Cultural Capital, Rational Choice and Educational Inequalities [PDF]

Aug 30, 2003 - hardly be seen as a simple matter of choice, even if the child does have some ...... factors may be impor

1 downloads 12 Views 1MB Size

Recommend Stories


Rational Choice and Social Theory
If you feel beautiful, then you are. Even if you don't, you still are. Terri Guillemets

Educational inequalities in hypertension
Courage doesn't always roar. Sometimes courage is the quiet voice at the end of the day saying, "I will

Barnga and Cultural Capital
When you talk, you are only repeating what you already know. But if you listen, you may learn something

Parental Cultural Capital and Educational Attainment in the Netherlands
No amount of guilt can solve the past, and no amount of anxiety can change the future. Anonymous

Cultural Capital
I cannot do all the good that the world needs, but the world needs all the good that I can do. Jana

cultural capital
Silence is the language of God, all else is poor translation. Rumi

Islanders' educational choice
Don’t grieve. Anything you lose comes round in another form. Rumi

PdF Infections and Inequalities
You miss 100% of the shots you don’t take. Wayne Gretzky

Human Capital, Occupational Choice and Inequality
Be grateful for whoever comes, because each has been sent as a guide from beyond. Rumi

Idea Transcript


Cultural Capital, Rational Choice and Educational Inequalities Alice Sullivan August 30, 2003

Introduction Background The educational disadvantage suffered by children from working class families has long been the focus of sociological attention. Traditionally, the reasons for this disadvantage have been taken as being fairly obvious, and therefore the focus was on promoting educational reforms such as the universal provision of free and compulsory education. The problem to be solved was the wastage of working class talent, rather than social class differentials in educational attainment per se (Lindsay, 1926). There is evidence that educational reforms have reduced absolute differences in rates of educational participation between the classes (Jonsson and Mills, 1993a; Jonsson and Mills, 1993b; Hellevik, 1997). However, it seems that the association between social class and educational attainment has remained intact despite these reforms (Halsey et al., 1980; Shavit and Blossfeld, 1993). The failure of educational reforms to eradicate the link between social class and educational attainment has led sociologists to focus on the question of why this association exists. The emergence of “culturalist” approaches can be seen as a natural consequence of the failure of educational reforms to dramatically alter the association between social class and educational outcomes. If the lifting of economic barriers to educational participation did not eradicate social class differences in educational outcomes, then it may be fruitful to look to cultural, rather than (or perhaps, as well as) economic differences between the classes to explain class inequalities in educational attainment. One cultural difference that has been invoked to explain the social class differential in educational attainment is that of language. Bernstein focuses on class differences in language to explain working class educational under-

i

ii achievement. Bernstein, for instance, claims that working class people have access to only a “restricted” language code whereas middle class people have access to an “elaborated” code. In the restricted code “The meanings are likely to be concrete, descriptive or narrative rather than analytical or abstract.” (Bernstein, 1973, p. 128). Bernstein claims that the school is based on the elaborated code, in that it transmits “de-contextualised” and “universalistic” meanings. This puts working class students at a disadvantage, because of the gap between their code, and the code of the school. Bernstein has been misinterpreted (for instance, by Labov (Labov, 1972)) as a deficit theorist, blaming working class children for their own educational failure. A more relevant criticism is it is not clear that working class people actually do have a “restricted code”, as opposed to perhaps simply having relatively smaller vocabularies than middle class people (see Lawton (Lawton, 1968)). Neither is there any empirical evidence that the education system is (or was) based on an elaborated code. Such evidence as there is (Cooper, 1976) points to the opposite conclusion. Advocates of theories of “cultural reproduction” claim that the function of the school is to legitimate economic inequalities by giving an appearance of meritocracy, see for instance Collins (Collins, 1971),Jencks (Jencks, 1972) and Bowles and Gintis (Bowles and Gintis, 1976). Pierre Bourdieu is the most influential advocate of the “cultural reproduction” approach. Bourdieu’s work has strong parallels with that of Bernstein, not least in his focus on the use of language. Bourdieu’s theory of cultural reproduction states that the school rewards pupils who possess the cultural traits associated with the middle classes. These include the ability to use “educated” language, as well as participation in “high-brow” cultural activities. Another parallel between Bernstein and Bourdieu is the lack of clarity in their prose styles, and the corresponding difficulty in deriving testable hypotheses from their theories. Perhaps because Bourdieu is the most prominent theorist of cultural reproduction, his theory has inspired a great deal of empirical work, despite the problems inherent in attempting to operationalise his concepts. It should be noted that there is nothing new in the observation that the middle class child often enjoys cultural as well as economic advantages. This insight need not go hand in hand with an acceptance of “cultural reproduction” theory as such. Floud et. al. (Floud et al., 1956) divide the resources associated with the home into “material” and “cultural” categories. Their measure of cultural resources includes parents’ knowledge of the selection

iii procedures of the grammar schools, parents’ visits to the child’s school, parents’ aspirations and preferences for the child’s education, newspapers and magazines read and library membership. In retrospect, this measure seems to combine some elements of both Bourdieu’s concept of “cultural capital” and Coleman’s concept of “social capital” (Coleman, 1990). The “cultural reproduction” approach stands in marked contrast to the “human capital” approach put forward by Becker (Becker, 1964) according to which education makes people more productive, and this is why people with higher levels of education are rewarded with relatively high pay. More generally, cultural reproduction is seen as being in conflict with the “rational choice” approach, which tends to stress the importance of economic constraints in explaining social class differentials in educational attainment. Boudon (Boudon, 1974) makes a crucial distinction between the “primary” and “secondary” effects of stratification. The primary effects of stratification are cultural inequalities that determine the academic abilities of pupils. The secondary effects of stratification are the different costs and benefits that are associated with different educational decisions for pupils from different social classes. The secondary effects explain any social class difference in educational participation that remains once one has controlled for performance at the previous stage. Boudon claims that the secondary effects of stratification are more important than the primary effects, as the secondary effects are exponential. It seems natural that rational choice theorists should focus on key transition points rather than on class differences in educational attainment within a given curriculum, since choices are made at these transition points, whereas one’s level of performance, particularly in the earliest years of schooling, can hardly be seen as a simple matter of choice, even if the child does have some autonomy in the question of how hard to work. However, there is no contradiction inherent in drawing upon both “cultural reproduction” and “rational choice” perspectives to explain educational inequalities. Furthermore, the question of whether economic or cultural factors are more important in explaining social class differentials in educational attainment can be seen as distinct from the question of whether the primary effects of stratification are greater than the secondary effects. It may be that economic factors have a strong impact on performance before any key transitions have been made, and it may also be that cultural factors affect the decisions made regarding these transitions. The literature I have cited so far has focused on social class inequalities,

iv and has ignored other structures of inequality, notably gender. Traditionally, gender inequalities in educational performance were seen as natural and unproblematic. All male samples were used in both quantitative and qualitative work, (see for instance, Halsey et. al. (Halsey et al., 1980), Willis (Willis, 1977)) and women and girls were therefore excluded as subjects of analysis. More recently, although mixed samples have been collected by quantitative researchers, the characteristics of mothers have sometimes been deemed irrelevant, and gender differentials in educational attainment have often not been seen as worthy of explanation. Feminist ethnographic work, though taking gender differences seriously, has often been based on an essentialist conception of gender differences as being intrinsic and fundamental. So, it has been argued that achievement and rationality are male values, and girls will inevitably fail in an educational system which stresses these values. From the fact that boys’ academic achievement has traditionally been higher than that of girls it is inferred that “achievement” is a male value. So, Arnot claims that girls have different values than boys, and girls who fail in school are merely expressing these values. They “. . . celebrate their femininity through a rejection of male culture that stresses the value of hierarchies. . . objective versus subjective knowledge, and individual competition above co operation” (Arnot, 1994, p. 95).The traditional view of gender differences has a great deal in common with the essentialist view. Both are ahistorical, and neither can account for the fact that girls are increasingly out performing boys educationally in many industrialised countries (Erikson and Jonsson, 1996; Shavit and Blossfeld, 1993). Of course, it is not surprising that girls’ educational performance has improved given womens’ increased labour market participation. However, the fact that girls are actually outperforming boys at GCSE has caused a certain amount of panic. A discourse of male disadvantage has emerged, according to which, boys are failing because of a “crisis of masculinity” or because too many teachers are women (Weiner et al., 1997). This discourse ignores the fact that boys’ educational attainment has not deteriorated. It has improved, but just not as much as that of girls.

v

Aims This thesis aims to examine the mechanisms behind social class and gender differences in educational outcomes, and to evaluate the usefulness of cultural reproduction theory and rational choice theory in explaining educational inequalities. Bourdieu’s theory of cultural reproduction is often vague and ambiguous (for further discussion of Bourdieu’s theory see chapter 1. The main concept that empirical researchers have taken from the theory is that of cultural capital. Cultural capital can be defined loosely as those cultural traits that help people to gain educational success. Empirical researchers have often failed to grapple with the problem of operationalising the concept of cultural capital, preferring to use proxy measures such as parental education. A key aim of this thesis is to clarify the concept of cultural capital. As broad as possible an operationalisation of cultural capital will be used in order to assess which aspects of cultural participation actually constitute cultural capital in the sense of having a pay off in terms of educational performance. Bourdieu’s theory of cultural reproduction is also unclear on the question of why cultural participation should lead to educational success. I will assess the possibility that relatively high levels of linguistic ability and/or cultural knowledge are the mechanisms through which cultural participation translates into educational success. Rational choice theory asserts that individuals act rationally in order to fulfil their desires in the light of their beliefs about the situation (or at least that, on average, people behave as though this was what they were doing). In itself, this is a rather vacuous theory. If it is to yield determinate predictions, ancillary assumptions regarding the beliefs and desires of agents must be made. Rational choice theorists who have attempted to give explanations of educational inequalities have made many assumptions about the motivations of agents, but have been loath to apply any empirical test to these assumptions. Data on actions is seen as “hard” and therefore reliable, whereas data on the beliefs and desires of agents is seen as “soft” and therefore unreliable. The doctrine of “revealed preference” has led rational choice theorists to attempt to infer motivations from actions. The problem with this is, of course, that one can infer any number of quite different motivations from any given action. This is reflected in the wide variety of explanations for social class inequalities in educational outcomes given by rational choice theorists. This thesis will attempt to directly examine whether pupils’ beliefs (about their academic abilities) and attitudes towards education vary according to

vi social class and gender. Plans for further education will be analysed in order to assess whether significant social class and gender differences remain once GCSE performance has been controlled for, and in order to attempt to explain any such differences. Various versions of rational choice and cultural reproduction explanations of educational inequalities have been both advocated or rejected on the basis of inadequate empirical evidence or even no empirical evidence at all. A common problem has been the tendency of researchers to attempt to answer questions using the data that is to hand, regardless of whether this data can actually provide a good test of hypotheses derived from the theory to be assessed. I have tried to collect appropriate data to answer the questions I am posing here. My sample consists of 465 pupils in their final year of compulsory education in four English schools. From the point of view of examining educational differentials, an advantage of the contemporary British context is that the first crucial educational “branching point” has been pushed back to the end of compulsory education, age 16. This means that in analysing results in the GCSE exams, the whole ability range is covered, and one does not risk conflating the “primary effects” of stratification with the “secondary effects” in Boudon’s sense. The GCSE (General Certificate of Secondary Education) exams are taken by all pupils at the end of their compulsory schooling, when they are approximately 16 years old. The GCSE examination was introduced in England and Wales in 1985, replacing CSEs and O levels. CSEs were of a lower academic standard than O levels, and were designed for pupils in the secondary modern schools. (Before 1965, Britain had a tripartite system of secondary education, and children were allocated to grammar, technical, or secondary modern schools at age 11. Subsequently, local authorities gradually introduced the comprehensive system, and there are now only 162 selective state schools in Britain). The introduction of the GCSE means that all fourteen to sixteen year olds, regardless of academic ability, follow a similar curriculum (although there is some choice of subjects). Typically, students study 8 or 9 subjects, and are assessed partly by examination and partly by coursework. Pass grades range from G to A*. (The A* grade was recently introduced due to fears that there was not enough differentiation at the top level). More generally, there has been an increasing focus on credential outcomes within the education system. A national curriculum has been introduced, with standard tests for seven, eleven and fourteen year olds, and the publication of “league tables” of schools’ results in these tests, along with increased parental choice

vii of school, has put great pressure on schools to focus on improving test and examination results overall. I will attempt to answer the following key questions. 1. Cultural Reproduction and GCSE performance • Are cultural practices transmitted from parents to children? • What is cultural capital and how important is it for educational attainment and participation? • Which elements of the dominant culture are actually associated with educational success? • To what extent can differences in cultural participation account for social class differences in educational outcomes? • Are there gender differences in cultural participation, and, if so, do these help to explain girls’ superior performance compared to boys at GCSE? • If cultural participation is associated with educational performance, why is this so? 2. Rational Choice and Post-16 Decisions • How important are the “secondary effects” of stratification at the key transition from GCSE to further education? Is it the case, for instance, that working class pupils are less likely to choose to do A levels once GCSE performance has been controlled for? • If there are significant secondary effects of stratification, how does the size of these effects compare to the size of social class differences in GCSE performance? • To what extent is subject choice associated with pupils’ sex? • How can we explain any social class and gender differences in educational decision making, controlling for GCSE results? • Do attitudes towards education and beliefs about ability differ according to social class and gender? • Do occupational aspirations differ according to social class and gender? • Are occupational aspirations associated with educational decisions, controlling for other relevant variables?

viii I have divided the key questions above into those associated primarily with a cultural reproduction approach and with initial academic performance (or the primary effects of stratification) and those associated with a rational choice approach and with educational decision making (or the secondary effects of stratification). Although this is a convenient classification, the comment I made above that the primary effects of stratification may well not be entirely cultural, nor the secondary effects entirely economic, will not be forgotten in the analysis to follow.

