the 2014 Scottish independence referendum (Fieldhouse and
Prosser 2016) however.

207
References

Abrams, Mark, and Richard Rose. 1960. Must Labour Lose? Harmondsworth: Penguin
Books.
Achterberg, Peter, and Dick Houtman. 2006. ‘Why do so many people vote “unnatur-
ally”? A cultural explanation for voting behaviour’. European Journal of Political
Research 45 (1): 75–92.
Adams, James, Jane Green, and Caitlin Milazzo. 2012. ‘Has the British public depolar-
ized along with political elites? An American perspective on British public opinion’.
Comparative Political Studies 45 (4): 507–30.
Adorno, Theodor W, Else Frenkel-Brunswik, Daniel J. Levinson, and Nevitt Sanford.
1950. The Authoritarian Personality. New York: Harper and Row.
Alderson, Arthur, Jason Beckfield, and Francois Nielsen. 2005. ‘Exactly how has income
inequality changed?’ International Journal of Comparative Sociology 46 (5–6): 405–23.
Aldrich, John H. 1993. ‘Rational choice and turnout’. American Journal of Political Science
37 (1): 246–78.
Aldrich, John H., Jacob M. Montgomery, and Wendy Wood. 2011. ‘Turnout as a habit’.
Political Behavior 33 (4): 535–63.
Alford, Robert. 1964. Party and Society: The Anglo-American Democracies. London: John
Murray.
Alt, James, Ivor Crewe, and Bo Särlvik. 1977. ‘Angels in plastic: The Liberal surge in
1974’. Political Studies 25 (3): 343–68.
Andersen, Robert, and Josh Curtis. 2012. ‘The polarizing effect of economic inequality
on class identification: Evidence from 44 countries’. Research in Social Stratification
and Mobility 30 (1): 129–41.
Andersen, Robert, and Tina Fetner. 2008. ‘Economic inequality and intolerance: Atti-
tudes towards homosexuality in 35 democracies’. American Journal of Political Science
52 (4): 942–58.
Andersen, Robert, James Tilley, and Anthony F. Heath. 2005. ‘Political knowledge and
enlightened preferences: Party choice through the electoral cycle’. British Journal of
Political Science 35 (2): 285–302.
Anderson, Christopher J., and M. Shawn Reichert. 1995. ‘Economic benefits and sup-
port for membership in the EU: A cross-national analysis’. Journal of Public Policy
15 (3): 231–50.
Anderson, Dewey, and Percy Davidson. 1943. Ballots and the Democratic Class Struggle.
Stanford: Stanford University Press.
References

Anderson, Lewis. 2016. ‘Inequality in children’s mental health and behavioural

problems: What are the trends?’ Centre for Social Investigation 19: briefing note.
Available online: http://csi.nuff.ox.ac.uk/?page_id=11.
Ansolabehere, Stephen, and Shanto Iyengar. 1995. Going Negative: How Political Advert-
isements Shrink and Polarize the Electorate. New York: Free Press.
Ansolabehere, Stephen, Jonathan Rodden, and James N. Snyder Jr. 2008. ‘The
strength of issues: Using multiple measures to gauge preference stability, ideological
constraint, and issue voting’. American Political Science Review 102 (2): 215–32.
Argyle, Michael. 1994. The Psychology of Social Class. Abingdon-on-Thames, UK: Taylor
& Francis.
Armstrong, David, and Ryan Bakker. 2006. ‘Take that you lousy dimension: A new
model for manifesto-based left-right party placement’. Unpublished working paper.
University of Oxford.
Arzheimer, Kai. 2009. ‘Contextual factors and the extreme right vote in Western
Europe, 1980–2002’. American Journal of Political Science 53 (2): 259–75.
Atherton, Kate, and Chris Power. 2007. ‘Health inequalities with the National Statistics
Socioeconomic classification: Disease risk factors and health in the 1958 British birth
cohort’. European Journal of Public Health 17 (5): 486–91.
Autor, David H., and David Dorn. 2013. ‘The growth of low-skill service jobs and the
polarization of the US labor market’. American Economic Review 103 (5): 1553–97.
Bakker, Ryan, and Sara Hobolt. 2013. ‘Measuring party positions’. In Geoffrey Evans
and Nan Dirk De Graaf (eds.), Political Choice Matters: Explaining the Strength of Class
and Religious Cleavages in Cross-National Perspective. Oxford: Oxford University Press,
27–45.
Bakker, Ryan, Catherine De Vries, Erica Edwards, Liesbet Hooghe, Seth Jolly, Gary
Marks, Jonathan Polk, Jan Rovny, Marco Steenbergen, and Milada Anna Vachudova.
2015. ‘Measuring party positions in Europe: The Chapel Hill expert survey trend file,
1999–2010’. Party Politics 21 (1): 143–152.
Baldassarri, Delia, and Andrew Gelman. 2008. ‘Partisans without constraint: Political
polarization and trends in American public opinion’. American Journal of Sociology
114 (2): 408–46.
Bale, Tim. 2016. ‘The loser takes it all. Labour and Jeremy Corbyn: A response to Steven
Richards’. The Political Quarterly 87 (1): 18–19.
Balestrini, Pierre P. 2012. ‘How citizens’ education, occupation, personal economic
expectations and national identity interact with one another to sway public opinion
on the EU’. Swiss Political Science Review 18 (3): 371–84.
Bara, Judith. 2006. ‘The 2005 manifestos: A sense of déjà vu?’ Journal of Elections, Public
Opinion and Parties 16 (3): 265–81.
Barry, Brian. 2001. Culture and Equality: An Egalitarian Critique of Multiculturalism.
Cambridge, UK: Polity Press.
Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age.
Princeton, NJ: Princeton University Press.
Baumberg, Ben, and Nigel Meager. 2015. ‘Job quality and the self-employed: Is it still
better to work for yourself?’ In Alan Felstead, Duncan Gallie, and Francis Green (eds.),
Unequal Britain at Work. Oxford: Oxford University Press, 105–219.

210
 References

Beck, Ulrich. 1992. Risk Society: Towards a New Modernity. London: Sage.
Beck, Ulrich, and Elisabeth Beck-Gernsheim. 2002. Individualization: Institutionalized
Individualism and its Social and Political Consequences. London: Sage.
Beckfield, Jason. 2003. ‘Inequality in the world polity: The structure of international
organization’, American Sociological Review, 68: 401–24.
Beckfield, Jason. 2006. ‘European integration and income inequality’. American Socio-
logical Review 71: 964–85.
Bell, Daniel. 1960. The End of Ideology. Glencoe, IL: Free Press.
Bell, Wendell, and Robert V. Robinson. 1980. ‘Cognitive maps of class and racial inequal-
ities in England and the United States’. American Journal of Sociology 86 (2): 320–49.
Bendyna, Mary, and Celinda Lake. 1994. ‘Gender and voting in the 1992 Presidential
election’. In Elizabeth A. Cook, Sue Thomas, and Clyde Wilcox (eds.), The Year of the
Woman: Myths and Realities. Boulder, CO: Westview, 237–54.
Bennie, Lynn, Jack Brand, and James Mitchell. 1997. How Scotland Votes: Scottish Parties
and Elections. Manchester: Manchester University Press.
Benoit, Kenneth, and Michael Laver. 2007. ‘Estimating party policy positions: Compar-
ing expert surveys and hand-coded content analysis’. Electoral Studies 26 (1): 90–107.
Berelson, Bernard, Paul F. Lazarsfeld, and William N. McPhee. 1954. Voting: A Study of
Opinion Formation in a Presidential Campaign. Chicago: University of Chicago Press.
Bernstein, J. 2016. ‘The Brexit and budget austerity: What’s the connection?’. The
Washington Post, 27 June 2016.
Bernstein, Robert, Anita Chadha, and Robert Montjoy. 2001. ‘Overreporting voting:
Why it happens and why it matters’. Public Opinion Quarterly 65 (1): 22–44.
Best, Henreich, and Maurizio Cotta (eds.). 2000. Parliamentary Representatives in Europe
1848–2000: Legislative Recruitment and Careers in Eleven European Countries. Oxford:
Oxford University Press.
Betz, Hans-Georg. 1994. Radical Right-Wing Populism in Western Europe. New York:
St. Martin’s Press.
Betz, Hans-Georg, and Susi Meret. 2013. ‘Right-wing populist parties and the working-
class vote’. In Jens Rydgren (ed.), Class Politics and the Radical Right. Abingdon:
Routledge, 107–21.
Beynon, Huw. 1973. Working for Ford. Harmondsworth, UK: Penguin Books.
Bhatti, Yosef, Kasper M. Hansen, and Hanna Wass. 2012. ‘The relationship between age
and turnout: A roller-coaster ride’. Electoral Studies 31 (3): 588–93.
Birch, Sarah. 2016. ‘Our new voters: Brexit, political mobilisation and the emerging
electoral cleavage’. Juncture 23 (2): 107–10.
Blanden, Jo, and Stephen Machin. 2007. Recent Changes in Intergenerational Mobility:
Report for the Sutton Trust. London: Sutton Trust.
Bolton, Paul. 2010. ‘Higher education and social class’. House of Commons Library: Research
Briefings. Available online: http://researchbriefings.parliament.uk/ResearchBriefing/
Summary/SN00620#fullreport.
Bonham, John. 1954. The Middle Class Vote. London: Faber and Faber.
Books, John W., and JoAnn B. Reynolds. 1975. ‘A note on class voting in Great Britain
and the United States’. Comparative Political Studies 8 (3): 360–76.
Bottero, Wendy. 2004. ‘Class identities and the identity of class’. Sociology 38 (5): 985–1003.

211
References

Brady, David. 2005. ‘The welfare state and relative poverty in rich western democracies,
1967–1997’. Social Forces 83 (4): 1329–64.
Brady, David, and Hang Young Lee. 2014. ‘The rise and fall of government spending in
affluent democracies, 1971–2008.’ Journal of European Social Policy 24 (1): 56–79.
Brand, Jack, James Mitchell, and Paula Surridge. 1994. ‘Social constituency and ideo-
logical profile: Scottish nationalism in the 1990s’. Political Studies 42 (4): 616–29.
Breen, Richard, and Jan O. Jonsson. 2005. ‘Inequality of opportunity in comparative
perspective: Recent research on educational and social mobility’. Annual Review of
Sociology 31: 223–43.
Brennan, Geoffrey, and Alan Hamlin. 1998. ‘Expressive voting and electoral equilib-
rium’. Public Choice 95 (1–2): 149–75.
Brooks, Clem, Paul Nieuwbeerta, and Jeff Manza. 2006. ‘Cleavage-based voting behavior
in cross-national perspective: Evidence from six postwar democracies’. Social Science
Research 35: 88–128.
Brynin, Malcom, and Kenneth Newton. 2003. ‘The national press and voting turnout:
British general elections of 1992 and 1997’. Political Communication 20 (1): 59–77.
Budge, Ian, Hans-Dieter Klingemann, Andrea Volkens, and Judith Bara. 2001. Mapping
Policy Preferences: Estimates for Parties, Governments and Electors 1945–1998. Oxford:
Oxford University Press.
Bukodi, Erzsebet, Robert Erikson, and John H. Goldthorpe. 2014. ‘The effects of social
origins and cognitive ability on educational attainment: Evidence from Britain and
Sweden’. Acta Sociologica 57 (4): 293–310.
Bukodi, Erzsebet, and John H. Goldthorpe. 2016. ‘Educational attainment – relative or
absolute – as a mediator of intergenerational class mobility in Britain’, Research in
Social Stratification and Mobility 43: 5–15.
Bukodi, Erzsebet, John H. Goldthorpe, Lorraine Waller, and Jouni Kuha.. 2015. ‘The
mobility problem in Britain: New findings from the analysis of birth cohort data’, The
British Journal of Sociology 66 (1): 93–117.
Bulmer, Martin (ed.). 1975. Working Class Images of Society. London: Routledge.
Busch, F. 2015. ‘Unemployment and insecurity in the UK labour market’. Centre for Social
Investigation 11: briefing note. Available online: http://csi.nuff.ox.ac.uk/?page_id=11.
Butler, David, and Gareth Butler. 2000. Twentieth Century British Political Facts
1900–2000. London: Macmillan.
Butler, David, and Dennis Kavanagh. 1984. The British General Election of 1983. Basing-
stoke: Macmillan.
Butler, David, and Donald E. Stokes. 1969. Political Change in Britain: Forces Shaping
Electoral Choice. London: Macmillan.
Butler, David, and Donald E. Stokes. 1974. Political Change in Britain: The Evolution of
Electoral Choice. 2nd edn. London: Macmillan.
Butler, Tim, and Michael Savage (eds.). 1995. Social Change and the Middle Classes.
London: UCL Press.
Cabrera-Alvarez, Pablo, Curtis Jessop, and Martin Wood. 2016. ‘Public opinion on the
EU referendum question: A new approach’. NatCen briefing paper.
Calvert, Peter. 1982. The Concept of Class: An Historical Introduction. London:
Hutchinson.

212
 References

Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960.
The American Voter. New York: Wiley.
Campbell, Rosie, and Philip Cowley. 2014. ‘What voters want: Reactions to candidate
characteristics in a survey experiment’. Political Studies 62 (4): 745–65.
Campbell, Rosie, Joni Lovenduski, and Sarah Childs. 2010. ‘Do women need women
representatives?’ British Journal of Political Science 40 (1): 171–94.
Cantril, Hadley. 1943. ‘Identification with social and economic class’. The Journal of
Abnormal Psychology 38 (1): 74–80.
Carey, Sean, and Jonathan Burton. 2004. ‘Research note: The influence of the press in
shaping public opinion towards the European Union in Britain’. Political Studies
52 (3): 623–40.
Carnes, Nicholas. 2012. ‘Does the numerical underrepresentation of the working class
in congress matter?’ Legislative Studies Quarterly 37 (1): 5–34.
Castles, Frances G., and Peter Mair. 1984. ‘Left-right political scales: Some “expert”
judgements’. European Journal of Political Research 12 (1): 147–57.
Centers, Richard. 1949. The Psychology of Social Classes. A Study of Class Consciousness.
Princeton, NJ: Princeton University Press.
Chan, Tak W., and John H. Goldthorpe. 2007. ‘Class and status: The conceptual
distinction and its empirical relevance’. American Sociological Review 72 (4): 512–32.
Cheng, Helen, John Brynner, Richard Wiggins, and Ingrid Schoon. 2012. ‘The meas-
urement and evaluation of social attitudes in two British cohort studies’. Social
Indicators Research 107 (2): 351–71.
Childs, Sarah, and Paul Webb. 2010. ‘Constituting and substantively representing
women: Applying new approaches to a UK case study’. Politics and Gender 6 (2):
199–223.
Clark, Alistair. 2012. Political Parties in the UK. Basingstoke: Palgrave Macmillan.
Clark, Terry N. 2001a. ‘The debate over “are social classes dying?” ’. In Terry N. Clark
and Seymour M. Lipset (eds.), The Breakdown of Class Politics. A Debate on Post-
Industrial Stratification. Baltimore, MD: The Johns Hopkins University Press,
273–319.
Clark, Terry N. 2001b. ‘What have we learned in a decade on class and party politics?’
In Terry N. Clark, and Seymour M. Lipset (eds.), The Breakdown of Class Politics: A Debate
on Post-Industrial Stratification. Baltimore, MD: The Johns Hopkins University Press, 6–39.
Clark, Terry N., and Seymour M. Lipset. 1991. ‘Are social classes dying?’ International
Sociology 6 (4): 397–410.
Clarke, Harold, David Sanders, Marianne Stewart, and Paul Whiteley. 2004. Political
Choice in Britain. Oxford: Oxford University Press.
Cockerell, Michael. 1988. Live from Number 10: The Inside Story of Prime Ministers and
Television. London: Faber & Faber.
Coleman, Richard P., and Lee Rainwater. 1979. Social Standing in America: New Dimen-
sions of Class. London: Routledge & Kegan Paul.
Conlon, Gavan, and Pietro Patrignani. 2011. ‘The returns to higher education qualifi-
cations’. BIS Research Paper Number 45.
Connelly, Roxanne, Vernon Gayle, and Paul S. Lambert, 2016. ‘A review of occupation
based social classifications for social survey research.’ Methodological Innovations 9: 1–14.

213
References

Converse, Phillip E. 1958. ‘The shifting role of class in political attitudes and behav-
iour’. In Eleanor E. Maccoby, Theodore M. Newcomb, and Eugene L. Hartley (eds.),
Readings in Social Psychology, 3rd edn. New York: Holt, 388–99.
Cook, Elizabeth. 1994. ‘Voter responses to women senate candidates’. In Elizabeth
Cook, Sue Thomas, and Clyde Wilcox (eds.), The Year of the Woman: Myths and
Realities. Boulder, CO: Westview, 217–36.
Corneo, Giacomo, and Hans-Peter Gruner. 2002. ‘Individual preferences for political
redistribution’. Journal of Public Economics 83 (1): 83–107.
Cotta, Maurizio, and Henreich Best (eds.). 2007. Democratic Representation in Europe:
Diversity, Change, and Convergence. Oxford: Oxford University Press.
Cozzarelli, Catherine, Anna V. Wilkinson, and Michael J. Tagler. 2001. ‘Attitudes
toward the poor and attributions for poverty’. Journal of Social Issues 57 (2):
207–27.
Crawford, Claire. 2014. ‘Socio-economic differences in university outcomes in the UK:
Drop-out, degree completion and degree class.’ IFS Working Paper W14/31.
Crewe, Ivor. 1986. ‘On the death and resurrection of class voting: Some comments on
how Britain votes’. Political Studies 34 (4): 620–38.
Crewe, Ivor. 1987. ‘A new class of politics?’ The Guardian, 15 June 1987.
Crewe, Ivor, Tony Fox, and James Alt. 1977a. ‘Non-voting in British general elections,
1966–October 1974’. British Political Sociology Yearbook 3: 38–109.
Crewe, Ivor, Bo Särlvik, and James Alt. 1977b. ‘Partisan dealignment in Britain
1964–1974’. British Journal of Political Science 7 (2): 129–90.
Curtice, John. 2016. ‘Brexit: Behind the referendum’. Political Insight, September
2016: 4–7.
Curtice, John. 1997. ‘Is the sun shining on Tony Blair? The electoral influence of British
newspapers’. The Harvard International Journal of Press/Politics 2 (2): 9–26.
Curtice, John, and Ann Mair. 2008. ‘Where have all the readers gone? Popular news-
papers and Britain’s political health’. In Alison Park, John Curtice, Katrina Thomson,
Miranda Phillips, Mark Johnson, and Elizabeth Clery (eds.), British Social Attitudes:
The 24th Report. London: Sage, 161–72.
Curtice, John, and Holli Semetko. 1994. ‘Does it matter what the papers say?’
In Anthony Heath, Roger Jowell, and John Curtice (eds.), Labour’s Last Chance?: The
1992 Election and Beyond. Aldershot: Dartmouth, 43–64.
Curtis, Josh. 2016. ‘Social mobility and class identity: The role of economic conditions
in 33 societies, 1999–2009’. European Sociological Review 32 (1): 108–21.
Cutler, Fred. 2002. ‘The simplest shortcut of all: Sociodemographic characteristics and
electoral choice’. Journal of Politics 64 (2): 466–90.
Cutts, David, Robert Ford, and Matthew J. Goodwin. 2011. ‘Anti-immigrant, politically
disaffected or still racist after all? Examining the attitudinal drivers of extreme right
support in Britain in the 2009 European elections’. European Journal of Political
Research 50 (3): 418–40.
Dalton, Russell. 2008. Citizen Politics: Public Opinion and Political Parties in Advanced
Industrial Democracies, 5th edn. Chatham, NJ: Chatham House Publishers.
Davis, James A. 1975. ‘Communism, conformity and categories: American tolerance in
1954 and 1972–73’. American Journal of Sociology 81 (3): 491–514.

214
 References

De Graaf, Nan Dirk, Anthony Heath, and Ariana Need. 2001. ‘Declining cleavages and
political choices: The interplay of social and political factors in the Netherlands.’
Electoral Studies 20 (1): 1–15.
De Graaf, Nan Dirk, Paul Nieuwbeerta, and Anthony Heath. 1995. ‘Class mobility and
political preference: Individual and contextual effects.’ American Journal of Sociology
100 (1): 997–1027.
De Lange, Sarah L. 2007. ‘A new winning formula? The programmatic appeal of the
radical right’. Party Politics 13 (4): 411–35.
Dekker, Paul, and Peter Ester. 1987. ‘Working-class authoritarianism: A re-examination
of the Lipset thesis’. European Journal of Political Research 15 (4): 395–415.
Dennis, Norman, Fernando Henriques, and Clifford Slaughter. 1956. Coal is Our Life.
London: Eyre and Spottiswoode.
Denny, Kevin, and Orla Doyle. 2009. ‘Does voting history matter? Analysing persist-
ence in turnout’. American Journal of Political Science 53 (1): 17–35.
Denver, David. 1995. ‘Non-voting in Britain’. In Joan Font, and Rosa Virós (eds.),
Electoral Abstention in Europe. Barcelona: Institut de Ciènces Politiques i Socials.
Dickens, Charles. 1837, 1999. The Pickwick Papers. Mark Wormald (ed.), New York:
Penguin.
Dolan, Kathleen. 1998. ‘Voting for women in the “year of the woman” ’. American
Journal of Political Science 42 (1): 272–93.
Dorling, Danny. 2016. ‘Brexit: The decision of a divided country’. British Medical Journal
354: i3697.
Dowds, Lizanne, and Ken Young. 1996. ‘National Identity’. In Roger Jowell, John
Curtice, Alison Park, Lindsay Brook, and Katrina Thomson (eds.), British Social Atti-
tudes: The 13th Report. Aldershot: Dartmouth.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper and Row.
Drever, Frances, Tim Doran, and Margaret Whitehead. 2004. ‘Exploring the relation
between class, gender, and self-rated general health using the new socioeconomic
classification. A study using data from the 2001 census’. Journal of Epidemiology and
Community Health. 58 (7): 590–6.
Dunleavy, Patrick. 1980. ‘The political implications of sectoral cleavages and the
growth of state employment: Part 1, the analysis of production cleavages’. Political
Studies 28 (3): 364–83.
Dunleavy, Patrick. 1986. ‘The growth of sectoral cleavages and the stabilization of state
expenditures’. Environment and Planning D: Society and Space 4 (2): 129–44.
Dunleavy, Patrick. 1987. ‘Class dealignment in Britain revisited’. West European Politics
10 (3): 400–19.
Dunleavy, Patrick, and Christopher T. Husbands. 1985. British Democracy at the Cross-
roads. London: Allen and Unwin.
Eder, Klaus. 1993. The New Politics of Class. London: Sage.
Eidlin, Barry. 2014. ‘Class formation and class identity: Birth, death, and possibilities
for renewal’. Sociological Compass 8 (8): 1045–62.
Elff, Martin. 2007. ‘Social structure and electoral behavior in comparative perspective:
The decline of social cleavages in Western Europe revisited’. Perspectives on Politics
5 (2): 277–94.