Outline Introduction

Bourdieu’s Theory of Cultural Reproduction In this chapter I will outline and discuss Bourdieu’s theory, and both Bourdieu’s attempts to apply empirical evidence to this theory, and the work of other researchers on this theme. It will be argued that previous operationalisations of the concept have been inadequate, and that a broader operationalisation of the concept of cultural capital is necessary.

Rational Choice Theories of Educational Inequality This chapter will discuss rational choice theory, and the various attempts to apply rational choice theory to the question of educational inequalities. I will argue that rational choice theorists have made various assumptions regarding the motivations of agents, but that these assumptions are in need of empirical test.

Methodology This chapter will describe my operationalisation of concepts, questionnaire design, sample, and data collection process.

Cultural Capital and Educational Attainment This chapter includes an empirical analysis of the link between parents’ cultural participation and pupils’ cultural participation, and between pupils’ cultural participation and GCSE attainment. I will show that reading and television viewing habits are the only types of cultural participation that are associated with GCSE

ix attainment, once other relevant factors have been controlled for, and the effect of these variables is mediated entirely by pupils’ cultural knowledge and vocabulary.

Students’ Beliefs and Desires This chapter will assess the question of whether students’ attitudes to education and beliefs about their own academic abilities vary according to social background and gender. I will conclude that pupils’ attitudes to education vary only slightly by social class, and that this effect is entirely mediated by school success as measured by GCSE results. However, pupils’ attitudes do vary by gender, and both the social background and gender of pupils affect their perception of their own abilities, with service class and male pupils’ overestimating their own abilities as compared to non-service class and female pupils.

Students’ Decisions This chapter will address the question of social class and gender differences in rates of educational participation post-16. Social class has no direct effect on participation in further education for my sample. So, in this case, the ‘secondary effects’ of stratification seem to be nonexistent. However, parental education does have a significant association with the decision to do A levels, controlling for pupils’ GCSE results. Girls are significantly less likely to choose hard science subjects than are boys, and this seems to be partly explained by girls’ relatively lower assessment of their own abilities.

Conclusions This chapter will summarise my results and provide some suggestions for future research.

Contents

1 Cultural Reproduction

1

1.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

1.2

Bourdieu’s theory . . . . . . . . . . . . . . . . . . . . . . . . .

2

1.2.1

Cultural Capital . . . . . . . . . . . . . . . . . . . . .

2

1.2.2

Habitus . . . . . . . . . . . . . . . . . . . . . . . . . .

9

1.2.3

Political Action . . . . . . . . . . . . . . . . . . . . . . 13

1.2.4

Science and Language . . . . . . . . . . . . . . . . . . 15

1.3

1.4

Empirical Evidence on Cultural Capital . . . . . . . . . . . . . 17 1.3.1

Bourdieu’s Own Evidence . . . . . . . . . . . . . . . . 17

1.3.2

Other Research on Cultural Reproduction . . . . . . . 19

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2 Rational Choice Perspectives

24

2.1

The Utility Model . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.2

Murphy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.2.1

Formalising Murphy . . . . . . . . . . . . . . . . . . . 29

2.3

Boudon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

2.4

Gambetta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 x

CONTENTS

xi

2.4.1

Formalising Gambetta . . . . . . . . . . . . . . . . . . 34

2.5

2.6

Breen and Goldthorpe’s Model . . . . . . . . . . . . . . . . . . 34 2.5.1

Applying the Utility Model to Breen and Goldthorpe . 38

2.5.2

Breen and Goldthorpe’s Assumptions . . . . . . . . . . 41

2.5.3

A Generic Model? . . . . . . . . . . . . . . . . . . . . . 44

2.5.4

Breen and Goldthorpe’s attempt to explain empirical changes . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3 Methodology 3.1

3.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.1.1

Cultural Capital . . . . . . . . . . . . . . . . . . . . . 50

3.1.2

Rational Choice Theory . . . . . . . . . . . . . . . . . 51

Survey Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 3.2.1

3.3

3.4

50

Pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Questionnaire Design and Coding . . . . . . . . . . . . . . . . 56 3.3.1

Cultural Capital Questions . . . . . . . . . . . . . . . . 56

3.3.2

Viewing Habits . . . . . . . . . . . . . . . . . . . . . . 66

3.3.3

Formal Culture and Music

3.3.4

Parents’ Cultural Capital . . . . . . . . . . . . . . . . . 70

3.3.5

Validity of Cultural Capital Questions . . . . . . . . . 71

3.3.6

Rational Choice Questions . . . . . . . . . . . . . . . . 73

3.3.7

Reliability of Rational Choice Questions . . . . . . . . 74

3.3.8

Parents’ Social Class and Qualifications . . . . . . . . . 75

. . . . . . . . . . . . . . . 68

Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

CONTENTS

xii

3.4.1

Administration of questionnaire . . . . . . . . . . . . . 78

3.5

Response Rates . . . . . . . . . . . . . . . . . . . . . . . . . . 80

3.6

Ethics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

3.7

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

4 Educational Attainment at GCSE

83

4.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

4.2

Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

4.3

4.2.1

Parental Cultural Capital . . . . . . . . . . . . . . . . 84

4.2.2

Pupils’ Cultural Capital . . . . . . . . . . . . . . . . . 85

4.2.3

Language and Knowledge . . . . . . . . . . . . . . . . 88

4.2.4

Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

4.2.5

GCSE attainment . . . . . . . . . . . . . . . . . . . . . 95

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

5 Beliefs and Desires

111

5.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

5.2

Rational Choice Theories . . . . . . . . . . . . . . . . . . . . . 113

5.3

Research Questions . . . . . . . . . . . . . . . . . . . . . . . . 115

5.4

Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

5.5

5.4.1

Attitudes to Education . . . . . . . . . . . . . . . . . . 116

5.4.2

Beliefs About Ability . . . . . . . . . . . . . . . . . . . 121

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

6 Pupils’ Choices 6.1

129

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

CONTENTS

xiii

6.2

Research Questions . . . . . . . . . . . . . . . . . . . . . . . . 131

6.3

Occupational Aspirations . . . . . . . . . . . . . . . . . . . . . 132

6.4

Attitudes to Jobs . . . . . . . . . . . . . . . . . . . . . . . . . 134

6.5

Occupational Aspirations and Values . . . . . . . . . . . . . . 140

6.6

6.7

6.5.1

Occupational values . . . . . . . . . . . . . . . . . . . . 141

6.5.2

Occupational Aspirations . . . . . . . . . . . . . . . . . 149

Educational Choices . . . . . . . . . . . . . . . . . . . . . . . 152 6.6.1

Choosing to stay on in different courses . . . . . . . . . 152

6.6.2

Further Qualifications . . . . . . . . . . . . . . . . . . 157

6.6.3

A Levels . . . . . . . . . . . . . . . . . . . . . . . . . . 158

6.6.4

Subject Choices . . . . . . . . . . . . . . . . . . . . . . 161

6.6.5

Gender and Subject Choice . . . . . . . . . . . . . . . 163

6.6.6

GNVQs . . . . . . . . . . . . . . . . . . . . . . . . . . 169

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

7 Conclusions

171

7.1

Cultural Reproduction . . . . . . . . . . . . . . . . . . . . . . 172

7.2

Rational Choice . . . . . . . . . . . . . . . . . . . . . . . . . . 173

7.3

Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

7.4

Family Size and Structure . . . . . . . . . . . . . . . . . . . . 178

7.5

Suggestions for future research . . . . . . . . . . . . . . . . . . 179

7.6

General Theoretical Conclusions . . . . . . . . . . . . . . . . . 181

A Correlations

184

B Questionnaire

187

CONTENTS C Educational Attitudes Scale

xiv 204

C.1 Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 D Coding

207

List of Figures 2.1

A generic utility model . . . . . . . . . . . . . . . . . . . . . . 26

2.2

Murphy’s model . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.3

Breen and Goldthorpe’s model . . . . . . . . . . . . . . . . . . 39

3.1

Histogram of knowledge score . . . . . . . . . . . . . . . . . . 60

3.2

Histogram of passive vocabulary score . . . . . . . . . . . . . . 61

3.3

Histogram of active vocabulary score . . . . . . . . . . . . . . 62

3.4

Histogram of language score . . . . . . . . . . . . . . . . . . . 63

3.5

Histogram of reading cultural capital score . . . . . . . . . . . 65

4.1

Parents’ cultural participation by parents’ class . . . . . . . . 85

4.2

Parents’ cultural participation by parents’ educational level . . 87

xv

List of Tables 2.1

Perceived Utilities in Breen and Goldthorpe’s Model . . . . . . 40

3.1

Knowledge of cultural figures . . . . . . . . . . . . . . . . . . 59

3.2

Coding of Books . . . . . . . . . . . . . . . . . . . . . . . . . 64

3.3

Pupils’ reading habits . . . . . . . . . . . . . . . . . . . . . . . 65

3.4

Coding of television programmes . . . . . . . . . . . . . . . . 67

3.5

Formal culture and music . . . . . . . . . . . . . . . . . . . . 69

3.6

Social class breakdown for each school . . . . . . . . . . . . . 78

4.1

Pupils’ cultural participation . . . . . . . . . . . . . . . . . . . 86

4.2

Language score . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4.3

Cultural knowledge scores . . . . . . . . . . . . . . . . . . . . 92

4.4

Gender differences in cultural participation . . . . . . . . . . . 94

4.5

GCSE results . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

4.6

GCSE results using SOC class schema . . . . . . . . . . . . . 100

4.7

English GCSE results . . . . . . . . . . . . . . . . . . . . . . . 103

4.8

Maths GCSE results . . . . . . . . . . . . . . . . . . . . . . . 105

5.1

Pupils’ attitudes to education — mean scores . . . . . . . . . 117

xvi

LIST OF TABLES

xvii

5.2

Pupils’ attitudes to education — analysis of variance . . . . . 118

5.3

Pupils’ attitudes to education using bivariate parental class and education . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

5.4

Pupils’ attitudes to education controlling for GCSE results . . 119

5.5

Pupils’ self-assessed ability . . . . . . . . . . . . . . . . . . . . 121

5.6

Pupils’ self-predicted grades . . . . . . . . . . . . . . . . . . . 122

5.7

Pupils’ self-predicted grades and actual grades . . . . . . . . . 122

5.8

Self-assessed ability by gender . . . . . . . . . . . . . . . . . . 123

5.9

Self-predicted grades by gender . . . . . . . . . . . . . . . . . 124

5.10 Pupils’ self-predicted grades and estimation of own abilities . . 124 6.1

Occupational aspirations by gender . . . . . . . . . . . . . . . 133

6.2

Occupational aspirations by parental class . . . . . . . . . . . 135

6.3

Occupational values by gender . . . . . . . . . . . . . . . . . . 136

6.4

Factor analysis of occupational values . . . . . . . . . . . . . . 138

6.5

Occupational value categories by gender . . . . . . . . . . . . 139

6.6

Occupational values by parental class . . . . . . . . . . . . . . 139

6.7

Models of desire for intellectual stimulation . . . . . . . . . . . 142

6.8

Material values . . . . . . . . . . . . . . . . . . . . . . . . . . 144

6.9

Control and respect . . . . . . . . . . . . . . . . . . . . . . . . 147

6.10 Job aspirations . . . . . . . . . . . . . . . . . . . . . . . . . . 150 6.11 Educational intentions and gender . . . . . . . . . . . . . . . . 153 6.12 Educational intentions and social class . . . . . . . . . . . . . 153 6.13 Reliability of A level choices . . . . . . . . . . . . . . . . . . . 154 6.14 Reliability of further education choices . . . . . . . . . . . . . 154

LIST OF TABLES

xviii

6.15 Pupils’ intention to puruse further education . . . . . . . . . . 155 6.16 Pupils’ intention to pursue A-levels . . . . . . . . . . . . . . . 159 6.17 A level subject choice by gender . . . . . . . . . . . . . . . . . 161 6.18 A level subject choice by class . . . . . . . . . . . . . . . . . . 162 6.19 GNVQ subject choice by gender . . . . . . . . . . . . . . . . . 163 6.20 Pupils’ choice of any hard science A levels . . . . . . . . . . . 164 6.21 Proportion of hard sciences

. . . . . . . . . . . . . . . . . . . 166

6.22 GNVQ subject choices . . . . . . . . . . . . . . . . . . . . . . 167 C.1 Attitudes to education scale . . . . . . . . . . . . . . . . . . . 205 C.2 Factor analysis of attitudes to education . . . . . . . . . . . . 206

Chapter 1 Bourdieu’s Theory of Cultural Reproduction 1.1

Introduction

This chapter will examine Bourdieu’s theory of cultural reproduction, and the use empirical researchers have made of this theory. Bourdieu’s work must be seen in the context both of the debate on class inequalities in educational attainment and of broader questions of class reproduction in advanced capitalist societies. Bourdieu is interested in the link between original class membership and ultimate class membership, and in the way in which this link is mediated by the educational system. According to Bourdieu, the education systems of industrialised societies function in such a way as to legitimate class inequalities. Success in the education system is facilitated by the possession of cultural capital and of higher-class habitus. Lower class pupils do not in general possess these traits, so the failure of the majority of these pupils is inevitable. This explains class inequalities in educational attainment. However, success and failure in the educational system is seen as being due to individual gifts (or the lack of them). Therefore, for Bourdieu, educational credentials help to reproduce and legitimate social inequalities, as higher class individuals are seen to deserve their place in the social structure. The first part of this chapter will consist of a general discussion of Bourdieu’s theory of education, with particular reference to the concepts of cultural 1

CHAPTER 1. CULTURAL REPRODUCTION

2

capital and habitus. I will argue that the concept of habitus is theoretically incoherent and has no possible use for empirical researchers. The concept of cultural capital, on the other hand, while not constructed in the most precise and lucid manner possible by Bourdieu, is substantive enough to be potentially extremely useful to empirical researchers. The second section of this chapter will therefore assess some of the empirical work concerning cultural capital and the problems of operationalising the concept.