215
References

Elff, Martin. 2009. ‘Social divisions, party positions, and electoral behaviour’. Electoral
Studies 28 (2): 297–308.
Erikson, Robert, and John H. Goldthorpe. 1992. The Constant Flux: A Study of Class
Mobility in Industrial Societies. Oxford: Clarendon Press.
Erikson, Robert, and John H. Goldthorpe. 2010. ‘Has social mobility in Britain
decreased? Reconciling divergent findings on income and class mobility’. The British
Journal of Sociology 61 (2): 211–30.
Erikson, Robert, John H. Goldthorpe, Michelle Jackson, Mair Yaish, and David Cox.
2005. ‘On class differentials in educational attainment’. Proceedings of the National
Academy of Sciences of the United States of America 102 (27): 9730–3.
Ermisch, John, and Cheti Nicoletti. 2007. ‘Intergenerational earnings mobility: Changes
across cohorts in Britain’. B.E. Journal of Economic Analysis and Policy 7 (2): 157–83.
Evans, Geoffrey. 1992. ‘Testing the validity of the Goldthorpe class schema’. European
Sociological Review 8 (4): 211–32.
Evans, Geoffrey. 1993a. ‘The decline of class divisions in Britain? Class and ideological
preferences in the 1960s and 1980s’. British Journal of Sociology 44 (3): 449–71.
Evans, Geoffrey. 1993b. ‘Cognitive models of class structure and explanations of social
outcomes’. European Journal of Social Psychology 23 (5): 445–64.
Evans, Geoffrey. 1993c. ‘Class conflict and inequality’. In Roger Jowell, Lindsay Brook,
Lizanne Dowds, and Daphne Ahrendt (eds.), International Social Attitudes: The 10th
BSA Report. Aldershot: Dartmouth.
Evans, Geoffrey. 1996. ‘Putting men and women into classes: An assessment of the
cross-sex validity of the Goldthorpe class schema’. Sociology 30 (2): 209–34.
Evans, Geoffrey. 1997. ‘Political ideology and popular beliefs about class and oppor-
tunity: Evidence from a survey experiment’. British Journal of Sociology 48 (3):
450–70.
Evans, Geoffrey (ed.). 1999. The End of Class Politics? Class Voting in Comparative
Context. Oxford: Oxford University Press.
Evans, Geoffrey, and Nan Dirk de Graaf (eds.). 2013. Political Choice Matters: Explaining
the Strength of Class and Religious Cleavages in Cross-National Perspective. Oxford:
Oxford University Press.
Evans, Geoffrey, and Anthony Heath. 1995. ‘The measurement of left-right and
libertarian-authoritarian values: A comparison of balanced and unbalanced scales’.
Quality and Quantity 29 (2): 191–206.
Evans, Geoffrey, Anthony Heath, and Mansur Lalljee. 1996. ‘Measuring left-right and
libertarian-authoritarian values in the British electorate’. British Journal of Sociology
47 (1): 93–112.
Evans, Geoffrey, Anthony Heath, and Clive Payne. 1999. ‘Class: Labour as a catch-all
party?’ In Geoffrey Evans and Pippa Norris (eds.), Critical Elections: British Parties and
Voters in Long-term Perspective. London: Sage, 87–101.
Evans, Geoffrey, and Jonathan Mellon. 2016a. ‘Working class votes and Conservative
losses: Solving the UKIP puzzle’. Parliamentary Affairs 69 (2): 464–79.
Evans, Geoffrey, and Jonathan Mellon. 2016b. ‘Identity, awareness and political
attitudes: Why are we still working class?’ British Social Attitudes: The 33rd Report.
London: Sage.

216
 References

Evans, Geoffrey, and Colin Mills. 1998. ‘Identifying class structure: A latent class
analysis of the criterion-related and construct validity of the Goldthorpe class
schema’. European Sociological Review 14 (1): 87–106.
Evans, Geoffrey, and Colin Mills. 2000. ‘In search of the wage-labour/service contract:
New evidence on the validity of the Goldthorpe class schema’. British Journal of
Sociology 51 (4): 641–61.
Evans, Geoffrey, and James Tilley. 2011. ‘Private schools and public divisions: The
influence of fee-paying education on social attitudes’. British Social Attitudes: The
28th Report. London: Sage, 37–52.
Evans, Geoffrey, and James Tilley. 2012a. ‘How parties shape class politics: Explaining
the decline of class party support’. British Journal of Political Science 42 (1): 137–61.
Evans, Geoffrey, and James Tilley. 2012b. ‘The depoliticization of inequality and redis-
tribution: Explaining the decline of class voting’. Journal of Politics 74 (4): 963–76.
Evans, Geoffrey, and James Tilley. 2013. ‘Ideological convergence and the decline of
class voting in Britain’. In Geoffrey Evans and Nan Dirk de Graaf (eds.), Political
Choice Matters: Explaining the Strength of Class and Religious Cleavages in Cross-National
Perspective. Oxford: Oxford University Press, 87–113.
Evans, Geoffrey, and Stephen Whitefield. 2006. ‘Explaining the rise and persistence of
class voting in postcommunist Russia, 1993–2001’. Political Research Quarterly 59: 23–34.
Evans, Mariah D. R., and Jonathan Kelley. 2004. ‘Subjective social location: Data from
21 nations’. International Journal of Public Opinion Research 16 (1): 3–38.
Feagin, Joseph. R. 1975. Subordinating the Poor: Welfare and American Beliefs. Englewood
Cliffs, NJ: Prentice-Hall.
Felstead, Alan, Duncan Gallie, and Francis Green (eds.). 2015. Unequal Britain at Work.
Oxford: Oxford University Press.
Fieldhouse, Edward, and Christopher Prosser. 2016. ‘When attitudes and behaviour
collide: How the Scottish independence referendum cost Labour’. Working paper.
Fielding, Steven. 2003. The Labour Party: Continuity and Change in the Making of ‘New’
Labour. Basingstoke: Palgrave Macmillan.
Finkel, Steven E. 1993. ‘Re-examining the “minimal effects” model in recent presiden-
tial campaigns’. The Journal of Politics 55 (1): 1–21.
Finlayson, Alan, and Judi Atkins. 2015. ‘British Political Speech Archive’. Available
online: http://www.britishpoliticalspeech.org/speech-archive.htm.
Fischer, Claude S., and Michael Hout. 2006. Century of Difference: How America Changed
in the Last One Hundred Years. New York: Russell Sage Foundation.
Flanagan, Scott C. 1987. ‘Value change in industrial societies’. American Political Science
Review 81 (4): 1289–319.
Fleishman, John. 1988. ‘Attitude organization in the general public: Evidence for a bi-
dimensional structure’. Social Forces 67 (1): 159–84.
Ford, Robert. 2016. ‘Older ‘left-behind’ voters turned against a political class with
values opposed to theirs’. The Guardian, 25 June 2016.
Ford, Robert, and Matthew J. Goodwin. 2010. ‘Angry white men: Individual and context-
ual predictors of support for the British National Party’. Political Studies 58 (1): 1–25.
Ford, Robert, and Matthew J. Goodwin. 2014. Revolt on the Right: Explaining Support for
the Radical Right in Britain. London: Routledge.

217
References

Ford, Robert, Matthew J. Goodwin, and David Cutts. 2012. ‘Strategic eurosceptics and
polite xenophobes: Support for the United Kingdom Independence Party (UKIP) in the
2009 European Parliament elections’. European Journal of Political Research 51 (2): 204–34.
Franklin, Mark. 1992. ‘The decline of cleavage politics’. In Mark Franklin, Thomas
Mackie, Henry Valen, et al. (eds.), Electoral Change: Responses to Evolving Social and
Attitudinal Structures in Western Countries. Cambridge: Cambridge University Press,
383–405.
Franklin, Mark, Thomas Mackie, Henri Valen, et al. (eds.). 1992. Electoral Change:
Responses to Evolving Social and Attitudinal Structures in Western Countries. Cambridge:
Cambridge University Press.
Franklin, Mark N. 1984. ‘How the decline in class voting opened the way to radical
change to British politics’. British Journal of Political Science 14 (4): 483–508.
Franklin, Mark N. 1985. The Decline of Class Voting in Britain: Changes in the Basis of
Electoral Choice, 1964 –1983. Oxford: Clarendon Press.
Franklin, Mark N. 2004. Voter Turnout and the Dynamics of Electoral Competition in
Established Democracies since 1945. Cambridge: Cambridge University Press.
Fraser, Ronald. 1968/9. Work (2 vols.). Harmondsworth, UK: Penguin Books.
Furnham, Adrian. 1982. ‘Why are the poor always with us? Explanations for poverty in
Britain’. British Journal of Social Psychology 21 (4): 311–22.
Furnham, Adrian. 1988. Lay Theories: Everyday Understanding of Problems in the Social
Science. Oxford: Pergamon.
Gabel, Matthew. 1998a. ‘Public support for European integration: An empirical test of
five theories’. Journal of Politics 60 (2): 333–54.
Gabel, Matthew. 1998b. Interests and Integration: Market Liberalization, Public Opinion
and the European Union. Ann Arbor: University of Michigan Press.
Gabel, Matthew, and Harvey D. Palmer. 1995. ‘Understanding variation in public
support for European Integration’. European Journal of Political Research 27 (1): 3–19.
Gallie, Duncan. 2015. ‘Class inequality at work: Trends to polarization?’ In Alan
Felstead, Duncan Gallie, and Francis Green (eds.), Unequal Britain at Work. Oxford:
Oxford University Press, 22–41.
Garry, John, and James Tilley. 2009. ‘The macro economic factors conditioning the
impact of identity on attitudes towards the EU’. European Union Politics 10 (3): 361–79.
Gavin, Neil T. 1996. ‘Class voting and the Labour Party in Britain: The analysis of
qualitative data on voting preference in the 1987 General Election’. Electoral Studies
15 (3): 311–26.
Gelman, Andrew, and Gary King. 1993. ‘Why are American presidential election cam-
paign polls so variable when votes are so predictable?’ British Journal of Political Science
23 (4): 409–51.
Gerber, Alan, Donald Green, and Ron Shachar. 2003. ‘Voting may be habit forming:
Evidence from a randomized field experiment’. American Journal of Political Science
47 (3): 540–50.
Glenn, Norval D., and Michael Grimes. 1968. ‘Aging, voting, and political interest’.
American Sociological Review 33 (4): 563–75.
Golder, Matt. 2003. ‘Explaining variation in the success of extreme right parties in
Western Europe’. Comparative Political Studies 36 (4): 432–66.

218
 References

Goldstein, Kenneth, and Travis N. Ridout. 2004. ‘Measuring the effects of televised
political advertising in the United States’. Annual Review of Political Science 7: 205–26.
Goldthorpe, John H. 1995. ‘The service class revisited’. In Tim Butler and Michael
Savage (eds.). Social Change and the Middle Classes. London: UCL Press, 313–29.
Goldthorpe, John H. 2004. The Economic Basis of Social Class. LSE STICERD Research
Paper No. CASE080. London: London School of Economics and Political Science.
Goldthorpe, John H. 2007. ‘Social class and the differentiation of employment con-
tracts’. In On Sociology, vol. 2. Stanford: Stanford University Press, 101–24.
Goldthorpe, John H. 2016. ‘Social class mobility in modern Britain: Changing struc-
ture, constant process’. Centre for Social Investigation 21: briefing note. Available
online: http://csi.nuff.ox.ac.uk/?page_id=11.
Goldthorpe, John H., David Lockwood, Frank Bechhofer, and Jennifer Platt. 1969. The
Affluent Worker in the Class Structure. Cambridge: Cambridge University Press.
Goldthorpe, John H., and Abigail McKnight. 2006. ‘The economic basis of social class’.
In Stephen L. Morgan, David B. Grusky, and Gary S. Fields (eds.). Mobility and Inequal-
ity: Frontiers of Research in Sociology and Economics. Stanford: Stanford University Press.
Goldthorpe, John H., and Colin Mills. 2008. ‘Trends in intergenerational class mobility
in modern Britain: Evidence from national surveys, 1972–2005’. National Institute
Economic Review 205 (1): 83–100.
Goodhart, David. 2013. The British Dream: Successes and Failures of Post-War Immigra-
tion. London: Atlantic Books.
Goodwin, Matthew, and Oliver Heath. 2016. ‘The 2016 referendum, Brexit and the left
behind: An aggregate-level analysis of the result’. The Political Quarterly 87 (3): 323–32.
Goodwin, Matthew, and Caitlin Milazzo. 2015. UKIP: Inside the Campaign to Redraw the
Map of British Politics. Oxford: Oxford University Press.
Gorecki, Maciej A. 2013. ‘Electoral context, habit-formation and voter turnout: A new
analysis’. Electoral Studies 32 (1): 140–52.
Green, Donald P., and Ron Shachar. 2000. ‘Habit formation and political behaviour:
Evidence of consuetude in voter turnout’. British Journal of Political Science 30 (4):
561–73.
Grunberg, Gerard, and Etienne Schweisguth. 1993. ‘Social libertarianism and economic
liberalism’. In Daniel Boy and Nonna Mayer (eds.), The French Voter Decides. Ann
Arbor: University of Michigan Press, 45–64.
Güveli, Ayse, Ariana Need, and Nan Dirk De Graaf. 2007a. ‘Socio-political, cultural and
economic preferences and behaviour of the social and cultural specialists versus the
technocrats. Social class or education?’ Social Indicators Research 81 (3): 597–631.
Güveli, Ayse, Ariana Need, and Nan Dirk De Graaf. 2007b. ‘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): 129–46.
Hamlin, Alan, and Colin Jennings. 2011. ‘Expressive political behaviour: Foundations,
scope and implications’. British Journal of Political Science 41 (3): 645–70.
Harrison, Eric, and David Rose. 2006. The European Socio-economic Classification User Guide.
Colchester, UK: Institute for Social and Economic Research (ISER), University of Essex.
Harrop, Martin, Judith England, and Christopher T. Husbands. 1980. ‘The bases of
National Front support’. Political Studies 28 (2): 271–83.

219
References

Hayes, Bernadette. 1995. ‘The impact of class on political attitudes: A comparative

study of Great Britain, West Germany, Australia and the United States’. European
Journal of Political Research 27 (1): 69–91.
Heath, Anthony, John Curtice, and Gabriella Elgenius. 2009. ‘Individualization and the
decline of class identity’. In Margaret Wetherell (ed.), Identity in the 21st Century: New
Trends in Changing Times. Basingstoke: Palgrave, 21–40.
Heath, Anthony, Geoffrey Evans, and Jean Martin. 1994. ‘The measurement of core
beliefs and values: The development of balanced socialist/laissez-faire and libertarian/
authoritarian scales’. British Journal of Political Science 24 (1): 115–32.
Heath, Anthony, Roger Jowell, and John Curtice. 1985. How Britain Votes. Oxford:
Pergamon Press.
Heath, Anthony, Roger Jowell, and John Curtice. 1987. ‘Trendless fluctuation: A reply
to Crewe.’ Political Studies 35 (2): 259–77.
Heath, Anthony, Roger Jowell, John Curtice, Geoffrey Evans, Julia Field, and Sharon
Witherspoon. 1991. Understanding Political Change: The British Voter 1964–1987.
Oxford: Pergamon.
Heath, Anthony, and Bridget Taylor. 1999. ‘New sources of abstention?’ In Geoffrey
Evans and Pippa Norris (eds.), Critical Elections: British Parties and Voters in Long-Term
Perspective. London: Sage, 164–79.
Heath, Anthony F., Nan Dirk De Graaf, and Yu Li. 2010. ‘How fair is the route to the
top? Perceptions of social mobility’. British Social Attitudes: The 27th Report: Exploring
Labour’s Legacy. London: Sage.
Heath, Anthony F., Roger M. Jowell, and John K. Curtice. 2001. The Rise of New Labour: Party
Policies and Voter Choices: Party Policies and Voter Choices. Oxford: Oxford University Press.
Heath, Oliver. 2007. ‘Explaining turnout decline in Britain, 1964–2005: Party identifi-
cation and the political context’. Political Behavior 29 (4): 493–516.
Heath, Oliver. 2015. ‘Policy representation, social representation and class voting in
Britain’. British Journal of Political Science 45 (1): 173–93.
Heath, Oliver. 2016. ‘Policy alienation, social alienation and working class abstention
in Britain, 1964–2010’. British Journal of Political Science. Available online: http://dx.
doi.org/10.1017/S0007123416000272.
Hill, Kim Q., and Jan E. Leighley. 1996. ‘Political parties and class mobilization in con-
temporary United States elections’. American Journal of Political Science 40 (3): 787–804.
Hobolt, Sara B. 2014. ‘Ever closer or ever wider? Public attitudes towards further
enlargement and integration in the European Union’. Journal of European Public Policy
21 (5): 664–80.
Hodge, Robert W., and Donald J. Treiman. 1968. ‘Class identification in the United
States’. American Journal of Sociology 73 (5): 535–47.
Hoggart, Richard. 1957. The Uses of Literacy. London: Chatto and Windus.
Hoggart, Richard. 1989. ‘Foreword’. In George Orwell, Road to Wigan Pier. New York:
Penguin Random House.
Hooghe, Liesbet, Ryan Bakker, Anna Brigevich, Catherine De Vries, Erica Edwards, Gary
Marks, Jan Rovny, Marco Steenbergen, and Milada Vachudova. 2010. ‘Reliability and
validity of the 2002 and 2006 Chapel Hill expert surveys on party positioning’.
European Journal of Political Research 49 (5): 687–703.

220
 References

Hooghe, Liesbet, and Gary Marks. 2004. ‘Does identity or economic rationality drive
public opinion on European integration?’ PS: Political Science and Politics 37 (3): 415–20.
Hooghe, Liesbet, and Gary Marks. 2005. ‘Calculation, community, and cues: Public
opinion on European integration’. European Union Politics 6 (4): 419–43.
House of Commons Information Office. 2009. Ministerial Salaries. Factsheet M6, Mem-
bers Services. Appendix A.
Hout, Michael. 2008. ‘How class works: Objective and subjective aspects of class since
the 1970s’. In Annette Lareau, and Dalton Conley (eds.), Social Class: How Does it
Work? New York: Russell Sage Foundation, 25–64.
Hout, Michael, Clem Brooks, and Jeff Manza. 1993. ‘The persistence of classes in post-
industrial societies’. International Sociology 8 (3): 259–77.
Hout, Michael, Clem Brooks, and Jeff Manza. 1995. ‘The democratic class struggle in
U.S. presidential elections, 1948–1992’. American Sociological Review 60 (6): 805–28.
Houtman, Dick. 2003. ‘Lipset and “working-class” authoritarianism’. The American
Sociologist 34 (1): 85–103.
Houtman, Dick, Peter Achterberg, and Anton Derks. 2008. Farewell to the Leftist Working
Class. New Brunswick: Transaction Publishers.
Huber, Joan, and William H. Form. 1973. Income and Ideology. New York: The Free Press.
Huddy, Leonie. 1994. ‘The political significance of voters’ gender stereotypes’.
In Michael X.D. Carpini, Leonie Huddy, and Robert Y. Shapiro (eds.), Research in
Micropolitics: New Directions in Political Sociology. Greenwich, CT: JAI Press, 169–93.
Huddy, Leonie, and Nayda Terkildsen. 1993. ‘Gender stereotypes and the perception of
male and female candidates’. American Journal of Political Science 37 (1): 119–47.
Ichheiser, Gustav. 1949. ‘Misunderstandings in human relations: A study in false social
perception’. American Journal of Sociology 55 (2): 1–70.
Inglehart, Ronald. 1997. Modernization and Postmodernization: Cultural, Economic and
Political Change in 43 Societies. Princeton, NJ: Princeton University Press.
Inglehart, Ronald, and Jacques-René Rabier. 1986. ‘Political realignment in advanced
industrial society: From class-based politics to quality-of-life politics’. Government and
Opposition 21 (4): 457–79.
Irwin, Sarah. 2015. ‘Class and comparison: Subjective social location and lay experi-
ences of constraint and mobility’. British Journal of Sociology 66 (2): 259–81.
Ivarsflaten, Elisabeth. 2005. ‘The vulnerable populist right parties: No economic
realignment fueling their electoral success’. European Journal of Political Research
44 (3): 465–92.
Ivarsflaten, Elisabeth. 2006. ‘Reputational shields: Why most anti-immigrant parties
failed in Western Europe, 1980–2005’. Annual Meeting of the American Political Science
Association, Philadelphia.
Ivarsflaten, Elisabeth. 2008. ‘What unites right-wing populists in Western Europe? Re-
examining grievance mobilization models in seven successful cases’. Comparative
Political Studies 41 (1): 3–23.
Iversen, Torben, and David Soskice. 2001. ‘An asset theory of social policy preferences’.
American Political Science Review 95 (4): 875–93.
Iyengar, Shanto, and Donald Kinder. 1987. News That Matters: Television and Public
Opinion. Chicago: University of Chicago Press.

221
References

Iyengar, Shanto, and Adam F. Simon. 2000. ‘New perspectives and evidence on political
communication and campaign effects.’ Annual Review of Psychology 51: 149–69.
Jackman, Mary R., and Robert W. Jackman. 1983. Class Awareness in the United States.
Berkeley: University of California Press.
Jackson, Ben. 2016. ‘Commentary: Hard Labour.’ The Political Quarterly 87 (1): 4–5.
Jackson, Brian, and Dennis Marsden. 1966. Education and the Working Class. Harmonds-
worth, UK: Penguin Books.
Jacobsen, Dag I. 2001. ‘Higher education as an arena for political socialization: Myth or
reality?’ Scandinavian Political Studies 24 (4): 351–68.
Janssen, Giedo, Geoffrey Evans, and Nan Dirk De Graaf. 2013. ‘Class voting and left-
right party positions: A comparative study of 15 western democracies, 1960–2005’.
Social Science Research 42: 376–400.
Johns, Robert, David Denver, James Mitchell, and Charles Pattie. 2010. Voting for a
Scottish Government: The Scottish Parliament Elections of 2007. Manchester: Manchester
University Press.
Johnson, Brian, and Alaa Al-Hamad. 2011. ‘Trends in socio-economic inequalities in
female mortality, 2001–08. Intercensal estimates for England and Wales’. Health
Statistics Quarterly 52 (1): 3–32.
Johnston, Ron, and Charles Pattie. 2003. ‘The growing problem of electoral turnout in
Britain? Voters and non-voters at the British 2001 General Election’. Representation
40 (1): 30–43.
Jones, Owen. 2011. Chavs: The Demonization of the Working Class. London: Verso
Books.
Kalmijn, Matthijs, and Gerbert Kraaykamp. 2007. ‘Social stratification and attitudes:
A comparative analysis of the effects of class and education in Europe’. British Journal
of Sociology 58 (4): 547–76.
Kantomaa, Marko T., Tuija H. Tammelin, Panayotes Demakakos, Hannah E. Ebeling,
and Anja M. Taanila. 2010. ‘Physical activity, emotional and behavioural problems,
maternal education and self-reported educational performance of adolescents’.
Health Education Research 25 (2): 368–79.
Katz, Michael B. 1989. The Undeserving Poor: From the War on Poverty to the War on
Welfare. New York: Pantheon Books.
Kavanagh, Dennis, and Peter Morris. 1994. Consensus Politics from Attlee to Major.
Oxford: Blackwell.
Kellas, James G. 1968. Modern Scotland: The Nation since 1870. New York: Praeger.
Kelley, Jonathan, and Maria D. R. Evans. 1993. ‘The Legitimation of Inequality: Occu-
pational earnings in nine nations’. American Journal of Sociology, 99 (1): 75–125.
Kelley, Jonathan, and Maria D. R. Evans. 1995. ‘Class and class conflict in six Western
nations’. American Sociological Review 60 (2): 157–78.
Key, Valdimer O. 1966. The Responsible Electorate: Rationality in Presidential Voting,
1936–1960. Cambridge, MA: Belknap Press.
King, Anthony. 1997. ‘Why Labour won – at last’. In Anthony King (ed.), New Labour
Triumphs: Britain at the Polls. Chatham, NJ: Chatham House Publishers.
Kirby, Simon, and Rebecca Riley. 2008. ‘The external returns to education: UK evidence
using repeated cross-sections’. Labour Economics 15 (4): 619–30.