1.2

Bourdieu’s theory

1.2.1

Cultural Capital

Introduction to Cultural Capital Bourdieu states that cultural capital consists of familiarity with the dominant culture in a society, and especially the ability to understand and use “educated” language. The possession of cultural capital varies with social class, yet the education system assumes the possession of cultural capital. This makes it very difficult for lower class pupils to succeed in the education system. “By doing away with giving explicitly to everyone what it implicitly demands of everyone, the educational system demands of everyone alike that they have what it does not give. This consists mainly of linguistic and cultural competence and that relationship of familiarity with culture which can only be produced by family upbringing when it transmits the dominant culture.” (Bourdieu, 1977a, p. 494) Bourdieu claims that, since the educational system presupposes the possession of cultural capital, which few students in fact possess, there is a great deal of inefficiency in “pedagogic transmission”. This is because students simply do not understand what their teachers are trying to get across. For Bourdieu, this is particularly apparent in the universities, where students, afraid of revealing the extent of their ignorance “. . . minimize the risks by throwing a smoke-screen of vagueness over the possibility of truth or error.” (Bourdieu and Passeron, 1990, p. 114)

CHAPTER 1. CULTURAL REPRODUCTION

3

But despite the fact that lower class pupils are seriously disadvantaged in the competition for educational credentials, the results of this competition are seen as meritocratic and therefore as legitimate. In addition, Bourdieu claims that social inequalities are legitimated by the educational credentials held by those in dominant positions. This means that the educational system has a key role in maintaining the status quo. “. . . it [education] is in fact one of the most effective means of perpetuating the existing social pattern, as it both provides an apparent justification for social inequalities and gives recognition to the cultural heritage, that is, to a social gift treated as a natural one.” (Bourdieu, 1974, p. 32) In sum, Bourdieu’s view is that cultural capital is inculcated in the higher class home, and enables higher class students to gain higher educational credentials than lower class students. This enables higher class individuals to maintain their class position, and legitimates the dominant position which higher class individuals typically goes on to hold. Of course, some lower class individuals will succeed in the educational system, but, rather than challenging the system, this will strengthen it by contributing to the appearance of meritocracy. Bourdieu can be criticised for not being precise enough about exactly which of the resources associated with the higher class home constitute cultural capital, and how these resources are converted into educational credentials. However, I think that the concept of cultural capital is substantive enough to be operationalised, although Bourdieu does not make it at all obvious how this should be done. The empirical evidence Bourdieu gives for the link between class and cultural capital and cultural capital and educational success is rather weak. Neither does Bourdieu give any evidence for the view that educational credentials serve to legitimate class inequalities. Bourdieu’s emphasis on the non-material resources possessed by the higherclass household is to be welcomed. We have evidence (see (Halsey et al., 1980)) that the dramatic fall in the material costs to families of education due to educational reforms, such as the universal provision of free and compulsory secondary education, have not diminished the degree of association between class origins and educational attainment. This suggests that the educational advantage which higher class parents pass on to their children may not be entirely caused by economic factors, and that the notion of cultural capital is therefore worthy of serious attention. One does not have to accept all the

CHAPTER 1. CULTURAL REPRODUCTION

4

trappings of Bourdieu’s grand theory in order to acknowledge that cultural factors may be important in generating inequalities in educational outcomes.

Cultural Capital vs. Other Forms of Capital Bourdieu has been accused of giving too much weight to symbolic relations at the expense of material ones (Willis, 1983). Yet Bourdieu refers to economic capital and social capital (social relationships and networks) as well as symbolic and cultural capital (Bourdieu, 1989, p. 230). I will try to assess how important Bourdieu thinks cultural capital is in relation to other forms of capital. Bourdieu is characteristically unclear on this point. “Apart from the fact that the increase in the proportion of holders of the most prestigious academic qualifications among the ruling classes may mean only the need to call upon academic approval in order to legitimate the transmission of power and privileges is being more and more felt, the effect is as though the cultural and educational mechanisms had merely strengthened or taken over from the traditional mechanisms such as the hereditary transmissions of economic capital, of a name or of capital in terms of social relationships. . . ” (Bourdieu, 1977a, p. 496) Have cultural and educational mechanisms “merely strengthened” traditional mechanisms or reproduction or have they “taken over from” such traditional mechanisms? Bourdieu slides from the former to the latter claim as if there were not much to choose between them. And as if that was not vague enough, Bourdieu is not actually claiming to describe reality, but merely states that “the effect is as though” this claim were true. In fact, (the first part of the sentence implies) educational credentials may not be necessary to secure privileges at all, but only to legitimate them. In short this passage is quite incoherent. We are left with no clear idea of Bourdieu’s view of the importance of cultural and educational capital in the transmission of privileges. It is easy to find passages in Bourdieu in which the role of educational credentials in social reproduction is stressed. For instance, “In ever more completely delegating the power of selection to the academic institution, the privileged classes are able to appear to

CHAPTER 1. CULTURAL REPRODUCTION

5

be surrendering to a perfectly neutral authority the power of transmitting power from one generation to another. . . ” (Bourdieu and Passeron, 1990, p. 167) In this passage, the power of the educational system as a mechanism of social selection is “ever more complete”. Bourdieu seems to be stating that the importance of educational capital is supreme and increasing. However, elsewhere Bourdieu admits the importance of inheriting economic capital, and of social networks which are extremely valuable in the job market. “The same academic qualifications receive very variable values and functions according to the economic and social capital (particularly the capital of relationships inherited from the family) which those who hold these qualifications have at their disposal. . . ” (Bourdieu, 1977a, p. 506) Bourdieu goes further than this, stating that possessors of economic capital are “able to do without” cultural capital, “. . . since academic qualifications are a weak currency and possess all their value only within the limits of the academic market.” (Bourdieu, 1977a, p. 507) For cultural capital to be an important mechanism of social reproduction it must be the case, not just that cultural capital facilitates the acquisition of educational credentials, but that educational credentials are an important mechanism by which wealth and power are transmitted. Bourdieu focuses on the first of these relationships at the expense of the latter, and this may account for the ambiguity in his views on the subject. But, on balance, I think that although Bourdieu often stresses the importance of cultural capital, the claim that he fails to recognise the importance of other forms of capital, and the ultimate primacy of economic capital, is unfair. In particular, Bourdieu clearly recognises that cultural capital forms a far more important part of the strategy of social reproduction for some sections of the higher classes (in particular, the professional classes), than for others (such as capitalists). It must be acknowledged, though, that his statements on the subject are inconsistent. The Cultural Arbitrary In addition to cultural capital, Bourdieu introduces the supplementary concept of the cultural arbitrary, which poses an additional obstacle to lower

CHAPTER 1. CULTURAL REPRODUCTION

6

class educational attainment. Bourdieu does not define the concept of the cultural arbitrary. However, he states that: “In any given social formation the legitimate PA 1 , i.e. the PA endowed with the dominant legitimacy, is nothing other than the arbitrary imposition of the dominant cultural arbitrary insofar as it is misrecognized in its objective truth as the dominant PA and the imposition of the dominant culture. . . ” (Bourdieu and Passeron, 1990, p. 22) Since Bourdieu uses the term “cultural arbitrary” without defining it, it is not clear precisely what he means when referring to arbitrariness, or to what extent he sees the cultural skills demanded and transmitted by the educational system as arbitrary. In some cases, the educational standards described by Bourdieu are clearly in some sense arbitrary. For instance Bourdieu claims that lower class students who achieve a degree of academic success by dint of hard work, face the obstacle that their achievement may be deemed to be too hard won, and not natural enough. In the educational system: “. . . application becomes pedantry and a respect for hard work grinding, limited pettiness, with the implication that it is intended to compensate for lack of natural talents.” (Bourdieu, 1974, p. 59) This aristocratic disdain for lower class attempts to appropriate higher class culture leads to a peculiar set of values in higher education. Namely: “. . . a tendency to prefer eloquence to truth, style to content.” (Bourdieu, 1967, p. 335) Bourdieu backs up this claim by reference to university examination reports (Bourdieu and Saint-Martin, 1974). Bourdieu claims that the criteria of university examiners reflect the values of the dominant classes, and that the more vague the demands of the examiners are, the less chance lower class pupils will have of adhering to these demands. 1

Pedagogic Action

CHAPTER 1. CULTURAL REPRODUCTION

7

Bourdieu’s comments on the theme of academic values are plausible. But one must ask how important the cultural arbitrary is in contributing to class inequalities in educational attainment. Although Bourdieu’s argument is rather compelling in relation to the evaluation of work in the arts and humanities departments of universities, it does not have the same force when applied to the sciences or to primary and secondary schools. At least Bourdieu provides no evidence that “eloquence” is preferred to “truth” in university physics exams. In addition, it seems to me that the National exams taken by school children (at least in this country) are examined using clear and explicit criteria. This problem reflects a general tendency of Bourdieu’s to focus on universities rather than on schools. This can only detract from his arguments since it means that Bourdieu is dealing with a population from which the lower classes have already been largely eliminated. I am not suggesting that Bourdieu sees every element of the cultural capital transmitted in the home and the education transmitted in the school as arbitrary. For instance Bourdieu notes that: “. . . logical and symbolic mastery of abstract operations and, more precisely, mastery of the laws of transformation of complex structures [are] a function of the type of practical mastery of language and the type of language acquired in the home. . . ” (Bourdieu and Passeron, 1990, p. 58) Presumably the skill of abstraction, and the complex language which it requires, is not “arbitrary” simply because it is associated with the higher classes. So, how do we decide which educational values and practices are arbitrary, and which valid? Bourdieu states that: “The sociological theory of PA distinguishes between the arbitrariness of the imposition and the arbitrariness of the content imposed, only so as to bring out the sociological implications of the relationship between two logical fictions, namely a pure power relationship as the objective truth of the imposition and a totally arbitrary culture as the objective truth of the meanings imposed. . . There is no PA which does not inculcate some meanings not deducible from a universal principle (logical reason or biological nature. . . )” (Bourdieu and Passeron, 1990, p. 9–10)

CHAPTER 1. CULTURAL REPRODUCTION

8

In relation to the content of teaching, “arbitrariness” is opposed to “objective truth” and “meanings deducible from a universal principle”. This is confusing, since “objective truth” and “meanings deducible from a universal principle” are not the same thing. If anything which is not objectively true is therefore “arbitrary”, does this mean that subjects (e.g. the humanities) where we can never be confident of objective truth are always utterly arbitrary? Or is the aim of truth enough to justify a discipline? What about subjects such as music or woodwork which do not aim at truth? Bourdieu’s notion of the “cultural arbitrary” is unclear. It is not possible to determine to what extent he is arguing that the dominant culture and the educational values that serve it are no better than any other culture. Is there an element of cultural relativism in his argument? It is very important to distinguish between the view that many lower class pupils may find highbrow literature (for instance) hard to understand because of the distance between ordinary language and literary language, and the view that highbrow literature is irrelevant to working class pupils because they come from a different culture. The latter view is explicitly based on the seemingly egalitarian notion that all cultures are ‘equal but different’, but implicitly, it is based on the insulting view that lower class children are inevitably incapable of appreciating any aspect of highbrow culture. Bourdieu does not make a clear enough distinction between those parts of the dominant culture which are in some way snobbish (i.e. exclusive for exclusivity’s sake) and arbitrary, and those which are universally valuable but not universally accessible. So, Bourdieu gives some interesting examples of arbitrary values in education, but does not give a precise definition of what constitutes arbitrariness in this context. Such a definition is essential if we are to distinguish between those elements of the dominant culture which should be taught in schools, and those which should be removed from the curriculum. Furthermore, it seems clear that lower class pupils would be disadvantaged by a lack of cultural resources even if the content of educational syllabuses and assessments were utterly rational. A sophisticated grasp of language alone would be a huge advantage in just about any conceivable educational system. Given this, the cultural arbitrary should surely be relegated to a minor role in any explanatory theory of class inequalities in educational attainment.

CHAPTER 1. CULTURAL REPRODUCTION

1.2.2

9

Habitus

Introduction to Habitus The notion of habitus is central to Bourdieu’s thought, yet it is never clearly defined. I will try to elucidate the concept, before going on to criticise it. Like cultural capital, habitus is transmitted within the home. However, whereas cultural capital consists of the possession of legitimate knowledge, habitus is a set of attitudes and values, and the dominant habitus is a set of attitudes and values held by the higher classes. “In fact, each family transmits to its children, indirectly rather than directly, a certain cultural capital and a certain ethos2 . The latter is a system of implicit and deeply interiorized values which, among other things, helps to define attitudes towards the cultural capital and educational institutions. The cultural heritage, which differs from both points of view according to social class, is the cause of the initial inequality of children when faced with examinations and tests, and hence of unequal achievement.” (Bourdieu, 1974, p. 43) So, habitus works with cultural capital to the disadvantage of lower class students. A major components of the dominant habitus is a positive attitude towards education. “. . . the system of dispositions towards the school, understood as a propensity to consent to the investments in time, effort and money necessary to conserve and to increase cultural capital.” (Bourdieu, 1977a, p. 495) So is habitus just a set of attitudes, directed primarily towards education and culture? Sometimes Bourdieu seems to suggest that the dominant habitus consists of more than this — that it includes (or at least gives rise to) competence in specific social settings, including for instance: “. . . the practice of the games and sports of high society or the manners and tastes resulting from good breeding. . . ” (Bourdieu, 1977a, p. 506) 2

Ethos is an element of habitus.