222
 References

Kitschelt, Herbert. 1994. The Transformation of European Social Democracy. Cambridge:

Cambridge University Press.
Kitschelt, Herbert. 1995. The Radical Right in Western Europe: A Comparative Analysis.
Ann Arbor: University of Michigan Press.
Kitschelt, Herbert, and Anthony J. McGann. 2005. ‘The radical right in the Alps: Evolu-
tion of support for the Swiss SVP and Austrian FPO’. Party Politics 11 (2): 147–71.
Klapper, Joseph T. 1960. The Effects of Mass Communication. Glencoe, IL Illinois: The Free
Press.
Klingemann, Hans-Dieter, Andrea Volkens, Judith Bara, Ian Budge, and Michael
D. McDonald. 2006. Mapping Policy Preferences II: Estimates for Parties, Governments
and Electors in the OECD, EU and Central and Eastern Europe 1990–2003. Oxford:
Oxford University Press.
Klingeren, Marijn V., Hajo G. Boomgaarden, and Claes H. De Vreese. 2013. ‘Going
soft or staying soft: Have identity factors become more important than economic
rationale when explaining Euroscepticism?’ Journal of European Integration 35 (6):
689–704.
Kluegel, James R., and Eliot R. Smith. 1986. Beliefs about Inequality: Americans’ Views of
What Is and What Ought To Be. New York: Aldine de Gruyter.
Kriesi, Hanspeter. 1989. ‘New social movements and the new class in the Netherlands’.
American Journal of Sociology 94 (5): 1078–116.
Ladd, Jonathan M., and Gabriel S. Lenz. 2009. ‘Exploiting a rare communication shift
to document the persuasive power of the news media’. American Journal of Political
Science 53 (2): 394–410.
Lancee, Bram, and Oriane Sarrasin. 2015. ‘Educated preferences or selection effects?
A longitudinal analysis of the impact of educational attainment on attitudes towards
immigrants’. European Sociological Review 31 (4): 490–501.
Langford, Ann, and Brian Johnson. 2010. ‘Trends in social inequalities in male mortal-
ity, 2001–08. Intercensal estimates for England and Wales’. Health Statistics Quarterly
47: 1–28.
Laver, Michael, and Ian Budge (eds.). 1992. Party Policy and Government Coalitions.
London: Macmillan.
Laver, Michael, and John Garry. 2000. ‘Estimating policy positions from political texts’.
American Journal of Political Science 44 (3): 619–34.
Laver, Michael, and W. Ben Hunt. 1992. Policy and Party Competition. New York and
London: Routledge.
Lawler, S. (ed). 2005. Class, Culture and Identity. Special Issue, Sociology 39 (5).
Lazarsfeld, Paul F., Bernard Berelson, and Hazel Gaudet. 1948. The People’s Choice: How
the Voter Makes Up His Mind in a Presidential Campaign. New York: University of
Columbia Press.
Leighley, Jan E., and Jonathan Nagler. 1992a. ‘Socioeconomic class bias in turnout,
1964–1988: The voters remain the same’. American Political Science Review 86 (3): 725–36.
Leighley, Jan E., and Jonathan Nagler. 1992b. ‘Individual and systemic influences on
turnout: Who votes? 1984’. The Journal of Politics 54 (3): 718–40.
Leighley, Jan E., and Jonathan Nagler. 2014. Who Votes Now?: Demographics, Issues,
Inequality, and Turnout in the United States. Princeton: Princeton University Press.

223
References

Lemieux, Anthony F., and Felicia Pratto. 2003. ‘Poverty and Prejudice’. In Stuart C. Carr
and Todd S. Sloan (eds.), Poverty and Psychology: From Global Perspective to Local. New
York: Kluwer Academic/Plenum.
Linos, Katerina, and Martin West. 2002. ‘Self-interest, social beliefs, and attitudes to
redistribution. Re-addressing the issue of cross-national variation’. European Socio-
logical Review 19 (4): 393–409.
Lipset, Seymour M. 1959. Political Man: The Social Bases of Politics. Garden City, NY:
Doubleday.
Lipset, Seymour M. 1981. Political Man. 2nd edn. Garden City, NY: Doubleday.
Lipset, Seymour M. 1991. ‘No third way: A comparative perspective on the left’.
In Daniel Chirot (ed.), The Crisis of Leninism and the Decline of the Left. Seattle:
University of Washington Press, 183–232.
Lottes, Ilsa, and Peter Kuriloff. 1994. ‘The impact of college experience on political and
social attitudes’. Sex Roles 31 (1): 31–54.
Lowe, Will, Kenneth Benoit, Slava Mikhaylov, and Michael Laver. 2011. ‘Scaling policy
preferences from coded political texts’. Legislative Studies Quarterly 36 (1): 123–55.
Lucassen, Geertje, and Marcel Lubbers. 2012. ‘Who fears what? Explaining far-right-
wing preference in Europe by distinguishing perceived cultural and economic ethnic
threats’. Comparative Political Studies 45 (5): 547–74.
Lynch, Peter. 2002. SNP: The History of the Scottish National Party. Cardiff: Welsh
Academic Press.
Marshall, Gordon, David Rose, Howard Newby, and Carolyn Vogler. 1988. Social Class
in Modern Britain. London: Hutchinson.
Marwick, Arthur. 1980. Class: Image and Reality. London: Collins.
McCutcheon, Allan. 1985. ‘A latent class analysis of tolerance for nonconformity in the
American public’. Public Opinion Quarterly 49 (4): 474–88.
McDonald, Michael D., and Ian Budge. 2014. ‘Getting it (approximately) right (and
centre and left)’. Electoral Studies 35 (1): 67–77.
McKenzie, Robert, and Allan Silver. 1968. Angels in Marble: Working Class Conservatives
in Urban England. Chicago: University of Chicago Press.
McLaren, Lauren. 2002. ‘Public support for the European Union: Cost/benefit analysis
or perceived cultural threat?’ Journal of Politics 64 (12): 551–66.
McLaren, Lauren. 2004. ‘Opposition to European integration and fear of loss of
national identity: Debunking a basic assumption regarding hostility to the integra-
tion project’. European Journal of Political Research 43 (6): 895–912.
McLaren, Lauren. 2006. Identity, Interests and Attitudes to European Integration. Basing-
stoke: Palgrave Macmillan.
McLaren, Lauren. 2007. ‘Explaining mass-level Euroscepticism: identity, interests, and
institutional distrust’. Acta Politica 42 (2): 233–51.
McLaren, Lauren. 2015. Immigration and Perceptions of National Political Systems in
Europe. Oxford: Oxford University Press.
McLaren, Lauren, and Mark Johnson. 2007. ‘Resources, group conflict and symbols:
Explaining anti-immigration hostility in Britain’. Political Studies 55 (4): 709–32.
McLean, Iain. 1970. ‘The rise and fall of the Scottish National Party’. Political Studies
18 (3): 357–72.

224
 References

McLeod, Jane D., and Karen Kaiser. 2004. ‘Childhood Emotional and Behavioral Prob-
lems and Educational Attainment’. American Sociological Review 69 (5): 636–58.
Meguid, Bonnie M. 2005. ‘Competition between unequals: The role of mainstream
party strategy in niche party success’. American Political Science Review 99 (3): 347–59.
Merton, Robert K. 1957. Social Theory and Social Structure. London: MacMillan.
Michaels, Walter B. 2006. The Trouble with Diversity: How We Learned to Love Identity and
Ignore Inequality. Basingstoke: Palgrave Macmillan.
Middendorp, Cees, Hans J. W. Luyten, and R. Dooms. 1993. ‘Issue voting in the
Netherlands: Two-dimensional issue-distances between own position and perceived
party position as determinants of the vote’. Acta Politica 28 (1): 39–59.
Mills, Colin. 2015a. ‘The Great British Class Survey: Requiescat in pace’. Sociological
Review 63 (2): 393–9.
Mills, Colin. 2015b. ‘Is class inequality at KS4 decreasing?’ Centre for Social Investigation
11: briefing note. Available online: http://csi.nuff.ox.ac.uk/?page_id=11.
Mitchell, James, Lynn Bennie, and Robert Johns. 2012. The Scottish National Party:
Transition to Power. Oxford: Oxford University Press.
Moene, Karl Ove, and Michael Wallerstein. 2001. ‘Inequality, social insurance, and
redistribution’. American Political Science Review 95 (4): 859–74.
Moller, Stephanie, Arthur S. Alderson, and Francois Nielsen. 2009. ‘Changing patterns
of income inequality in U.S. counties, 1970–2000’. American Journal of Sociology
114 (4): 1037–101.
Moorhouse, H. F. 1976. ‘Attitudes to class and class relationships in Britain’. Sociology
10 (3): 469–96.
Mount, Ferdinand. 2004. Mind the Gap: Class in Britain Now. London: Short Books.
Mudde, Cas. 2007. Populist Radical Right Parties in Europe. Cambridge: Cambridge Uni-
versity Press.
Napier, Jaime L., and John T. Jost. 2008. ‘The “anti-democratic personality” revisited:
A cross-national investigation of working class authoritarianism’. Journal of Social
Issues 64 (3): 595–617.
National Equality Panel. 2010. An Anatomy of Economic Inequality in the UK: Report of the
National Equality Panel. London: Government Equalities Office.
Neckerman, Kathryn M., and Florencia Torche. 2007. ‘Inequality: Causes and conse-
quences’. Annual Review of Sociology 33: 335–57.
Newman, Saul. 1992. ‘The rise and decline of the Scottish National Party: Ethnic
politics in a post-industrial environment’. Ethnic and Racial Studies 15 (1): 1–35.
Newton, Kenneth. 1991. ‘Do people read everything they believe in the papers? News-
papers and voters in the 1983 and 1987 elections’. British Elections & Parties Yearbook
1 (1): 49–74.
Newton, Kenneth. 2006. ‘May the weak force be with you: The power of the mass media
in modern politics’. European Journal of Political Research 45 (2): 209–34.
Newton, Kenneth, and Malcom Brynin. 2001. ‘The national press and party voting in
the UK’. Political Studies 49 (2): 262–85.
NHS Information Centre for Health and Social Care. 2009. ‘GP Earnings and Expenses
Enquiry 2006/7: Final Report’. Available online: http://digital.nhs.uk/catalogue/
PUB02373/gp-earn-expe-enqu-2006-07-fina-rep.pdf.

225
References

Nieuwbeerta, Paul. 1995. The Democratic Class Struggle in Twenty Countries, 1945–1990.
Amsterdam: Thesis Publishers.
Nieuwbeerta, Paul. 1996. ‘The democratic class struggle in post-war societies: Class
voting in twenty countries, 1945–1990’. Acta sociologica 39 (4): 345–83.
Nisbet, Robert, A. 1959. ‘The decline and fall of social class’. Pacific Sociological Review
2 (1): 11–17.
Nordlinger, Eric A. 1967. The Working-Class Tories: Authority, Deference and Stable Dem-
ocracy. London: MacGibbon & Kee.
Norris, Pippa. 2005. Radical Right: Voters and Parties in the Electoral Market. New York:
Cambridge University Press.
Norris, Pippa, John Curtice, David Sanders, Margaret Scammell, and Holli E. Semetko.
1999. On Message: Communicating the Campaign. London: Sage.
Norris, Pippa, and Joni Lovenduski. 1995. Political Recruitment: Gender, Race and Class in
the British Parliament. Cambridge: Cambridge University Press.
Norton, Philip. 1994. ‘The growth of the constituency role of the MP’. Parliamentary
Affairs 47 (4): 705–21.
Oddsson, Gudmundur A. 2010. ‘Class awareness in Iceland’. International Journal of
Sociology and Social Policy 30 (5/6): 292–312.
Oesch, Daniel. 2006. Redrawing the Class Map. Stratification and Institutions in Britain,
Germany, Sweden and Switzerland. Basingstoke: Palgrave Macmillan.
Oesch, Daniel. 2008. ‘Explaining workers’ support for right-wing populist parties in
Western Europe: Evidence from Austria, Belgium, France, Norway, and Switzerland’,
International Political Science Review 29 (3): 349–73.
Oesch, Daniel. 2013. Occupational Change in Europe: How Technology and Education
Transform the Job Structure. Oxford: Oxford University Press.
O’Leary, Nigel C., and Peter J. Sloane. 2005. ‘The return to university education in Great
Britain’. National Institute Economic Review 193 (1): 75–89.
Oskarson, Maria. 2005. ‘Social structure and party choice’. In Jacques Thomassen (ed.),
The European Voter. A Comparative Study of Modern Democracies. Oxford: Oxford
University Press, 84–105.
Ossowski, Stanisław. 1963. Class Structure in the Social Consciousness. New York: Free
Press.
Pahl, Ray, David Rose, and Liz Spencer. 2007. ‘Inequality and Quiescence: A Continuing
Conundrum’, ISER Working Paper 2007–22. Colchester: University of Essex.
Pakulski, Jan, and Malcolm Waters. 1996a. The Death of Class. London: Sage.
Pakulski, Jan, and Malcolm Waters. 1996b. ‘The reshaping and dissolution of class’.
Theory and Society 25 (5): 667–91.
Parkin, Frank. 1968. Middle Class Radicalism. Manchester: Manchester University Press.
Parry, Geraint, George Moyser, and Neil Day. 1992. Political Participation and Democracy
in Britain. Cambridge: Cambridge University Press.
Pattie, Charles, and Ron Johnston. 1998. ‘Voter turnout at the British general election
of 1992: Rational choice, social standing or political efficacy?’ European Journal of
Political Research 33 (2): 263–83.
Pattie, Charles, and Ron Johnston. 2001. ‘A low turnout landslide: Abstention at the
British General Election of 1997’. Political Studies 49: 286–305.

226
 References

Pérez-Ahumada, Pablo. 2014. ‘Class consciousness in a mature neoliberal society:

Evidence from Chile’. Research in Social Stratification and Mobility 38: 57–75.
Pierson, Paul. 2001. The New Politics of the Welfare State. Oxford: Oxford University Press.
Piketty, Thomas. 2014. Capital in the Twenty-First Century. Harvard: Harvard University
Press.
Pimlott, Ben. 1992. Harold Wilson. London: Harper & Collins.
Pitkin, Hanna. 1967. The Concept of Representation. Berkeley: University of California Press.
Plutzer, Eric. 2002. ‘Becoming a habitual voter: Inertia, resources, and growth in young
adulthood’. American Political Science Review 96 (1): 41–56.
Plutzer, Eric, and John F. Zipp. 1996. ‘Identity politics and voting for women candi-
dates’. Public Opinion Quarterly 60 (1): 30–57.
Poole, Keith T., and Howard L. Rosenthal. 2011. Ideology and Congress. Piscataway,
NJ: Transaction Publishers.
Popkin, Samuel L. 1991. The Reasoning Voter. Chicago: University of Chicago Press.
Price, Robert, and George Sayers Bain. 1988. ‘The labour force’. In A. H. Halsey (ed.),
British Social Trends Since 1900: A Guide to the Changing Social Structure of Britain.
Palgrave Macmillan UK, 162–201.
Prosser, Christopher. 2014. ‘Building policy scales from manifesto data: A referential
content validity approach’. Electoral Studies 35: 88–101.
Prosser, Christopher, Jon Mellon, and Jane Green. 2016. ‘What mattered most to you
when deciding how to vote in the EU referendum?’. British Election Study, 11July 2016.
Przeworski, Adam. 1985. Capitalism and Social Democracy. Cambridge: Cambridge
University Press.
Przeworski, Adam, and John Sprague. 1986. Paper Stones: A History of Electoral Socialism.
Chicago: University of Chicago Press.
PWC. 2005. ‘The economic benefits of higher education qualifications’. Price Water-
house Cooper: report for the Royal Society of Chemistry and the Institute of Physics.
PWC. 2007. ‘The economic benefits of a degree’. Price Waterhouse Cooper: report for
Universities UK.
Ragnarsdóttir, Berglind Hólm, Jón Gunnar Bernburg, and Sigrún Ólafsdóttir. 2013. ‘The
global financial crisis and individual distress: The role of subjective comparisons after
the collapse of the Icelandic economy’. Sociology 47 (4): 755–75.
Ray, John J. 1982. ‘Authoritarianism/libertarianism as the second dimension of social
attitudes’. The Journal of Social Psychology 117 (1): 33–44.
Ray, John J. 1983. ‘The workers are not authoritarian: Attitude and personality data
from six countries’. Sociology and Social Research 67 (2): 166–89.
Reardon, Sean F., and Kendra Bischoff. 2011. ‘Income inequality and segregation’.
American Journal of Sociology 116 (4): 1092–153.
Rehm, Philipp, and Tim Reilly. 2010. ‘United we stand: Constituency homogeneity and
comparative party polarization’. Electoral Studies 29 (1): 40–53.
Richards, Steve. 2016. ‘Leadership, loyalty and the rise of Jeremy Corbyn’. The Political
Quarterly 87 (1): 12–17.
Rennwald, Line, and Geoffrey Evans. 2014. ‘When supply creates demand: Social-
democratic parties’ electoral strategies and the evolution of class voting’. West
European Politics 37 (5): 1108–35.

227
References

Roberts, Elizabeth. 1984. A Woman’s Place: An Oral History of Working-Class Women.

Oxford: Blackwell.
Roberts, Kath, Nick Cavill, Caroline Hancock, and Harry Rutter. 2013. Social and Eco-
nomic Inequalities in Diet and Physical Activity. London: Public Health England.
Roberts, Robert. 1971. The Classic Slum. Harmondsworth, UK: Penguin Books.
Robertson, David. 1984. Class and the British Electorate. Oxford: Blackwell.
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.
Rose, David, and Eric Harrison (eds.). 2010. Social Class in Europe: An Introduction to the
European Socio-Economic Classification. London: Routledge.
Rose, David, and David Pevalin. 2003. A Researcher’s Guide to the National Statistics Socio-
economic Classification. London: Sage.
Rose, Richard, and Ian McAllister. 1986. Voters Begin to Choose: From Closed Class to Open
Elections in Britain. London: Sage.
Rosenstone, Steven, and John M. Hansen. 1993. Mobilization, Participation, and Democ-
racy in America. New York: Macmillan.
Ross, Lee. 1977. ‘The intuitive psychologist and his shortcomings: Distortions in
the attribution process’. In L. Berkowitz, Advances in Experimental Social Psychology,
vol. 10. New York: Academic Press, 173–220.
Runciman, Walter Gary. 1966. Relative Deprivation and Social Justice: A Study of Attitudes
to Social Inequality in Twentieth-Century England. London: Routledge & Kegan Paul.
Russell, Andrew, and Edward Fieldhouse. 2005. Neither Left Nor Right: The Liberal
Democrats and the Electorate. Manchester: Manchester University Press.
Rydgren, Jens. 2008. ‘Immigration sceptics, xenophobes or racists? Radical right-wing
voting in six West European countries’. European Journal of Political Research 47 (6):
737–65.
Sandbrook, Dominic. 2006. Never Had It So Good: A History of Britain from Suez to the
Beatles, 1956–1963. London: Abacus.
Sandbrook, Dominic. 2012. Seasons in the Sun: The Battle for Britain, 1974–1979.
London: Penguin.
Särlvik, Bo, and Ivor Crewe. 1983. Decade of Dealignment: The Conservative Victory of
1979 and Electoral Trends in the 1970s. Cambridge: Cambridge University Press.
Sartori, Giovanni. 1969. ‘From the sociology of politics to political sociology’. In Seymour
M. Lipset (ed.), Politics and the Social Sciences. Oxford: Oxford University Press, 65–100.
Savage, Mike. 2000. Class Analysis and Social Transformation. Milton Keynes: Open
University Press.
Savage, Mike, Gaynor Bagnall, and Brian J. Longhurst. 2001. ‘Ordinary, ambivalent and
defensive: Class identities in the Northwest of England’. Sociology 35 (4): 874–92.
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 Great British class experiment’. Sociology 47 (2): 219–50.
Savage, Mike, Elizabeth Silva, and Alan Warde. 2010. ‘Dis-identification and class
identity’. In Elizabeth Silva and Alan Warde (eds.), Cultural Analysis and Bourdieu’s
Legacy: Settling Accounts and Developing Alternatives. London: Routledge, 60–74.

228
 References

Schneider, Silke. 2008. ‘Anti-immigrant attitudes in Europe: Outgroup size and

perceived ethnic threat’. European Sociological Review 24 (1): 53–67.
Scott, Janny, and Leonhardt, David. 2005. ‘Shadowy lines that still divide’. The New
York Times, 15 May: p.11.
Sears, David O. 1975. ‘Political socialization’. In Fred I. Greenstein and Nelson
W. Polsby (ed.), Handbook of Political Science: Micropolitical Theory. Reading:
Addison-Wesley.
Sears, David O. 1983. ‘The persistence of early political pre-dispositions: The roles of
attitude objects and life-stage’. In Lad Wheeler and Phillip Shaver (ed.), Review of
Personality and Social Psychology No. 4. Beverly Hills: Sage.
Sears, David O. 1993. ‘Symbolic politics: A socio-psychological theory’. In Shanto
Iyengar and William J. McGuire (ed.), Explorations in Political Psychology. Durham,
NC: Duke University Press.
Sears, David O., Richard R. Lau, Tom R. Tyler, and Harris M. Allen. 1980. ‘Self-interest
versus symbolic politics in policy attitudes and presidential voting’. American Political
Science Review 74 (3): 670–84.
Seawright, David. 2000. ‘A confessional cleavage resurrected? The denominational vote
in Britain’. In David Broughton and Hans-Martien ten Napel (eds.), Religion and Mass
Electoral Behaviour in Europe. London: Routledge, 44–60.
Seawright, David, and John Curtice. 1995. ‘The decline of the Scottish Conservative
and Unionist Party 1950–92: Religion, ideology or economics?’ Contemporary Record
9 (2): 319–42.
Seldon, Anthony (ed.). 2007. Blair’s Britain, 1997–2007. Cambridge: Cambridge
University Press.
Seldon, Anthony. 2008. Speech given to the Head Teachers conference at Wellington
College.
Seyd, Patrick. 1997. ‘Tony Blair and New Labour’. In Anthony King (ed.), New Labour
Triumphs: Britain at the Polls. Chatham, NJ: Chatham House.
Shibutani, Tamotsu. 1955. ‘Reference groups as perspectives’. American Journal of
Sociology 60 (6): 562–9.
Shirazi, Rez, and Anders Biel. 2005. ‘Internal-external causal attributions and perceived
government responsibility for need provision: A 14-culture study’. Journal of Cross-
Cultural Psychology 36 (1): 96–116.
Sides, John, and Jack Citrin. 2007. ‘European opinion about immigration: The role
of identities, interests and information’. British Journal of Political Science 37 (3): 477–504.
Siegel, Alberta E., and Sidney Siegel. 1957. ‘Reference groups, membership groups, and
attitude change’. Journal of Abnormal and Social Psychology 55 (3): 360–4.
Sigelman, Carol, Lee Sigelman, Barbara Walkosz, and Michael Nitz. 1995. ‘Black candi-
dates, white voters: Understanding racial bias in political perceptions’. American
Journal of Political Science 39 (1): 243–65.
Sillitoe, Alan. 1958, 2008. Saturday Night and Sunday Morning. London: Harper Perennial.
Silver, Brian D., Barbara A. Anderson, and Paul R. Abramson. 1996. ‘Who overreports
voting?’ American Political Science Review 80 (2): 613–24.
Skeggs, Beverley. 1997. Formations of Class and Gender. London: Sage.
Skeggs, Beverley. 2004. Class, Self, Culture. London: Routledge.

229
References

Spahn, Bradley, and Matthew Hindman. 2014. ‘Eccentric circles and non-habitual
voting: Turnout patterns in a 2-million-voter national panel dataset’. Midwest Polit-
ical Science Association Meeting: April 4–7, Chicago, IL.
Spies, Dennis. 2013. ‘Explaining working-class support for extreme right parties:
A party competition approach’. Acta Politica 48 (3): 296–325.
Steenbergen, Marco, and Gary Marks. 2007. ‘Evaluating expert surveys’. European
Journal of Political Research 46 (3): 347–66.
Stuart, Graham, Neil Carmichael, Alex Cunningham, Bill Esterson, Pat Glass, Siobahn
McDonagh, Ian Mearns, Caroline Nokes, Dominic Raab, David Ward, and Craig
Whittaker. 2014. ‘Underachievement in education by white working class children’.
House of Commons Education Committee: Report. Available online: http://www.publica
tions.parliament.uk/pa/cm201415/cmselect/cmeduc/142/142.pdf.
Stubager, Rune. 2008. ‘Education effects on authoritarian-libertarian values: A question
of socialisation’. British Journal of Sociology 59 (2): 327–50.
Stubager, Rune. 2009. ‘Education-based group identity and consciousness in the
authoritarian-libertarian value conflict’. European Journal of Political Research 48 (2):
204–33.
Stubager, Rune. 2010. ‘The development of the education cleavage: Denmark as a
critical case’. West European Politics 33 (3): 505–33.
Stubager, Rune, James Tilley, Geoffrey Evans, and Joshua Robinson. 2016. ‘In the eye of
the beholder: What determines how people sort others into social classes?’ Paper
presented at the MPSA annual conference, Chicago, 7–10 April.
Sturgis, Patrick. 2002. ‘Attitudes and measurement error revisited: A reply to Johnston
and Pattie’. British Journal of Political Science 32 (4): 691–8.
Surridge, Paula. 2007. ‘Class belonging: A quantitative exploration of identity and
consciousness’. The British Journal of Sociology 58 (2): 207–26.
Surridge, Paula. 2016. ‘Education and liberalism: Pursuing the link’. Oxford Review of
Education 42 (2): 146–64.
Sutton Trust. 2009. ‘The educational backgrounds of leading lawyers, journalists, vice
chancellors, politicians, medics and chief executives’. The Sutton Trust Submission to
the Milburn Commission on Access to the Professions. Available online: http://www.
suttontrust.com/wp-content/uploads/2009/04/ST_MilburnSubmission.pdf.
Sveinsson, Kjartan P. (ed.). 2009. Who Cares about the White Working Class? London:
Runnymede.
Swaddle, Kevin, and Anthony Heath. 1989. ‘Official and reported turnout in the British
general election of 1987’. British Journal of Political Science 19 (4): 537–51.
Tate, Katherine. 1993. From Protest to Politics: The New Black Voters in American Elections.
Cambridge, MA: Harvard University Press.
Terkildsen, Nayda. 1993. ‘When white voters evaluate black candidates: The processing
implications of candidate skin color, prejudice, and self-monitoring’. American Jour-
nal of Political Science 37 (4): 1032–53.
Thau, Mads. 2016. ‘The changing role of social groups in party competition: From
cleavage to consensus in Britain’. Paper presented at the Midwest Political Science
Association annual meeting, Chicago, 7–10 April.