CHAPTER 1. CULTURAL REPRODUCTION

10

So what does Bourdieu mean by a “set of dispositions”? His definition in (Bourdieu, 1977b, p. 214) is unhelpful. He states that the word disposition encompasses three meanings: 1. “the result of organising an action”. 2. a “way of being” or “habitual state”. 3. a “tendency”, “propensity” or “inclination”. If Bourdieu insists on using the word habitus to signify three such different things, it would be helpful if some attempt was made to indicate to the reader which sense is intended each time the term is used. Needless to say, Bourdieu does not do this. No wonder then that the concept of habitus is condemned by Nash as “ambiguous and overloaded” (Nash, 1990, p. 446). Unlike cultural capital, there has rarely been any attempt by empirical researchers to use the concept of habitus. This is because the concept is too nebulous to be operationalised.

Structure and Agency: the Role of Habitus Given the messiness of the concept of habitus, one might ask why Bourdieu introduces it into his theory at all. The answer is that Bourdieu thinks that the concept of habitus solves a fundamental problem in sociology — the conflict between structure and agency. Bourdieu attacks crude structuralism on the grounds that “certain structuralists” see “agents as the simple ‘supports’ of structures invested with the mysterious power of determining other structures.” (Bourdieu, 1977a, p. 487) However, Bourdieu also criticises methodological individualism. Certain “atomistic” mobility researchers are singled out for attack on the grounds that they do not recognise that social mobility can coexist with stable class structures. Bourdieu protects himself here by failing to name the researchers he is referring to, and I cannot think of any mobility researcher who fails to recognise this simple point. In fact, Bourdieu’s comment reveals a crude view of methodological individualism (and perhaps also of structuralism) neither of which he defines, of course, despite the fact that they mean so many different things to different people.

CHAPTER 1. CULTURAL REPRODUCTION

11

According to Bourdieu, if we wish to avoid the dichotomy between individualism and structuralism: “This means that our object becomes the production of the habitus, that system of dispositions which acts as mediation between structures and practice; more specifically, it becomes necessary to study the laws that determine the tendency of structures to reproduce themselves by producing agents endowed with the system of predispositions which is capable of engendering practices adapted to the structures and thereby contributing to the reproduction of the structures” (Bourdieu, 1977a, p. 487) Following (though not acknowledging) Boudon, Bourdieu notes that working class students are more likely to drop out of the educational system than higher class students, even if we control for previous achievement. He claims that this is a more important mechanism of selection than exam failure. “Thus, previous performances being equal, pupils of working-class origin are more likely to ‘eliminate themselves’ from secondary education by declining to enter it than to eliminate themselves once they have entered, and a fortiori more likely not to enter than to be eliminated from it by the explicit sanction of examination failure.” (Bourdieu and Passeron, 1990, p. 153) Bourdieu claims that this phenomenon can be explained in terms of the lower class habitus. The habitus is in some way formed by the objective chances of success shared by the class. The habitus in turn determines the actions of the members of the class. “. . . the principle underlying the production of the most durable academic and social differences, the habitus — the generative unifying principle of conducts and opinions which is also their explanatory principle, since at every moment of an educational or intellectual biography it tends to reproduce the system of objective conditions of which it is the product.” (Bourdieu and Passeron, 1990, p. 161) But if attitudes are due to a particular habitus, how does this habitus form, and how is it absorbed by individuals? Bourdieu is characteristically vague:

CHAPTER 1. CULTURAL REPRODUCTION

12

“. . . everything happens as if parental attitudes towards their children’s education. . . were primarily the interiorization of the fate objectively allotted (and statistically quantifiable) as a whole to the social category to which they belong.” (Bourdieu, 1974, p. 33) Again we have the non-committal phrase “everything happens as if”. Bourdieu is not suggesting that parents consult statistics on class differences in survival rates in education before making decisions on their children’s education. So how exactly is their “fate” “interiorized”? The process becomes more mysterious the deeper one probes: “. . . the negative predispositions towards the school which result in the self-elimination of most children from the most culturally unfavoured classes and sections of a class — such as selfdepreciation, devaluation of the school and its sanctions, or a resigned attitude to failure and exclusion — must be understood as an anticipation, based upon the unconscious estimation of the objective probabilities of success possessed by the whole category, of the sanctions objectively reserved by the school for those classes or sections of a class deprived of cultural capital.” (Bourdieu, 1977a, p. 495) The objections to this are obvious. Firstly, How can an estimation be unconscious? If habitus is not generated by conscious individuals, where does it come from? Secondly, even if an individual knows the objective probabilities of success possessed by the whole category, why does she not recognise that, by changing her attitude to the educational system, she can escape the fate of the rest of her category? Thirdly, Bourdieu seems to be arguing that people’s behaviour is the result of accepting the “objective probabilities” of future success. However, as Jenkins points out, “Something which happens at time ‘x’ cannot be accounted for by the likely state of affairs — as predicted by statistics — at the time ‘x + 1’.” (Jenkins, 1992, p. 81) Expectations about the future must be based upon the present. The actions based on these expectations create social reality, rather than “objective probabilities” creating expectations which lead to action. It might be argued that it is uncharitable to interpret Bourdieu as putting forward an explanation of current events in terms of future events. But even if we interpret Bourdieu more kindly as arguing that lower class pupils do

CHAPTER 1. CULTURAL REPRODUCTION

13

not pursue demanding educational options because they are aware of the current tendency of the class as a whole not to pursue such options, it must be admitted that this is a feeble explanation. If we were happy to accept explanations of the characteristics of individuals which simply refer us to the characteristics of the group of which these individuals are members, without explaining these characteristics, there would be little need for sociology. In sum, the notion of habitus utterly fails in Bourdieu’s stated purpose of avoiding both structuralist determinism and “atomism”. The notion of habitus is completely deterministic leaving no place for individual agency or even individual consciousness. Yet Bourdieu denies the charge of determinism on three counts. Firstly, the same habitus will produce different practices in different social fields. Secondly, the habitus can be changed by changed circumstances. Thirdly, the habitus may be controlled by the “awakening of consciousness and socioanalysis”(Bourdieu, 1990, p. 116). The first two points depend on forces external to the social group and are therefore no defence against determinism. The speciousness of these arguments can be illustrated by the fact that the same charge will produce different motion in different electric fields - which hardly shows that electromagnetism has a role for individual freedom. The third point is rather vague - Bourdieu does not suggest how we are supposed to awaken our (presumably slumbering) consciousnesses. So, habitus is a vague concept, which fails to fulfil the role that Bourdieu claims for it within his theory. Bourdieu may be right that the attitudes towards education associated with different social classes affect the tendency of individuals within those classes to pursue education. But this is barely worth saying if we cannot be any more precise. So the concept of habitus adds nothing to Bourdieu’s analysis of class inequalities in educational attainment. As for the claim that the concept of habitus transcends the dichotomy between structure and agency, this is totally unjustified.

1.2.3

Political Action

A criticism which is linked to Bourdieu’s determinism is the view that Bourdieu does not allow for working class cultural production, and hence has nothing to say about a radical politics of education. For instance, Giroux

CHAPTER 1. CULTURAL REPRODUCTION

14

states that Bourdieu’s theory: “. . . is a theory of reproduction that displays no faith in subordinate classes and groups, no hope in their ability or willingness to reinvent and reconstruct the conditions under which they work and learn.” (Giroux, 1982, ) One could simply retort that sociology should never be driven by a demand for “faith” and “hope” — that is the role of dubious metaphysics. However, it is fair to ask whether Bourdieu suggests any way of improving the education system. In fact, Bourdieu does propose an education system which would do justice to lower class pupils — namely “universal pedagogy”. This form of pedagogy: “. . . would take nothing for granted initially, would not count as acquired what some, and only some of the pupils in question had inherited, would do all things for all and would be organised with the explicit aim of providing for all with the means of acquiring that which, although apparently a natural gift, is only given to the children of the educated classes. . . ” (Bourdieu, 1974, p. 38) This means that cultural capital would be taught explicitly in the school, rather than simply transmitted within the higher class home. However, Bourdieu does not suggest that “universal pedagogy” could be implemented now, since: “. . . only a school system serving another system of external functions and, correlatively, another state of the balance of power between the classes, could make such pedagogic action possible” (Bourdieu and Passeron, 1990, p. 127) In my view Bourdieu is right to argue that the education system must be seen in terms of a wider system of social relations, and that changes to the education system alone cannot transform those relations. However, while Giroux exaggerates the autonomy possessed by the educational system, Bourdieu perhaps underplays it. Although Bourdieu is right to argue that the possibilities for educational reform, and for the effectiveness of such reforms are

CHAPTER 1. CULTURAL REPRODUCTION

15

limited by the balance of power between the classes, this does not mean that the idea of universal pedagogy has no implications for educational reformers. If the lack of cultural capital is an important cause of lower class educational disadvantage, then this should be taken on board by schools. In particular, schools should put more emphasis on the explicit teaching, not just of basic literacy, but of a sophisticated grasp of language. If Bourdieu’s theory were correct, this would certainly increase the absolute level of attainment of lower class pupils (a worthwhile aim in itself) and might well have some success in eroding class inequalities in attainment.

1.2.4

Science and Language

I have pointed out that Bourdieu fails to express his theory clearly. In my view, this failure is bound up with Bourdieu’s rejection of what he describes as a “. . . positivist conception of science. . . ”(Bourdieu, 1990, p. 19–20). Of course, Bourdieu does not define what he means by positivism. Instead, he uses the common ploy of denouncing all research that attempts to test hypotheses empirically as positivist without actually saying what he thinks is wrong with this type of methodology. The rejection of the importance of deriving hypotheses from a theory and attempting to test these hypotheses allows Bourdieu to be unapologetic in the use of poorly defined concepts. “Especially in the Anglo–Saxon tradition, people criticise the researcher for using concepts that function as signposts pointing to phenomena that are worth examining but that often remain obscure and vague, even if they are suggestive and evocative.” (Bourdieu, 1990, p. 40) Against such Anglo–Saxon criticisms, Bourdieu asserts that, because the social world is complex, theories about it must be complicated, and must be expressed in complicated language. “I think that, literary and stylistic qualities apart, what Spitzer says about Proust’s style is something I could say about my own writing. He says, firstly, that what is complex can only be said in a complex way; secondly, that reality is not only complex, but also structured. . . if you want to hold the world in all its complexity and at the same time order and articulate it. . . you have to use

CHAPTER 1. CULTURAL REPRODUCTION

16

heavily articulated sentences that can be practically reconstructed like Latin sentences. . . ” (Bourdieu, 1990, p. 51–52) The absurdity of this argument is easily shown. Firstly, the aim of science is not to “hold the world in all its complexity”, and the history of science tells us that a simple theory will be preferred to a more complicated theory if the simpler theory has equal or superior predictive power, e.g. Copernicus’ defeat of Ptolemaic astronomy. Furthermore, it simply is not true that a difficult concept or theory must be expressed in difficult language. In my view, the real purpose served by the obscurity of Bourdieu’s prose is to protect his own work from refutation. Bourdieu’s strategy in dealing with criticism is to claim that his critics have not understood his work, and to imply that his critics are just jealous because they are not as clever as him. “. . . they criticise not my analyses, but an already simplified, if not maimed, representation of my analyses. This is because they invariably apply to them the very modes of thought, and especially distinctions, alternatives and oppositions, which my analyses are aimed at destroying and overcoming.” (Bourdieu, 1990, p. 107) The point that the critic may not agree that Bourdieu has succeeded in destroying such oppositions is ignored by Bourdieu, who never deals with specific criticisms in a direct way. When a criticism is made of Bourdieu, the explanation for this is always to be found in the inadequacies of the critic. So, behind “positivist methodology” lies an “epistemology of resentment” which allows its advocates to “prohibit others from doing what they themselves are unable to do, so that they can impose their own limits on others.”(Bourdieu, 1990, p. 35) “Positivism” is simply “. . . a tradition often appealed to by the most mediocre of researchers in order to “pare the lion cubs’ claws”, as Plato put it — in other words, to disparage and reduce the creations and innovations of the scientific imagination.” (Bourdieu, 1990, p. 40) So, although Bourdieu declares a “headlong, rather crazy commitment to science” (Bourdieu, 1990, p. 26), his rejection of scientific values is made plain. Furthermore, his impenetrable prose style should not be seen simply as an irritation for the reader, but rather as being closely bound up with this rejection of scientific values, since clarity makes a theory amenable to testing, whereas obscurity protects it from falsification.

CHAPTER 1. CULTURAL REPRODUCTION

1.3

17

Empirical Evidence on Cultural Capital

This section consists of a survey of some of the empirical work relating to Bourdieu’s theory. Most of this work focuses on the link between cultural capital and educational attainment. I will conclude that the evidence is mixed, largely due to widely varying operationalisations of cultural capital. Evidence on the link between educational attainment and social reproduction and mobility will also be examined. There has been less focus on this part of Bourdieu’s theory, but such evidence as there is suggests that educational capital is more a vehicle of social mobility than of social reproduction.

1.3.1

Bourdieu’s Own Evidence

Bourdieu is adamant that he does not engage in theory for its own sake, and that empirical work is central to his enterprise. “Let me say outright and very forcefully that I never ‘theorise’, if by that we mean engage in the kind of conceptual gobbledegook. . . that is good for textbooks and which, through an extraordinary misconstrual of the logic of science, passes for Theory in much of Anglo-American social science. . . There is no doubt a theory in my work, or, better, a set of thinking tools visible through the results they yield, but it is not built as such. . . It is a temporary construct which takes shape for and by empirical work.” (Waquant, 1989, p. 50) Unfortunately, the claim that Bourdieu’s theoretical framework is subordinate to the needs of empirical research is not backed by the evidence provided by Bourdieu regarding cultural reproduction. For Bourdieu’s theory to be backed empirically, he would need to show that: 1. Parental cultural capital is inherited by children. 2. Children’s cultural capital is converted into educational credentials. 3. Educational credentials are a major mechanism of social reproduction in advanced capitalist societies.