230
 References

Thomas, James. 1998. ‘Labour, the tabloids, and the 1992 general election’. Contempor-
ary British History 12 (2): 80–104.
Tilley, James. 2005. ‘Libertarian-authoritarian value change in Britain, 1974–2001’.
Political Studies 53 (2): 442–53.
Tilley, James. 2015. ‘We don’t do God? Religion and party choice in Britain’. British
Journal of Political Science 45 (4): 907–27.
Tilley, James, and Anthony Heath. 2007. ‘The decline of British national pride’. British
Journal of Sociology 58 (4): 661–78.
Tolleson Rinehart, Sue. 1992. Gender Consciousness and Politics. New York: Routledge.
Tolsma, Jochem, and Maarten H. J. Wolbers. 2010. ‘Onderwijs als nieuwe sociale
scheidslijn? De gevolgen van onderwijsexpansie voor sociale mobiliteit, de waarde
van diploma’s en het relatieve belang van opleiding in Nederland’. Tijdsschrift Voor
Sociologie 31 (3): 239–59.
Turner, Alwyn. 2015. A Classless Society: Britain in the 1990s. London: Aurum Press.
Universities UK. 2006. Patterns of Higher Education Institutions in the UK: 6th Report.
Available online: http://www.universitiesuk.ac.uk/policy-and-analysis/reports/
Documents/2006/patterns-and-trends-uk-higher-education-06.pdf.
Van de Werfhorst, Herman G., and Nan Dirk De Graaf. 2004. ‘The sources of political
orientations in post-industrial society: Social class and education revisited’. British
Journal of Sociology 55 (2): 211–35.
Van de Werfhorst, Herman G., and Gerbert Kraaykamp. 2001. ‘Four field-related edu-
cational resources and their impact on labor, consumption and sociopolitical orien-
tation’. Sociology of Education 74 (4): 296–317.
Van Heerde-Hudson, Jennifer, and Rosie Campbell. 2015. Parliamentary Candidates UK
Dataset (v. 1).
Vanneman, Reeve, and Lynn W. Cannon. 1987. The American Perception of Class.
Philadelphia, PA: Temple University Press.
Verba, Sidney, and Norman H. Nie. 1972. Participation in America. New York: Harper
and Row.
Verba, Sidney, Norman H. Nie, and Jae-On Kim. 1978. Participation and Political Equal-
ity: A Seven-Nation Study. Cambridge: Cambridge University Press.
Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality:
Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.
Volkens, Andrea, Pola Lehmann, Theres Matthieß, Nicolas Merz, Sven Regel, and Annika
Werner. 2015. The Manifesto Data Collection. Manifesto Project (MRG/CMP/MARPOR).
Version 2015a. Berlin: Wissenschaftszentrum Berlin für Sozialforschung (WZB).
Weakliem, David. 1989. ‘Class and party in Britain, 1964–83’. Sociology 23 (2): 285–97.
Weakliem, David. 2001. ‘Social class and voting: The case against decline’. In Terry
N. Clark and Seymour M. Lipset (eds.), The Breakdown of Class Politics: A Debate on Post-
Industrial Stratification. Baltimore, MD: The Johns Hopkins University Press, 197–223.
Weakliem, David. 2002. ‘The effects of education on political opinions: An inter-
national study’. International Journal of Public Opinion Research 13 (2): 141–57.
Weakliem, David, and Anthony Heath. 1994. ‘Rational choice and class voting’. Ration-
ality and Society 6 (2): 243–70.

231
References

Weakliem, David, and Anthony F. Heath. 1999. ‘The secret life of class voting: Britain,
France, and the United States since the 1930s’. In Geoffrey Evans (ed.), The End of
Class Politics? Class Voting in Comparative Context. Oxford: Oxford University Press,
97–126.
Webb, Paul. 2004. ‘Party responses to the changing electoral markets in Britain’. In
Peter Mair, Wolfgang Muller, and Fritz Plasser (eds.), Political Parties and Electoral
Change. London: Sage.
Weeden, Kim, Young-mi Kim, Matthew Di Carlow, and David B. Grusky. 2007. ‘Social
class and earnings inequality’. American Behavioral Scientist 50 (5): 702–36.
Westergaard, John. 1995. Who Gets What? The Hardening of Class Inequality in the Late
Twentieth Century. Cambridge, UK: Polity Press.
White, Chris, Folkert van Galen, and Yuan Huang Chow. 2003. ‘Trends in social class
differences in mortality by cause, 1986 to 2000’. Health Statistics Quarterly 20: 25–37.
Whiteley, Paul, Harold Clarke, David Sanders, and Marianne Stewart. 2001. ‘Turnout’.
Parliamentary Affairs 54 (4): 775–88.
Willmott, Peter, and Michael Young. 1960. Family and Class in a London Suburb.
London: Mentor.
Williams, Raymond. 1961. The Long Revolution. Harmondsworth, UK: Penguin Books.
Willis, Paul. 1977. Learning to Labour. Farnborough, UK: Saxon House.
Wolfinger, Raymond E., and Steven J. Rosenstone. 1980. Who Votes? New Haven, CT:
Yale University Press.
Wring, Dominic, and David Deacon. 2010. ‘Patterns of press partisanship in the 2010
General Election’. British Politics 5 (4): 436–54.
Young, Michael. 1958. The Rise of the Meritocracy, 1870–2023: An Essay on Education and
Equality. London: Thames and Hudson.
Zucker, Gail S., and Bernard Weiner. 1993. ‘Conservatism and perceptions of poverty:
An attributional analysis’. Journal of Applied Social Psychology 23 (4): 925–44.
Zweig, Ferdynand. 1952. The British Worker. Harmondsworth, UK: Penguin Books.

232
Index

abstention 13, 127, 170, 172, 176, 198, 199 Clarke, Harold 12, 146, 164, 188
Adorno, Theodor 69 class
agenda setting 18, 92, 113 awareness 10, 40–1, 44, 57
Al-Hamad, Alaa 26, 38 conflict 9, 41, 53, 55–6, 84, 91, 97–100, 103,
American Dream 37 106, 110, 113
authoritarianism; see also social conservatism dealignment 12, 145, 163
10, 60, 68–9, 72–3, 75, 79, 167 disidentification 9, 42
division 1–2, 7–10, 12, 16–18, 22, 27, 36,
BBC 16, 37, 145, 165 53–7, 97, 98, 108, 123–4, 146, 165, 187,
Beck, Ulrich 24, 41 191–2
Bernstein, Jared 201 envy 123
Betz, Hans-Georg 186 identification 2, 9, 16, 40–4, 47, 49
Bevan, Aneurin 127 identities 8–9, 40–4, 56–8, 191, 197
Bevin, Ernest 127 non-voting 170, 176, 187–8, 193, 195, 198
Birch, Sarah 201 subjective class 2, 40, 44, 49
birth cohorts 26, 29, 38, 77 coalitions 190, 197
Blair, Tony 92, 108–9, 114, 119, 123, 124–7, Coleman, Richard 50
132, 137, 153, 192, 199 companies 27, 32, 39, 85
Blairism 117, 205 Comparative Manifesto Project 140–1
Blanden, Jo 29 Conservative Party; see also parties 10–11, 92,
Bolton, Paul 27 101–2, 104, 109–12, 114, 116–17, 119,
Bonham, John 2, 192 121–4, 127–8, 132, 147–9, 153, 155,
bourgeoisie 60, 80–2, 94, 113, 174, 188 160, 165–6, 170, 175, 178, 184, 189,
Brand, Jack 180–1 192, 195, 200
Brexit 201–6 Converse, Phillip 12, 102
British Election Study 92, 114, 202 Corbyn, Jeremy 194–5, 206
British Political Speech Archive 124–6 Cowley, Philip 127
broadsheets; see also newspapers 94–6, 101, credentialism; see also education 26
103, 105–8, 110, 115 criminals, attitudes toward 59, 69, 72, 74
Brown, Gordon 109
Busch, Felix 27 Daily Express; see also newspapers 92
Butler, David 49, 50, 92–3, 146 Daily Mail; see also newspapers 92, 94, 97
Daily Mirror; see also newspapers 92, 100–1,
cabinet 31–3, 39, 103, 127, 137, 206 103, 105–6, 108–11, 113–15, 124
Calvert, Peter 2 Daily Telegraph; see also newspapers 92, 94,
Cameron, David 109–10, 114, 181, 194 112–13
Campbell, Rosie 127 death penalty 59, 73–6, 78, 80, 82, 86, 155
Cantril, Hadley 40 Democrats; see also parties 14, 164, 190
career politicians 10 Dickens, Charles 91
censorship 77–8, 80, 87 disillusionment with
Centers, Richard 40, 50 Labour party 186, 206
CEO; see also managers, owners 22, 27, 32–3, 39 political parties 170
chavs 57 Dorling, Danny 201
Churchill, Winston 123–5, 140 Douglas-Home, Alec 103, 137
Index

earnings; see also income 3, 17, 22–3, 31–2, Guardian; see also newspapers 92, 94, 96–7,
37–9, 110 100, 104–10, 114–15
economy 1, 6, 36, 60–2, 68, 83, 110, Gunter, Ray 104
122, 190
economic crisis 27 health 9, 22, 24, 25, 29–31, 36, 107,
economic cycle 25 117, 191
economic recession 57 health inequalities 25, 29–31, 36, 192
Eden, Anthony 109, 137, 148 health policy 31, 107, 117
editorials; see also newspapers 86, 96–105, 108, Heath, Anthony 17, 36, 38–9, 58, 72, 85, 146,
110–14, 192 165, 171, 186, 190
education 5–8, 17, 26–7, 29, 30, 34, 37–8, Heath, Edward 107
47–8, 57, 70–1, 75, 86–7, 155, 159, 166, Heath, Oliver 127, 133, 135, 137, 175,
183–4, 195, 200, 204 188–9, 201
elections 14, 92, 97, 99, 103, 107, 119–21, 130, Hodge, Robert 40
134, 136, 146, 164–5, 170, 174, 178, 182, Hoggart, Richard 36
198–9, 202, 205–6 homosexuality; see also gay rights 59–60, 76,
Electoral Commission 171 78, 167
electorate 7, 11, 14, 107–9, 112, 131, 136, housing 21, 68, 104, 106, 114
146, 149–50, 165, 170, 179, 182, 186, 189, Hout, Michael 164
192, 195, 199, 203–4 Huber, Joan 34
European Elections 182
General Election 14–15, 85, 98, 108, 114, Ichheiser, Gustav 34
137, 141, 145, 148, 152, 171, 178, 191, IMF 117
204–5 immigration 10, 14, 59–60, 68–72, 78–9,
elites 8, 13, 107, 132, 136 83, 84–6, 109, 181–7, 189, 194, 198,
equality of opportunity 33 202–4, 206
Ermisch, John 29 income 2, 4, 9, 10, 16–18, 23–4, 27–9, 33, 37–9,
Europe 14, 17, 86, 181–2, 190, 200–1 40–1, 49, 56, 60–1, 63, 84, 85, 109, 113,
European Common Market 71, 146 172, 191
European Economic Area 204 distribution 23–4
European elections 182 household 15, 23, 28, 30, 39, 41, 92
European Parliament 14 inequality 21, 56, 84, 191
European Union 59, 68–72, 79, 82–4, 86, 97, Industrial Charter 117
182–7, 190, 198, 201–4, 206 inequality 9–10, 21–3, 26–7, 29, 31–4, 37–42,
58, 63–4, 84, 110, 114, 138, 191
fairness 33, 59, 63, 99–100, 114 inflation 105, 117
families 114, 120, 122–3, 126, 137, Inglehart, Ronald 18, 41
142, 192 intelligentsia 80–2, 94–5, 174, 197, 205, 207
farmers 3, 5, 138, 142 interventionism; see also socialism 60, 117
Ford, Robert 86, 182, 201 Ivarsflaten, Elizabeth 86, 181–2, 198
foremen; see also workers 5
Form, William 34 Jackman, Mary 40, 50
Franklin, Mark 146, 165 Jackman, Robert 40, 50
free school meals 26 job
FTSE 27, 63 autonomy 4, 17
security 3–4, 10, 16, 24, 29, 37, 57
Gallie, Duncan 22 Johnson, Brian 26, 38
Gallup 15, 148, 152, 166 Jones, Owen 14
gay rights 80 just world theory; see also fairness 34
GCSEs 26–7, 37
gender 127, 137 Kelley, Jonathan 42, 57
globalisation 9, 21, 201 Keynesian economics 136
Goldthorpe, John 3, 5, 17, 22, 38, 50 Kinnock, Neil 108, 125
Goldthorpe schema 3, 17, 38
Goodwin, Matthew 86, 182 Labour Force Survey 6, 15, 26
general practitioners 31–3, 39 labour market; see also workers 1, 4, 24, 26,
Greene, Hughie 145, 165 29, 31