CHAPTER 1. CULTURAL REPRODUCTION

18

Of course, Bourdieu does not deny that privilege can be inherited through means other than the acquisition of educational credentials. Inheritance of property, and occupational advantage gained through social networks are obvious examples of this. So, Bourdieu’s theory is not refuted by empirical evidence that there is no absolute correspondence between credentials and occupational outcomes (see for instance Dale and Pires 1984). However, it is crucial to Bourdieu’s theory that cultural capital actually does facilitate educational success, and that educational success actually is associated with occupational advantage, even if this is only a means of legitimating class inequalities. Bourdieu claims that (1) and (2) are shown: “. . . by the fact that, among the pupils of the grandes ´ecoles, a very pronounced correlation may be observed between academic success and the family’s cultural capital measured by the academic level of the forbears over two generations on both sides of the family. . . ” (Bourdieu, 1977a, p. 497) Bourdieu is not entitled to assume that a high parental level of education reveals a high level of parental cultural capital. In fact, Bourdieu’s use of parental educational credentials as a measure of cultural capital begs the question of whether educational credentials simply constitute “. . . embodied cultural capital that has received school sanctioning.” (Bourdieu and Boltanski, 1981, p. 145). In addition, the use of bivariate analyses is crude. Clearly, a simple association between two variables is not convincing evidence of a causal relationship. Bourdieu fails to show that parental cultural capital is inherited by the children, and that this is the mechanism through which higher class pupils tend to attain higher educational credentials than lower class pupils. His evidence is quite consistent with educational privilege being passed down through mechanisms other than cultural capital, such as parental encouragement and material resources. Bourdieu also presents evidence that both social class and educational attainment are strongly associated with participation in cultural activities such as book reading and buying, and cinema, theatre, concert and museum attendance. (Bourdieu and Boltanski, 1981, p. 490–2). However, on their own, these figures do not really back up Bourdieu’s theory. They do not constitute evidence that participation in cultural activities is the mechanism by which middle class parents ensure good qualifications for their children.

CHAPTER 1. CULTURAL REPRODUCTION

19

In sum, Bourdieu assumes much of what he sets out to prove. It is circular to treat educational level as a proxy for cultural capital if one is trying to assess whether cultural capital does in fact help to determine the educational levels reached by individuals.

1.3.2

Other Research on Cultural Reproduction

Bourdieu’s concept of cultural capital is not clearly defined, and it is not particularly surprising that it has been operationalised in various different ways. I have argued that Bourdieu’s own operationalisation of the concept is quite inadequate. Yet Bourdieu is not the only author to use parental education as a proxy for cultural capital. For instance, Halsey, Heath and Ridge (Halsey et al., 1980) use this proxy, as do Robinson and Garnier (Robinson and Garnier, 1985) and Jan o. Jonsson (Jonsson, 1987). Since Bourdieu’s definition of cultural capital is not precise, it is not clear what an “authentic” operationalisation would consist of. However, Bourdieu does explicitly state the importance of linguistic competence. Cultural “competence” and “familiarity” can reasonably be interpreted as knowledge of and participation in the dominant culture. Despite this, previous investigations of cultural capital have not included data on linguistic ability, and DiMaggio and Mohr (DiMaggio, 1982), (DiMaggio and Mohr, 1985) and (Mohr and DiMaggio, 1995) are unusual in using data on cultural knowledge. Data on cultural activities other than reading has often tended towards highly exclusive activities such as gallery attendance, which are foreign to a large proportion even of the middle and upper classes. For example, De Graaf (De Graaf, 1986) uses a measure of the number of visits per month to museums, galleries, concerts, theatres and historical buildings. In general, surveys include data on either pupils’ or parents’ cultural participation, but not both. Most commonly, the proxy of parental education is used instead of data on parental cultural capital, although this proxy clearly begs the question of whether occupational status and educational attainment actually do reflect the possession of culturalcapital. Given that researchers have operationalised the concept of cultural capital in different ways, it is not surprising that empirical studies of the effect of cultural capital on educational attainment have varied in their conclusions. As well as those already mentioned, note (Crook, 1997), (Egerton, 1997), (Graetz, 1988), (Kalmijn and Kraaykamp, 1996), (Katsillis and Rubinson, 1990) and (Savage and Egerton, 1997). Which cultural attributes should be seen as constituting capital cannot be

CHAPTER 1. CULTURAL REPRODUCTION

20

determined without empirical investigation, since the term cultural capital implies an analogy with economic capital, and therefore, a return. The return on cultural capital takes the form of educational credentials and, ultimately, occupational success. Therefore, in subsequent chapters, I will develop a broad operationalisation of cultural capital in order to examine which elements actually yield returns in the sense of contributing to educational success. If participation in cultural activities does lead to academic success, one may ask why this should be. It may be suggested that the culture of the school reflects the dominant culture. This could occur if teachers are prejudiced in favour of pupils who display “cultured” traits, and therefore give them higher grades (Farkas et al., 1990). This view is perhaps most relevant in the US, where grades awarded by teachers are an important outcome of schooling. It is a less plausible explanation in nations such as Britain, where the key outcome of schooling is the results gained in national examinations. Alternatively, the dominant culture could be ingrained in the curriculum. However, it has been pointed out that, although this may be true of France, there is little emphasis on highbrow culture in schools in countries such as Britain, the Netherlands, and the US (De Graaf et al., 2000). An alternative explanation is that participation in cultural activities leads to the development of knowledge or skills, which in turn enable pupils to succeed at school. For instance, one might expect reading novels to contribute to both linguistic competence and cultural knowledge. Crook (Crook, 1997) and de Graaf et. al. (De Graaf et al., 2000) follow de Graaf (De Graaf, 1986; De Graaf, 1988) in breaking cultural capital into two constituent parts, reading and beaux arts participation. Beaux arts participation refers to participation in formal cultural activities outside the home, such as gallery, theatre and concert attendance. Both Crook (Crook, 1997) and de Graaf et. al. (De Graaf et al., 2000) find that reading is associated with academic success whereas beaux arts participation is not, and infer from this that the effect of cultural capital on educational attainment is due to the “educative resources” such as analytic and cognitive skills which are developed by reading, rather than to the communication of status via participation in formal culture. However, this inference may be questioned, since one could argue that participation in beaux arts may contribute to the development of skills and knowledge, or that pupils’ reading is as likely to prejudice teachers in their favour as is participation in other cultural activities. Therefore, as a further test of this hypothesis, it will be useful to test pupils on the sorts of abilities

CHAPTER 1. CULTURAL REPRODUCTION

21

and knowledge that may be developed through cultural participation, in order to see whether these skills are in fact the means through which cultural participation promotes educational success. In sum, many researchers examining cultural capital have used what data was available to them, even though this data has not been ideally suited to the purpose. In my view, it is far preferable to begin with an exploration of the theory of cultural reproduction, and of the mechanisms through which cultural capital may operate, in order to develop a sound operationalisation of the concept of cultural capital, and then to collect appropriate data.

1.4

Conclusions

Bourdieu’s project is extremely ambitious, and I have argued that, on many scores, it is a failure. I have stated that habitus is an unhelpful concept. It has some intuitive plausibility, but is at once too all-inclusive and too vacuous to be of any use to empirical researchers. I have also argued that Bourdieu’s claim that the notion of habitus solves the conflict between structure and determinism on the one hand and agency and individualism on the other is quite unjustified. In fact Bourdieu’s theory has no place not only for individual agency, but even for individual consciousness. Even in the area where I have been most sympathetic, namely the concept of cultural capital, I have argued that the concept is not clearly defined. In addition, I argued that the related concept of the cultural arbitrary is of very limited usefulness. Bourdieu has some valuable insights into arbitrary practices in higher education. However, he does not distinguish clearly enough between standards which are prejudicial to lower class pupils and students because they are arbitrary, and standards which are prejudicial to lower class pupils and students because they do not have the resources to meet those standards. Bourdieu’s theory is vaguely expressed throughout. One might even accuse him of what he accuses undergraduates — a tendency to “. . . minimise the risks by throwing a smoke-screen of vagueness over the possibility of truth or error.” However, I have defended Bourdieu against what I took to be unfair criticisms. Bourdieu has been accused of emphasising the importance of cultural capital at the expense of economic capital. Although it is easy to see how

CHAPTER 1. CULTURAL REPRODUCTION

22

one would get this impression from some passages in Bourdieu, on balance I think that Bourdieu is aware of the prime importance of economic capital Bourdieu has also been accused of not having anything to say about a radical politics of education. This is true in that Bourdieu does not see the education system as independent from the rest of society, and therefore sees the transformative power of radical educational policies as rather limited. However, I believe that Bourdieu is, to some extent, correct in his pessimism. The education system has a degree of autonomy, but critics of Bourdieu such as Giroux exaggerate this autonomy somewhat naively. However, Bourdieu may be accused of understating the autonomy of the educational system, and of ignoring the policy implications of the theory of cultural capital. But my main defence of Bourdieu is that despite the huge flaws in the grand structure of his theory, some of his ideas have the potential to be very fruitful. In particular, the concept of cultural capital is one that has inspired empirical researchers. However, many of the empirical findings on cultural capital contradict one another. This may be partly due to the fact that these studies were carried out at different times in different countries. It may be that cultural capital is more important in some countries than in others, or operates differently in different countries at different times. However, in my view the main reason for the variable findings presented here is the different methodologies used in each study, and in particular, the array of different operationalisations of cultural capital that are used. Bourdieu himself uses parental educational attainment as a measure of cultural capital. In my view this is a crude indicator of the concept, and little improvement is made on it by most of the studies cited here. Bourdieu describes cultural capital as a set of activities and competencies associated with the higher classes. From the examples Bourdieu gives, these seem to include participation in high culture, but also a sophisticated grasp of language, and specific cultural skills and knowledge. In order to test the link between cultural capital and academic attainment thoroughly, these different elements of cultural capital should be distinguished, in order to find which (if any) elements of cultural capital are important. The link between educational attainment and occupation is not examined empirically by Bourdieu, and he is not consistent in the strength of the role he wishes to claim for educational credentials in maintaining social reproduction. Given the evidence showing that the association between class origins and destinations persists even when educational credentials are controlled for (Marshall et al., 1997), an overriding role should not be claimed for ed-

CHAPTER 1. CULTURAL REPRODUCTION

23

ucational credentials. Evidence from France (Robinson and Garnier, 1985) suggests that educational capital is a mechanism for social mobility rather than social reproduction. Therefore, Bourdieu may have overstated the role of educational credentials in social reproduction, at least in his stronger statements on this subject. The work that I will go on to describe in subsequent chapters uses measures of as broad a range of cultural resources associated with the higher class home as possible. I will use measures of both parents’ and pupils’ cultural participation, and of pupils’ linguistic ability and cultural knowledge in order to examine whether any element of cultural capital is important in the transmission of educational advantage.

Chapter 2 Rational Choice Perspectives Rational choice theory aims to explain human behaviour on the basis that people act rationally to achieve their aims in the light of their beliefs about the situation. This of course leaves much scope for disagreement among rational choice theorists. In this chapter, I will examine the application of rational choice theory to the question of inequalities (it must be acknowledged that the inequalities addressed are largely class inequalities) in educational attainment, with reference to the work of Murphy, Boudon, Gambetta and Breen and Goldthorpe. Breen and Goldthorpe (Breen and Goldthorpe, 1997) give a formalisation of their model. This technique is useful in highlighting the key variables and their effects. I will describe a generic “utility” model, which I will argue covers the models of decision making about post compulsory education of all the authors mentioned here (with the partial exception of Murphy). I will go on to discuss Breen and Goldthorpe’s formalisation in detail. I will argue that Breen and Goldthorpe fail in their stated aim of giving an explanation that does not depend on social class differences in preferences, although they give a common mechanism for the formation of such preferences.

2.1

The Utility Model

Breen and Goldthorpe give a formalisation of their model. I will now describe a basic “utility” model, which I will use to analyse the work of all the authors mentioned here, and in particular, to discuss Breen and Goldthorpe’s 24

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

25

formalisation in detail. Let me stress that this model is only intended as an aid to examining other models of educational decision making, not as an explanatory model in its own right. The utility model is a model of the decision to stay on in education after having already achieved a certain level of education. For the sake of argument let us say it is the decision to stay on in the sixth form having passed GCSE’s. Suppose that each individual attaches a “utility” to each of the following possible educational outcomes: 1. Staying on in education and “passing” (i.e. achieving some minimum level chosen by the researcher, not the individual). 2. Staying on in education and “failing”. 3. Leaving education. By utility I mean the value of the educational outcome in life after education (in the job market — but also possibly in the marriage market or for other social or individual purposes). The utility of the outcome should be contrasted with the “cost” of education, by which I mean costs incurred in the course of trying to gain the qualification, set off against any benefits of doing the course (as opposed to benefits of success in gaining the qualification), for example, direct financial costs, earnings forgone, like or dislike of the course, approbation/disapproval of peers and family etc.). Let us denote the utilities of each of the possible outcomes by U1 , U2 and U3 respectively, and denote the cost of education by C. Suppose that the individual believes that they would have probability π of passing were they to stay on in education. This can be represented diagrammatically as shown in figure 2.1. It should be noted that these utilities, costs and probabilities are subjective — i.e. they are those perceived by the individual. It is reasonable to say therefore that the expected net utility (utility − costs) if an individual chooses to stay on is given by: (expected utility if they pass)(probability of them passing)+ (expected utility if they fail)(probability of them failing)−(costs of education)

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

26

Pass .r ..... U .......... 1 .......... ..........

...... .......... .......... .......... .......... . . . . . . . . . .... .......... ........... .......... ................... .......... .......... .......... .......... . . . . . . . . . .......... .. .......... .......... .......... .......... . . . . . . . . .......... . ....... . . .......... . . . . . . . .......... ....... . . . . . . .......... . . . . ............... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... ......