234
 Index

Labour Party 10–14, 18, 103, 108, 117, 127–8, and voting 11, 13, 16, 18, 57, 60, 107, 127,
137–8, 140–1, 145, 149, 167, 176, 178, 194, 147–8, 151–7, 159, 161–3, 165–6, 173–7,
201, 205 179, 181, 183, 186, 188–90, 193, 195, 198,
landlords; see also housing 105 200, 202–4
Langford, Ann 26 Milburn, Alan 21, 37
left wing; see also social liberalism, socialism 60, Miliband, Ed 114, 124, 194
80–4, 92, 117, 137, 140, 157, 160, 166, Mills, Colin 16, 26, 29
183–4, 190, 194, 198, 206 minorities; see also race 44, 87, 138
Leighley, Jan 172, 190, 199 monopolies 105
Lexicoder 137, 141, 142 Moorhouse, H.F. 50
Liberal Party 92, 104, 153–6, 159–61, Mount, Ferdinand 14
175, 178
Lipset, Seymour 18, 60, 69, 72, 147 Nagler, Jonathan 199
Liverpool 145 National Equality Panel 22
Lowe, Will 139 National Front 182, 189
National Health Service 39, 102, 107
Machin, Stephen 29 national identity 68, 70, 86, 184
Macmillan, Harold 123, 137, 148, 153 National Statistics Socio-Economic
Major, John 123 Classification 17, 22–3, 37, 194,
managers; see also CEOs, owners 5, 17, 22, 37, 206–7
39, 60–1, 106, 109, 134, 138, 148 New Labour; see also Labour Party, parties 13,
manifestos 11, 15, 116–23, 125–6, 136–7, 123, 130
139–40, 192 Newman, Saul 178
Manifesto Project on Political newspapers 10, 11, 13, 15, 91–102, 105–6,
Representation 118, 120–2, 140–1 108–9, 111–13, 136, 161, 163
manufacturing 4 Nicoletti, Cheti 29
marital status 40 North Sea oil; see also Scotland 178
Marshall, Gordon 15, 17, 40, 50–1
Marx, Karl 2, 80 occupational structure 5, 78
May, Theresa 205 ONS Longitudinal Study 26
McAllister, Ian 146 Orwell, George 36
media 11, 15, 18, 84, 91–3, 96–7, 101, 106, owners; see also CEOs, managers 2, 13, 60–2,
112, 116, 127, 135–6, 145, 163–4, 175, 64, 80, 107, 114, 117, 155, 186
181, 194 Oxbridge 128
median voter; see also elections 10
Meret, Susi 186 Pahl, Ray 41
meritocracy; see also fairness 39 panel data 113, 175
middle class parental characteristics 27–9, 38
and identity 9, 16, 41–58, 68, 138, 175, 191, Parliamentary Candidates UK
193, 197 Dataset 128
and income 2, 4, 9–10, 16–18, 23–4, 29, 33, pensions 4, 24
37–8, 41, 49–52, 56, 60–1, 63–5, 84, 109, polarization
113, 191 of class 52–3
and jobs 3–4, 6, 17, 23–4, 31–3, 35, 37–8, 41, of labour markets 4, 22, 37
45, 47, 51–2, 56, 60, 64, 66, 69, 71, 80, of parties 12, 130, 140, 199
83, 94–5, 104, 127, 133, 175–6, 180, 183, policy
192, 202 convergence 11–12
junior middle class 5, 23–5, 37, 44, 61–2, differentiation 117, 136
148, 152, 189–90 politics
new middle class 4–6, 17, 23–5, 36–7, 44, 61, competition 2, 130
63–4, 66, 68–9, 74–5, 77, 94, 96, 148, exclusion 11, 204, 207
151–2, 154–5, 159, 163, 166, 174–5, 180–1, non-participation 16, 112, 200, 205
186, 195–6, 200, 202 participation 14, 16, 18, 167, 171–2, 175,
old middle class 4–5, 21, 61–6, 69, 87, 94, 96, 189, 199–200, 204–5
113, 148, 152, 163, 166, 174, 194 parties 10–11, 16, 91, 100, 105, 112, 116,
and participation 16, 18, 167, 189, 200, 122, 131, 136, 138, 147, 158, 164, 170,
204, 226 191, 202, 204–5

235
Index

politics (cont.) Scottish National Party 14, 138, 148, 153,

representation 2, 57, 117–18 163, 165, 170, 177–81, 187–9, 193, 200
rhetoric 2, 11, 13, 101, 110–11, 117, 123–4, Scottish nationalism 178–80
132, 136, 145, 161, 163–4, 181, 198 Second World War 3, 13, 92, 98, 147
poor; see also poverty 21, 37, 39, 56, 63, 69, 80, shop assistants 31
106–10, 113–14, 120, 122, 125–6, 137, 164, sick pay 24
166, 203–4 Skills and Employment Surveys 22
populism; see also immigration, social social attitudes
conservatism, socialism 182, 186 attitudes, general 6, 16, 92
post-material; see also social liberalism 18, 21, 167 social conservativism 81, 84, 87, 160–1,
post-war consensus 114, 117, 119, 164 167, 194
poverty 21, 29, 39, 41 social liberalism 10, 60, 69, 81–3, 87,
Prescott, John 127 159–60, 167, 184, 192, 194, 206
private schools social atomisation 9
general population 27, 38, 107, 114 Social Democratic Party; see also Liberal Party,
MPs 128–9 parties 18, 194
privatisation; see also Thatcherism 62 social desirability biases 55
professional 5–6, 10, 17, 37, 39, 60–1, 68–70, social dominance 34, 44, 102
86, 95, 106, 109, 128, 138, 148, 152, 166, social mobility 9, 21, 28–9, 36–7, 191
200, 205 Mobility and Child Poverty Commission 21
proletariat; see also workers 80–3, 87, 94–5, social origins 33
174, 197, 205, 207 socialism 68, 117, 123
promotion prospects 3–4, 26, 123 socio-economic group; see also National Socio-
Prosser, Chris 118, 140 economic Classification 3, 38
Protestant work ethic 34 sovereignty; see also European Union 187, 203
Przeworski, Adam 12, 147, 188, 194 speeches 11, 15, 117, 123–6, 132, 136,
public ownership 60, 62, 83, 117, 155, 186 139–42, 192
leader speeches 11, 140
qualifications; see also education 5–8, 17–18, Sprague, John 12, 188, 194
27–8, 38, 47–9, 56, 70, 83, 87, 188–90, 197, status rankings 2
200, 202, 204, 206–7 Steptoe and Son 145
stiffer sentences; see also criminals 74, 80
race 82, 137, 188–9, 206 Stokes, Donald 49–50, 72, 93, 146
racism 86, 181–2, 194 strikers; see also trade unions 104, 105
radical right 14, 86, 170, 181–2, 198, 200 The Sun; see also newspapers 86, 92, 94, 112
Rainwater, Lee 50 The Sutton Trust 21
redistribution; see also left wing, socialism surveys
10, 60, 63, 68, 80, 83, 109, 155, 157, British Election Study 15, 23, 153, 171, 180
186, 198 British Social Attitudes survey 15, 23, 148
reference group theory 11, 41–2, 122, 126 Gallup 148
referendums; see also elections 179–81, 187, Labour Force Survey 15
198, 200–7 ONS Longitudinal Study 26
religion 77, 82, 85, 188–9, 206 Skills and Employment survey 22–3
religiosity 77
religious denomination 62 tabloids; see also newspapers 94–6, 103, 115
revolution; see also socialism 1, 2 taxes 31, 122
right wing 60, 80–2, 116, 118–19, 140, 157, technicians 5, 138
160, 166, 186, 192 Thatcher, Margaret 106–8, 117, 121, 123, 189
risk society 9, 24, 41 Thatcherism 108, 117
Robertson, David 146–7 Thau, Mads 120
Rolls-Royce 107 Three-day week 117
Rose, Richard 146 The Times; see also newspapers 94, 96–7, 100,
Runnymede Trust 16 104–11, 113–15
top down
salaried workers 4, 39 influence 7, 199, 203
Scotland 14, 138, 148, 163, 170, 177–82, 187, model of political change 11–12, 147, 164,
189–90, 193, 200 193, 197, 200

236
 Index

and supply side approach 12 workers 2–5, 17, 21–3, 29, 31–2, 51–2, 60–2,
trade unions 1, 61–4, 66, 72, 80–3, 85, 87, 105, 80, 104–5, 107–8, 113–14, 119–21, 123,
114, 121, 125, 128, 136, 166, 173, 189–90 125, 137–8, 142, 148, 152, 166, 188, 195
Treiman, Donald 40 agriculture 3
trendless fluctuation in class voting 12, factory 31–2, 137
146, 152 low-skilled 21
turnout; see also elections 13, 18, 112, 170–5, manual labourers 2, 5, 17, 52, 107, 138,
177, 186, 188–9, 193, 197–200, 204–5, 207 148, 166
non-manual labourers 2, 5, 138, 148, 166
underclass; see also poor, poverty 122 personal service 5, 17, 166, 195
unemployment 1, 3, 9, 22, 24–5, 27–9, 36–7, self-employed 5, 17
39, 57, 64, 121, 125, 136, 191 working class
rates 24, 25, 27 authoritarianism 10, 60, 68–9, 72, 75, 79
United Kingdom Independence Party; see also conservativism 12, 101–3, 148
parties 14, 86, 148, 153, 163, 165, 170, and identity 9, 16, 41–58, 68, 138, 175, 191,
178, 181–90, 193, 198, 205–6 193, 197
United States 21 and income 2, 4, 10, 16, 18, 21–4, 27, 29, 33,
political participation 13–14, 18, 172, 199 37–8, 41, 49, 50–2, 56, 60–1, 63–5, 84–5,
university; see also Oxbridge 10, 27, 128, 130, 109, 113, 172, 191
134, 136, 204 and jobs 1, 3, 4, 6, 23–4, 27, 31–3, 35, 37–8,
41, 44–5, 47, 51–2, 56, 60, 64, 66, 69, 71,
wages 21, 32, 106, 109–10, 166 80, 83, 85, 94–5, 104, 127–8, 130, 133, 135,
mean percentile wage 23 137, 170, 175–6, 180, 183, 192, 202, 205–7
real wages 21 and participation 14, 16, 18, 167, 172, 189,
Weakliem, David 164–5, 190 199, 200, 204–5
Weber, Max 2 and voting 11–16, 18, 57, 60, 107, 127, 145,
welfare 1, 41, 60, 102, 104, 106, 117, 119, 147–8, 151–7, 161–3, 165–6, 170, 172–9,
122, 136 181, 183, 186–90, 193, 195, 198–200,
welfare state 1, 41, 136 202–6
Westminster 11, 178 working conditions 4, 104, 166
white British; see also race 26, 27
Wilson, Harold 123, 145, 165 Zweig, Ferdynand 3, 50

237