Stay on

r

r

Fail

r U2

rU

Leave

3

Figure 2.1: A generic utility model

which in turn is given by: π(U1 ) + (1 − π)(U2 )C = π(U1 − U2 ) + U2 − C whereas the expected net utility if they choose to leave is given by U3 . Thus if one introduces a parameter p given by p = (expected utility of staying on) − (expected utility of leaving) we can say that: p = π(U1 − U2 ) + U2 − C − U3 . p is interpreted as follows: if p is greater than zero, the individual expects more net utility from staying on than from leaving education; if p is less than zero, the reverse is true. Note that this is as far as one should attribute any meaning to p. It is only its difference from zero that is relevant and not its size. Thus we could equally well have defined p by: (expected utility of staying on) expected utility of staying on) + (expected utility of leaving) and then interpreted p as follows: if p is greater than one half, the individual expects more net utility from staying on than from leaving; if p is less than one half, the reverse is true. Of course, if one uses this alternative definition of p the results obtained will be the same — however, the formulae will

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

27

be superficially more complex. Therefore, I would like to use the original formula for p, but in deriving Breen and Goldthorpe’s equations I must use the second formula. Nevertheless, the differences are purely superficial. 1 To complete the model one now assumes that if an individual has p greater than 0 they will stay in education and if p is less than zero they will leave education. As it stands, this model, of course, makes no predictions. One would have to describe how the various parameters vary across class, gender etc. in order to make predictions. However, one can observe how these variables must vary in order to account for social inequalities. If we make the reasonable assumption that for all individuals U1 > U2 (i.e. that there is greater utility in passing than failing) then it must be true that for some individuals U2 − C − U3 is negative. Otherwise p will always be positive and so it will be in no-one’s interest to leave education. In words, if one refers to the difference between the utility of staying on and failing and the utility of leaving as the guaranteed utility of education then it must be the case for some people that the costs of education are greater than the guaranteed utility of education. (Of course, if the costs of education are positive and the guaranteed utility of education is negative then this will automatically be the case.) One can therefore describe the primary problem for rational choice theorists as explaining why the costs of education outweigh the guaranteed utility of education for working class pupils more often than is the case for middle class pupils. (Note however, that although the above condition is necessary, it is not sufficient to explain class differentials. It would be possible for the expectation 1

The name p for this variable is best thought of as standing for “parameter”. I call it p because Breen and Goldthorpe do. However, with their definition, p always takes values between zero and one. Given, on top of this, its name it is tempting to think of p as representing some probability — perhaps the probability of the student staying on in education. However, there is no justification whatsoever for thinking that this arbitrary formula represents any probability at all. Similarly, the fact that one student may have a larger value of p than another student does not even in imply that it is more in the interest of the first student to stay on. At times in their paper, Breen and Goldthorpe make both of these misinterpretations. For example, they prove that, on the basis of their model, p is always greater for service class than working class students and deduce that “children from middle class backgrounds will more strongly prefer (in the sense of perceiving it to be in their best interests) to remain in school to a further level of education rather than leave” (Breen and Goldthorpe, 1997, p. 12). Given the arbitrary nature of the formula for p such a deduction is totally unjustified. Similarly Breen and Goldthorpe’s argument on pages 287–289 rests on the implicit assumption that p is a probability.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

28

of success in exams to be high enough for working class pupils to compensate for a difference between the costs and guaranteed utility of education.)

2.2

Murphy

Murphy takes issue with the view that class differentials in educational attainment have anything to do with inequality of opportunity (Murphy, 1981; Murphy, 1990). Murphy states that the equation of class differentials in educational attainment with class inequality is due to a failure to take class differences in educational aspiration seriously. In Murphy’s view, working class youth simply demand lower levels of education than do middle class youth. This difference in the levels of demand for education is not due to inequality of opportunity, but is simply a matter of taste — i.e. preferences differ by social class. Murphy claims that the refusal to acknowledge this cultural difference between the social classes is a form of liberal ethnocentrism. It is clear how a rational choice approach could lead to such a view. Agents act to maximise the fulfilment of their preferences, given their beliefs about the situation. Class differentials in educational attainment have not narrowed significantly, despite the removal of obstacles that stood in the way of working class pupils in the past (for instance, through the provision of universal free education, compulsory up to age 16). Therefore, it might be argued, working class pupils must simply prefer lower levels of educational attainment to those demanded by middle class pupils. However, this version of the rational choice approach leaves several issues unexamined. Firstly, to what extent are class differentials in attainment actually due to decisions made by pupils, and to what extent do unequal resources (cultural or economic) determine educational ability? It seems implausible that class differences in initial attainment (at GCSE for instance) are due to rational action by pupils. Of course Murphy could claim that even young children make a decision on whether to work hard at school or not, and that the primary effects of stratification are due to working class children making a rational choice not to work hard at school, since they dislike schoolwork. However, in my view, the younger the children we are considering, the more dubious it is to present their actions in this way. The problems of a lack of information on the future benefits of education, and of weakness of will when

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

29

immediate effort is demanded for a distant payoff, are compounded massively in the case of five year old ‘decision-makers’. Secondly, what causes the class differential in demand for education? Murphy claims that working class pupils demand less education than middle class pupils because they simply dislike the educational process more than middle class pupils. But there are other possible explanations for the difference in the amount of education demanded by pupils from different social classes, even at the same level of demonstrated ability. Working class pupils may have lower occupational aspirations than middle class pupils do, and may therefore feel less need for qualifications. Another possibility is that working class pupils may have a lower level of belief in the instrumental value of education in attaining a desired occupation than middle class pupils do, or they may evaluate their chances of success in further and higher education modestly as compared to middle class pupils. Murphy does not give any reason for preferring his explanation to these alternatives.

2.2.1

Formalising Murphy

It should be noted that unlike other rational choice theorists of education, Murphy does not use Boudon’s distinction between the primary and secondary effects of stratification. Murphy sees the effort made by pupils during their school careers as a result of decisions made on the basis of preferences. This contrasts with the more usual approach taking tested ability or exam results at the first key educational “branching point” as a given primary effect. For Murphy, ability is a function of effort, and effort is a function of social class. Working class pupils make less effort in school simply because they dislike education. Thus they see the costs of education (for Murphy, “costs” always include psychic costs) as being higher than do middle class children — moreover they see the costs of education as a function of effort, i.e. working class pupils dislike making an effort in school more than middle class pupils do. Thus Murphy’s model of educational decision making pre-GCSE can be represented pictorially as shown in figure 2.2. π1 denotes the probability of passing if the pupil makes an effort, π2 denotes the probability of passing if the pupil does not make an effort. U1 denotes the utility of passing, U2 the utility of failing, and C is the cost of making an effort. One can read off from this that the expected net utility from making an effort

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

π1

30

Pass .r ..... U .......... 1 .......... ..........

...... .......... .......... .......... .......... . . . . . . . . . .. .......... .......... ...... .................... ...... .......... .......... ...... . . . . . .......... . .......... ...... .......... ...... . . . . .......... . .... . .......... . . . . .......... .... . . . .......... . . .... . . . . . . .... . . . . . .... . . . . . .... . . . . . .. ...... ...... ...... ...... ........... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... .... ...... .......... ...... .......... ...... .......... ...... .......... . . . . . ...... . . . . ...... ...... .......... ...... .......... ...... .......... ...... ...... .................... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... ......

Make an effort, cost C

r

1 − π1

Fail

r U2

r

π2

Pass r U1

r

Not make an effort

1 − π2

Fail

r U2

Figure 2.2: Murphy’s model

is given by: π1 U1 + (1 − π1 )U2 − C and that the expected net utility from not making an effort is given by: π2 U1 + (1 − π2 )U2 So the expected utility from making an effort minus the expected utility from not making an effort is given by: p = (π1 − π2 )(U1 − U2 ) − C This does not completely fit the generic model. Nevertheless, Murphy’s theory as applied to post GCSE decisions does fit the model. Murphy takes the values of the utilities to be independent of social class and explains the difference between the classes by 1. The difference in (psychic) costs. These differences in costs are a function of differences in preferences, which in turn are a function of class. Working class pupils find education more unpleasant than middle class pupils do.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

31

2. The difference in expected chances of success (π) post 16 is also a function of social class via ability. However, it should be borne in mind that for Murphy the social class difference in ability is an indirect consequence of the class difference in costs at the previous educational level. This is because working class pupils make less effort at the previous educational level.

2.3

Boudon

Boudon explains class differentials in educational attainment with reference to two mechanisms — the primary and secondary effects of stratification. The primary effects of stratification are cultural inequalities that determine the academic abilities of pupils. The secondary effects of stratification are the different costs and benefits that are associated with different educational decisions for pupils from different social classes. For Boudon, the benefits associated with each educational option vary with social class because ambition is relative to the social starting point of an individual. So, a working class child who wants to be a lawyer must be more ambitious than a middle class child who wants to be a lawyer. Therefore, high prestige educational options may be essential in avoiding social demotion for middle class pupils, whereas working class pupils can avoid social demotion without pursuing such options. This leads to middle class pupils being more likely to pursue such options than working class pupils at any given level of ability. For his argument to work, Boudon must make two assumptions. Boudon does not make these assumptions explicit, but Breen and Goldthorpe do. Firstly, we must assume that people’s priority is to avoid social demotion, rather than to pursue social mobility. Otherwise, prestigious educational options might be more attractive to students from working class backgrounds than to students from middle class backgrounds, since the social distance likely to be travelled as a result of successful completion of a prestigious course will be far greater for the working class student. Secondly, we must assume that failure in a high prestige option is believed to be more likely to lead to social demotion than not attempting to pursue such an option.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

32

Boudon also states that the social costs of taking an educational option may vary by social class. He gives one example of this. “Thus not choosing a prestigious curriculum may represent a high social cost for a youngster from a middle-class family if most of his friends have chosen it; but choosing the same course may represent a high cost for a lower-class youngster if most of his friends have not.” (Boudon, 1974, p. 30) This sort of explanation rests on an extreme and fallacious version of methodological individualism — the view that the explanation of a group’s behaviour is identical to the sum of the explanations of the behaviour of each individual in the group. In this case, this leads to the explanation of each child’s behaviour in terms of the behaviour of the rest of the group. This leaves us none the wiser as to why the working class child finds that all his friends are taking a low-prestige curriculum in the first place. (This is not to deny that there may be a contextual effect such that a critical mass of working class children taking a particular option may influence the remainder of the group in favour of taking that option.) More importantly, it is easy to come up with hypothetical costs and benefits which “explain” pupils’ decisions in terms of rational choice theory. But if we are to derive concrete hypotheses from rational choice theory, assumptions must be made about the beliefs and desires of agents, and these assumptions should be made explicit. Boudon’s failure to make his assumptions explicit seems to derive from a failure to realise that assumptions are necessary. This problem may stem from the version of rational choice theory that Boudon espouses in (Boudon, 1974), which is a “utilitarian” rather than a “desire-fulfilment” approach. “Now, the ‘rationality theory’ that was developed in chapter 2 is incompatible with the value theory. The latter assumes that people — at least some people — behave against their interests because of the values they are committed to, whereas the former assumes that people behave according to their interests in the sense that they attempt to maximise the utility of their decisions. Of course the ‘rationality factor’ and the ‘value factor’ cannot be added to each other, since they are contradictory.” (Boudon, 1974, p. 111) If we phrase rational choice theory in terms of the efficient fulfilment of desires, it is clear that any hypotheses we form must be based on assumptions about beliefs and desires. Boudon may think that in defining rational choice

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

33

in terms of utility maximisation he is avoiding this problem. In fact, it just means that we have to make assumptions about what makes people happy, and these assumptions will be, if anything, harder to back up than assumptions about what people want. Value driven (or wertrational) action only contradicts a form of rational choice theory that assumes that the desires acted on by an agent must be individualistic and hedonistic. So, if rational choice hypotheses must be based on some view of beliefs and desires, and if values are just a special kind of belief, then Boudon is wrong to claim that explanations of action in terms of the values of a group are incompatible with rational choice theory. Since Breen and Goldthorpe’s work is an extension and formalisation of Boudon’s, it is not necessary to formalise Boudon separately.

2.4

Gambetta

Gambetta (Gambetta, 1987) notes that the decisions of working class students are more sensitive to poor academic performance than those of middle class pupils. However, Gambetta disputes Boudon’s explanation of this phenomenon in terms of differing costs and benefits of education across social classes. Gambetta recognises that, although a working class child may find the decision to drop out of school early to be a low cost option, since it does not entail social demotion, whereas for a middle class child it would be a high cost option, since it would probably entail social demotion, correspondingly, the relative benefit to a working class pupil of taking a high prestige educational option should be much higher than it would be for a middle class pupil, since the social distance travelled is much further. Gambetta prefers an explanation in terms of class differences in norms and preferences for education. “. . . it could either be that relatively more subjects in the middle class feel a greater normative pressure to resist the temptation to abandon school after a failure or, on the other hand, that relatively more subjects in the working class do not attach as high a value to education.” (Gambetta, 1987, p. 173) Of course, class differences in norms and preferences are a possible alternative explanation to Boudon’s. However, Gambetta fails to provide evidence that

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

34

norms regarding education vary according to social class. Gambetta does not pursue a hard-line rational choice approach to the study of educational decision making. Instead, he is happy to borrow from structuralism and sub-intentional theories. In my view, although Gambetta states that he is aware of the dangers of an eclectic approach, which may “fail to offer simplifying insights”, (Gambetta, 1987, p. 167) this does not always stop him from falling into these pitfalls. Gambetta’s conclusion is that many mechanisms are at work. “Educational decisions are the result of three main processes: of what one can do, of what one wants to do and, indirectly, of the conditions that shape one’s preferences and intentions. They are the result partly of causality and partly of intentionality. It seems hardly the case that decisions are generated by either of these two forces alone.” (Gambetta, 1987, p. 169) This seems hardly a very interesting conclusion.

2.4.1

Formalising Gambetta

Gambetta can easily be fitted into the generic model. Although he describes many mechanisms which may cause class differentials, all of the mechanisms which can be seen as part of a rational choice approach relate primarily to the cost, C. Specifically in explaining why the costs exceed the guaranteed utility from education he emphasises social pressures which increase the cost of education for working class pupils. He does not take the approach of suggesting that U3 > U2 for working class pupils. In common with Boudon and Goldthorpe, Gambetta sees perceived chances of success, π, as a function of ability which in turn is a function of class. This also contributes to class differentials.

2.5

Breen and Goldthorpe’s Model

Breen and Goldthorpe’s stated aim is to explain the following empirical phenomena:

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

35

“(i) increasing educational participation rates; (ii) little change in class differentials in these rates; and (iii) a recent and very rapid erosion of gender differentials in educational attainment levels.” (Breen and Goldthorpe, 1997, p. 275) Breen and Goldthorpe aim to explain these empirical phenomena as the product of rational choices by families. They especially wish to avoid giving cultural or normative differences any place in this explanation. They construct their model formally in order that it should have testable implications. Breen and Goldthorpe follow Boudon’s (Boudon, 1974) distinction between primary and secondary effects of stratification, and concur in Boudon’s claim that it is the secondary effects (i.e. the different costs and benefits that are associated with different educational decisions for pupils from different social classes) that play the crucial role. They make explicit Boudon’s implicit assumptions that: 1. People’s priority is to avoid social demotion, rather than to pursue social mobility 2. Failure in a high prestige option is believed to be more likely to lead to social demotion than not attempting to pursue such an option. This second assumption means assuming either that it is in fact the case that failure in a high prestige option is more likely to lead to social demotion than not attempting such an option, or assuming that although this is not in fact the case, people mistakenly believe that it is. In the latter case, the mistaken belief will itself require explanation. The former assumption on the other hand is open to empirical attack. Payne (Payne, 1987) examines NCDS (National Child Development Survey) data on young people who reached 16 in 1974 and finds no disadvantage in staying on in education post-16, even to students who gain no further qualifications. “. . . those who left at 17 or 18 with qualifications no better than those of minimum age leavers suffered no long term disadvantage in comparison with the latter, despite their loss of potential work experience, and some groups had lower unemployment rates in the long term than minimum age leavers with equally good qualifications.” (Payne, 1987, p. 425)

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

36

In addition, there is evidence that an uncompleted college degree is not useless in the labour market (see Becker (Becker, 1964) and Layard and Psacharopoulos (Layard and Psacharopoulos, 1974)). Breen and Goldthorpe explicitly assume that values, norms and beliefs regarding education do not vary by social class, and that classes differ in terms of two factors only: 1. Average ability 2. Resources Breen and Goldthorpe focus on three factors that will be taken into account by parents and children. 1. The cost of remaining in school 2. The likelihood of success in continued education 3. The value that is attached to educational outcomes. This is determined by beliefs about the chances that each outcome gives of access to each social class. (Although these beliefs need not be accurate according to Breen and Goldthorpe, they must not be systematically distorted by social class.) Their model contains three social classes. The service class, which consists of professionals, administrators and managers; the working class; and the underclass, which consists of the unemployed and “those with only a precarious place in the labour market and in only the lowest grades of employment if not unemployed” (Breen and Goldthorpe, 1997, p. 281). Three mechanisms are postulated for the secondary effects of stratification. The first and most important of these is what Breen and Goldthorpe call ‘relative risk aversion’. Again, in line with Boudon, Breen and Goldthorpe state that aspirations must be seen as relative to the social starting point of an individual. As I stated earlier, Breen and Goldthorpe explicitly assume that the priority in all classes is avoiding downward social mobility. So middle class pupils will prefer prestigious educational options more strongly than working class pupils, due to the greater value middle class pupils attach to middle class occupational outcomes.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

37

The second mechanism is due to the different levels of educational ability shown by pupils in different social classes. Pupils’ own knowledge of their ability is likely to shape the subjective probability they attach to success in a particular curriculum. The third mechanism is that middle class families have more resources to devote to their children’s education. So, more middle class than working class families can allow their children to continue in education for longer. Breen and Goldthorpe’s work is an advance on Boudon’s approach, in that it makes explicit assumptions, and these have testable implications. The main one of these is that before pursuing a particular curriculum, working class children will require a higher probability of success than middle class children. This view would have to be revised if either: 1. Working class pupils do not require a higher level of perceived success than middle class children before pursuing a particular curriculum; or 2. Pupils’ expectations of success are formed differently according to social class, rather than being formed purely according to objective evidence. However, one would need to gather subjective data on expectations to test Breen and Goldthorpe’s view — and this is something that rational choice theorists have not generally been inclined to do. Or, as Manski puts it: “The standard economic model assumes that a youth’s schooling choice c is a function f (.) of his or her expected returns to schooling r; that is c = f (r). Suppose that one wishes to learn the decision rule f (.) mapping expectations into choices. If one observes the choices and expectations of a sample of youth, then one can infer the decision rule. But if one observes only the choices of these youth, then clearly one cannot infer f (.).” (Manski, 1993, p. 44–45) One problem with Breen and Goldthorpe’s model is that one cannot infer from it how pupils from ‘underclass’ backgrounds will behave, since there is no risk of social demotion for these pupils. One could argue that these pupils will have low aspirations since they come from humble backgrounds, but equally, one could argue that they will be able to take ambitious educational options since they face no risk of social demotion if they fail.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

38

I would also argue that ‘relative risk aversion’ is a misleading term to describe Breen and Goldthorpe’s model. Risk aversion refers to a preference for a large probability of a small gain rather than a small probability of a large gain (e.g. I am risk averse if I prefer a 90% chance of winning $50 to a 10% chance of winning $5,000). This concept of risk aversion does not apply to Breen and Goldthorpe’s model. On this model, a service class pupil will pursue an educational option on the grounds that it represents their best chance of remaining in the service class, even if this means a greater risk of demotion to the underclass than not pursuing this option would entail. So, perhaps Breen and Goldthorpe’s model is more accurately described as one of ‘loss aversion’ than of ‘risk aversion’. Pupils perceive themselves as already having a particular social class status (i.e. that of their parents), and they will risk everything to maintain that status. In addition, I dispute Breen and Goldthorpe’s claim that their model does not rest on an assumption of class differences in preferences. It is clear that, on Breen and Goldthorpe’s model, the working class attach a higher value to a working class outcome than do the service class, and so on. Thus, preferences for each outcome class vary by class of origin.

2.5.1

Applying the Utility Model to Breen and Goldthorpe

I wish to derive Breen and Goldthorpe’s formulae for p from the utility model described above. The formulae for the service class pupil and for the working class pupil are as follows: pS =

πα + (1 − π)β1 πα + (1 − π)β1 + γ1

π + (1 − π)(β1 + β2 ) π + (1 − π)(β1 + β2 ) + (γ1 + γ2 ) Where the various variables are defined by: pW =

1. π is the perceived chance of ‘passing’ given that one has stayed on 2. α is the probability of entering the service class given that one has ‘passed’. 3. β1 is the probability of entering the service class given that one has ‘failed’

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

39

4. β2 is the probability of entering the working class given that one has ‘failed’ 5. γ1 is the probability of entering the service class given that one has not stayed on 6. γ2 is the probability of entering the working class given that one has not stayed on. The situation is summarised diagrammatically by figure 2.3 (figure 1 in (Breen and Goldthorpe, 1997)). α.............................................r S

.......... ..........

Stay on

Pass ..r..................................................................................................................................................r ........ W π ............................................... 1 − α ... . . . . . . . . . ........

.... .......... .......... .......... .......... ............................. . . . . . . . . . .......... .......... .. .......... .......... .......... .......... .......... . .......... . . . . . . . . . . . . . . . . . . . .......... .. .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... ................... .......... ................................................................................................................ ................ .......... ........ .......... ........ .......... ........ .......... ........ .......... ........ .......... ........ .......... ........ .......... .......... ........ .......... ........ ..... ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ... ........ .......... ........ .......... ........ .......... ........ .......... ........ .......... . . . . . . . . . ........ .... ........ .......... ........ .......... ........ ................... .................................................................................................................. .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... ........

r

r

rS

β1

1−π

β2

r

Fail 1 − β1 − β2

rU rS

γ1

r

Leave 1 − γ1 − γ2

rW

γ2

rW rU

Figure 2.3: Breen and Goldthorpe’s model

The central assumption of Breen and Goldthorpe’s ‘relative risk aversion’ model is: “. . . families in both classes alike seek to ensure, so far as they can, that their children acquire a class position at least as advantageous as that from which they originate or, in other words they seek to avoid downward social mobility.” (Breen and Goldthorpe, 1997, p. 283) We can express this in terms of utilities as shown in table 2.1:

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES For service class Perceived utility of service class occupations Perceived utility of working class occupations Perceived utility of underclass occupations

40

1

For working class 1

0

1

0

0

Table 2.1: Perceived Utilities in Breen and Goldthorpe’s Model

If one applied this to members of the underclass they would perceive all occupational classes as being of equal utility. Breen and Goldthorpe would certainly resist the application of the concept of utility to their work. In particular they would resist the view that they give an explanation that at root relies on different perceptions of utility between social classes. However, as we shall see, their model can be derived from the above table. Suppose that the probabilities of entering the various classes given the three educational outcomes are as described in figure 1 of Breen and Goldthorpe: Then one can say: U1S = α, U1W = 1,

U2S = β1 ,

U2W = β1 + β2 ,

U3S = γ1 U3W = γ1 + γ2

where the superscripts S and W refer to the service and working class respectively, and the U ’s refer to the expected utility of each possible educational outcome (labelled 1 [staying on and passing], 2 [staying on and failing] and 3 [leaving]). One can now write the expected utility from staying on, which we recall is: π(U1 ) + (1 − π)U2 and the expected utility from leaving which is U3 . Thus for service class pupils the parameter p is given by: pS =

πα + (1 − π)β1 πα + (1 − π)β1 + γ1 )

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

41

and similarly for working class pupils it is given by: pW =

π + (1 − π)(β1 + β2 ) . π + (1 − π)(β1 + β2 ) + (γ1 + γ2 )

These are Breen and Goldthorpe’s formulae 1 and 2 (Breen and Goldthorpe, 1997). Thus, their model fits into the utility model completely. This reveals the implicit reliance of their model on the differences between utilities as perceived by each class. Therefore Breen and Goldthorpe have clearly failed in their stated aim of providing an explanation of class differentials which does not ultimately rely on class differences in preferences — i.e. they have failed to avoid a ‘cultural’ or ‘normative’ account.

2.5.2

Breen and Goldthorpe’s Assumptions

Further to the table of utilities, it is still necessary to make some assumptions about the parameters π, α, β etc. if one is to extract any predictions from the model. Breen and Goldthorpe make four assumptions about the parameters in their model, which I will now discuss: 1. α > β1 ,

α > γ1

2. γ1 + γ2 > β1 + β2 3.

γ2 γ1

≥ 1,

γ2 γ1



β1 β2

4. α > 0.5 The first assumption is uncontroversial — it simply states that remaining at school and succeeding affords a better chance of access to the service class than does remaining at school and failing or leaving school. The second assumption is most important to the conclusions drawn by Breen and Goldthorpe. It states that remaining at school and failing increases the chances of entering the underclass as against just leaving school. Breen and Goldthorpe make no attempt to justify this rather striking assumption with empirical evidence. As I pointed out earlier, Payne’s (Payne, 1987) analysis of NCDS (National Child Development Survey)data seems to contradict the assumption in the case of the UK.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

42

Assumptions 3 and 4 are less central to Breen and Goldthorpe’s model than assumption 2. However, the dubiousness of these assumptions still deserves to be pointed out. Assumption 3 is stated as: “Those who leave school immediately have a better chance of entry into the working class than the service class. This may or may not be the case among those who remain at school and fail though, if it is, their odds of entering the working class rather than the service class are no greater than for those who leave school immediately” (Breen and Goldthorpe, 1997, p. 282) The first part of this assumption constrains the relative sizes of the respective classes. For instance, the assumption would be false in the case where 70% of the population is middle class, and 30% working class, and only 30% of the population stay on in education at a given decision point. In this example, even if all 30% who stay on at school end up in the service class, more than half of those who leave school immediately must also. (What we mean by ‘passing’ and ‘failing’ are also highly relevant here.) Since assumption 3 can only be true given a favourable combination of class structure and staying on rate (and even then the assumption will not necessarily be true) it would seem reasonable to demand some empirical justification for the assumption. This is not provided. It seems very unlikely that assumption 3 would hold for higher level educational decisions, such as whether or not to go to university. The second part of assumption 3 fails to take into account the possibility that those who fail a course of study may still gain labour market benefits above and beyond what they would have gained from any work experience they missed out on due to study. Assumption 4 is stated as “Staying on at school and passing the examination makes entry to the service class more likely than entry to the working class.” (Breen and Goldthorpe, 1997, p. 282) The plausibility of the assumption depends on how we define a pass or fail. If passing the examination means getting at least 3 Cs at A level, the assumption seems plausible. If however, we define a pass as one E grade at A level, then the assumption that this makes entry to the service class more likely than entry to the working class becomes questionable.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

43

Given the expressions for expected utilities we derived above, we can rewrite these assumptions as: 1. U1S > U2S ,

U1S > U3S .

2. U3W > U2W 3.

U W −U3S U3S

≥ 1 ≡ U3W − U3S > U3W ≡ U3W > 2U3S and

U W −U3S U3S



U2S U2W −U2S

4. U1S > 0.5 In words: 1. For the service class, the utility of staying on and passing is greater than the utility of staying on and failing. For the service class, the utility of staying on and passing is greater than the utility of leaving. 2. For the working class, the utility of leaving is greater than the utility of staying on and failing. 3. The utility of leaving is more than twice as big for the working class as for the service class. Putting the additional assumption into words does not make it any clearer than it is as a formula. 4. For the service class, the utility of staying on and passing is greater than 0.5 (note that 1 is the maximum available utility) Assumptions three and four only look more contrived and implausible when phrased in terms of utilities. One striking fact about the assumptions once they are phrased in terms of utilities is that they contain an asymmetry between the working and service classes. In order to avoid this asymmetry, one would have to add the following assumptions. 1*. For the working class, the utility of staying on and passing is greater than the utility of staying on and failing. For the working class, the utility of staying on and passing is greater than the utility of leaving. 2*. For the service class, the utility of leaving is greater than the utility of staying on and failing. Assumptions 1 and 1* are certainly reasonable assumptions. It is clear that assumption 2 is the driving force behind Breen and Goldthorpe’s argument.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

44

However, clearly assumptions 3 and 4 (which are unfounded) must also play a key role in their argument.

2.5.3

A Generic Model?

Let us examine Breen and Goldthorpe’s claim that their model is generic.

2

“The model that we present is intended to be generic: that is, as one applicable in principle to the entire range of decisions that young people may be required to make over the course of their educational careers as regards leaving or staying on or as regards which educational option to pursue.” (Breen and Goldthorpe, 1997, p. 279) It is a doubtful empirical question whether assumptions 2, 3, and 4 hold when the decision in question is whether to stay on at school post 16. It is clear, however, that assumptions 2, 3 and 4 cannot hold when applied to all educational decisions. Assumption 2 (that staying on and failing increases the chance of entering the underclass as against leaving) would seem not to hold in the case of US college students. As I pointed out earlier, research shows that individuals with uncompleted college degrees fare better in the labour market than individuals who did not go to college. Of course, there is always a difficulty in evaluating such research. (For instance, it may be that individuals who attempt college but do not graduate are simply more able than those who do not, and would have done better in the labour market even if they had not gone to college.) But the risk of entering the underclass will certainly decrease as we consider more ambitious educational options, becoming insignificant when we consider decisions such as whether to train to be a barrister having completed an undergraduate degree. Assumption 3 states that those who leave immediately have a better chance of entry into the working class than the service class. Again, this is clearly 2

In private communication, John Goldthorpe has clarified to me that the word ‘generic’ is intended in an extremely weak sense here, and that the model is only applicable to decisions which lead to class differentials, and even then, applying the model to the full range of decisions would necessitate changing the outcome classes. However, I do not think that this weak usage of the word ‘generic’ is at all obvious from the text, so perhaps it may be useful to explain why a stronger claim would be false.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

45

not true of more advanced options, such as the decision of whether to leave education after one’s undergraduate degree or to take a postgraduate course. Assumption 4 claims that staying on and passing the examination makes entry to the service class more likely than entry to the working class. This assumption would be implausible in the case of less prestigious options. E.g. consider the decision of whether to stay on to take a vocational qualification such as a GNVQ or leave the education system. There is a special problem with decisions where more than one alternative is considered simultaneously. For example, the decision of whether to stay on in the sixth form to do A-levels or leave might arguably meet assumptions 3 and 4. However, one must question whether any individual actually makes this decision. I would argue that the decisions made by real individuals are much finer e.g. the decision to do A-levels or a GNVQ. In this more relevant example condition 4 might be plausible in that passing A-levels (if by this we mean getting reasonably good grades) might make the service class the most likely destination. It is not necessarily clear however that someone who does a BTEC has a better chance of entry to the working class than to the service class. If Breen and Goldthorpe’s model is to apply to this decision, then this must be the case by assumption 3. Suppose for the sake of the exposition that this is the case. Since, by assumption 3, it is true that the probability of entering the working class is greater than the probability of entering the service class if one takes a BTEC, and assuming the probability of failing a BTEC is not too high — it is not at all unlikely that a pupil who passes a BTEC also has a greater probability of entering the working class than the service class. Now consider the case of a pupil who decides between taking a BTEC and leaving education. For this decision assumption 4 reads: the probability of the pupil entering the service class if they do a BTEC is at least one half. This is of course impossible if the probability of them entering the working class is greater than the probability of him entering the service class. Thus assumptions 3 and 4 cannot apply across the range of educational decisions. Furthermore they do not apply to realistic educational decisions. Goldthorpe claims that: “The assumption that the educational decision problem has three possible outcomes, each of which results in direct entry to the labour market, is innocuous” (Breen and Goldthorpe, 1997, p. 287)

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

46

If this statement is taken to mean that applying this assumption to the full range of educational decisions will be unproblematic, then it is certainly false. In addition, extending assumptions 3 and 4 to a model containing many social classes would be highly problematic. So, Breen and Goldthorpe’s model would have to be adapted considerably to apply to the full range of educational decisions. In my view, this makes the claim that it is a generic model somewhat misleading.

2.5.4

Breen and Goldthorpe’s attempt to explain empirical changes

Increase in Participation in Education The first empirical phenomenon Breen and Goldthorpe try to explain is the increase in participation in education by children of both working and service class origins. They explain this by the fact that the relative costs of education have declined over time in all economically advanced countries. However, they go on to say that: “. . . in so far as education is regarded as a ‘positional’ good, p could be expected to rise steadily simply as a consequence of educational expansion itself.” (Breen and Goldthorpe, 1997, p. 294) By this they presumably mean that increased rates of educational participation have made extended educational participation increasingly necessary for individuals who wish to enter the service class. For an earlier formalisation of this, highly plausible, view see (Halsey et al., 1980, p. 113-116). However, according to Breen and Goldthorpe’s model the working class is indifferent to the distinction between a service and a working class occupation. Thus this change would not necessarily affect them. Indeed this change could occur simultaneously with a growth in the chance of entering the underclass if one leaves education. As argued previously assumption 2 then requires a greater chance of entering the underclass if one pursues education and fails. This would lead to the working class being less likely to persist in education. Thus the only explanation of rising participation in education that can be derived from Breen and Goldthorpe’s model is that of decreasing costs.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

47

Persistence of Class Differentials The second phenomenon that Breen and Goldthorpe attempt to explain is the fact that class differentials in educational participation have remained stable despite this increase in overall participation rates. Given the fact that the underclass (if this is, as they suggest, to represent the unemployed and those with a tenuous position in the labour market) has grown over the past two decades, and given Breen and Goldthorpe’s assumption 2 — which states that the risk of falling into the underclass is greater if one fails in education than if one chooses to leave — it seems likely that the working class (if they are trying to minimise their chances of falling into the underclass) would be even less inclined to risk pursuing further education, all other things being equal. This contradicts the empirical fact that the working class has increasingly pursued further education as have the service class. Thus, to account for empirical facts, one is forced to reject either assumption 2 or the assumption that the working class’ educational decisions are driven by the desire to minimise their chances of falling into the underclass.

Gender Breen and Goldthorpe state that in the past, women’s class position did not depend strongly on their educational qualifications. Therefore, Breen and Goldthorpe’s, model could not be applied to women pre 1970s. However, due to increased labour market participation by women this is no longer the case and the model increasingly applies to women. Breen and Goldthorpe seem to claim that their model explains the decrease in gender inequality in educational participation. However, in the light of the statement that the model did not apply to women in the past, it is hard to see how this could be the case. They can only explain why the participation of women is currently similar to that of men. They cannot use their model to explain the change. Clearly, Breen and Goldthorpe’s model is designed to explain class differences, and the idea that a key aim of their paper is to explain the erosion of gender differences in educational participation is laughable. This is a clear example of gender being seen as an obligatory “addon” in sociological analysis, rather than being taken seriously as a category of analysis.

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

2.6

48

Conclusions

Different rational choice theorists have very different approaches to the explanation of educational inequality, as they do to other questions. Murphy uses class differentials in preferences as an explanation of class inequality in attainment, whereas Boudon claims that rational choice theory is in conflict with theories that postulate differences in preferences and values between social classes. Gambetta states that structural constraints that limit agents’ scope for choice also limit the explanatory force of rational choice theory. In my view, the rational choice assumption cannot be in direct conflict with any claims about beliefs, desires or constraints, since rational choice is a claim about a decision rule, not about the beliefs, desires or constraints which are factored into the decision. No determinate hypotheses can be derived from the decision rule alone, and the failure to fully realise this seems to lead to some confusion among rational choice theorists. I have applied a generic model to the various approaches considered here, and I hope this has been helpful in bringing out the key differences. I have given particular attention to Breen and Goldthorpe’s model of educational decision-making. My main conclusions on this model are as follows: Breen and Goldthorpe describe their model as being one of “risk-aversion”, but it is better understood as a model of “loss-aversion”. Breen and Goldthorpe fail to give an explanation of class differentials in educational participation rates that does not rely on class differences in preferences, although the preferences of each social class do arise from a common mechanism (the desire to avoid social demotion) on their model. In fact, Breen and Goldthorpe’s model implies that an individual’s class of origin affects the value they attach to each destination class, such that individuals from service class origins attach a higher value to achieving a service class “destination” than do individuals from working class origins. The assumptions on which Breen and Goldthorpe’s model is based are highly problematic, and are not backed up by empirical data. For instance, the assumption that remaining in education and failing increases the chance of entering the underclass as compared to just leaving education is crucial to the model, but what empirical evidence is available seems to contradict rather than support this view. Breen and Goldthorpe’s model would have to be adapted enormously if it was to apply to all decisions about educational participation or to all class

CHAPTER 2. RATIONAL CHOICE PERSPECTIVES

49

schemas. Therefore, to describe the model as “generic” is misleading. Breen and Goldthorpe attempt to explain empirically observed changes in educational participation rates in terms of their model. This attempt is unsatisfactory even from the perspective of their model. Crucially, none of the authors cited here has sought to provide empirical evidence for the assumptions they have made regarding the motivations of actors. In subsequent chapters I will attempt to bring empirical evidence to bear on the questions of whether beliefs and attitudes regarding education vary according to social class and gender, and how class and gender differentials in educational decision making can be explained.

Chapter 3 Methodology 3.1

Introduction

In previous chapters, I have discussed the attempts of two theoretical approaches — cultural reproduction theory and rational choice theory — to explain educational inequalities. I have argued that neither of these approaches has been subjected to adequate empirical test. This chapter will describe the survey that I have carried out, and the relevance of this empirical work to the theoretical questions at hand. I will discuss the operationalisation of key concepts such as cultural capital, subjective beliefs about ability, and attitudes towards education. I will also describe the sample, and the practical questions involved in conducting the survey. (The survey questionnaire is included in appendix B).

3.1.1

Cultural Capital

According to Bourdieu’s theory of cultural reproduction, cultural resources associated with the middle-class home facilitate the acquisition of educational credentials (see for example (Bourdieu and Passeron, 1990; Bourdieu, 1977a)). However, I have argued that the concept of cultural capital is ill defined, and has been operationalised in various, sometimes rather arbitrary ways. In order to find out which, (if any), cultural resources are important in determining educational attainment, I will give as broad as possible an opera50

CHAPTER 3. METHODOLOGY

51

tionalisation of cultural capital. Most surveys of cultural capital focus on culturally ´elite activities, such as attendance at galleries and classical concerts. I asked these conventional questions, but I also asked about more popular activities, for example, which TV programmes pupils watch. This is because questions on participation in extremely high-culture activities can only distinguish between a tiny minority who engage in these activities and the rest — leaving the bulk of the population undifferentiated. This point is particularly important when you consider that I am asking these questions to 16-year-olds. In addition to questions about cultural activities, my questionnaire included tests of vocabulary and cultural knowledge. These items were included because they represent a possible mechanism through which cultural participation may affect educational attainment. Initially, I planned to ask pupils questions about the education system itself, on the grounds that information about different educational options might be an important advantage conferred by the middle class home. (This might be seen either as an element of cultural capital, or as an advantage in decision making on the rational choice model, or both.) Unfortunately, I was forced to abandon this aim on the grounds that my questionnaire would be overlong if I tried to do too much. However, I still think that this is an important area for research, (this view was reinforced when one girl asked me what a university was).

3.1.2

Rational Choice Theory

A type of explanation in competition with the cultural capital approach is the rational choice approach. For example, Murphy claims that working-class children simply do not demand as much education as middle class children do (Murphy, 1981). But why do working-class children “demand less” education than middle-class children — i.e. why are they less likely to stay on at school post-16 and go on to university? Possible explanations include: • Working class pupils’ perceived probabilities of success in education are low as compared to middle class pupils of similar ability. • In comparison to middle-class pupils, working-class pupils undervalue education either as a means to success in the job market or as a good in itself

CHAPTER 3. METHODOLOGY

52

• Breen and Goldthorpe’s view that middle class pupils attach a stronger value to gaining middle class jobs than do working class pupils (Breen and Goldthorpe, 1997). Rational choice theorists have typically shied away from using “soft” data on pupils’ motivations, relying instead on the doctrine of revealed preference, which claims that individuals’ motivations are most reliably revealed by their actions. An obvious problem with this view is that the pattern of pupils’ observed choices is compatible with any of the above explanations of this pattern. I will assess these competing explanations using data on students’ beliefs, attitudes, plans, and aspirations. In order to assess the possibility that working class pupils’ perceived probabilities of success in education are low as compared to middle class pupils of similar ability, it was necessary to collect data on pupils’ subjective view of their own academic abilities, as well as data on their actual academic performance (GCSE results). This has also allowed me to assess the possibility that beliefs about ability are distorted by gender. In order to assess the possibility that working class pupils place a lower value on education than do middle class pupils, I collected data on pupils’ attitudes towards education. This has also allowed me to examine the question of whether there are gender differences in attitudes towards education. For Breen and Goldthorpe’s view (Breen and Goldthorpe, 1997) to be supported, it would have to be the case that social class differences in educational plans are explained by social class differences in occupational aspirations. Therefore, I asked pupils about their occupational aspirations.

3.2

Survey Design

I surveyed pupils in their final year of compulsory schooling (i.e. “year 11” pupils, about 16 years old) in England, in 1998. I chose to survey year 11 pupils because I wished to follow up on the actual GCSE results attained by pupils and on what they ended up doing after their GCSEs. If I had used a younger year group, I would have had to wait too long for this data. If I had used an older year group, a large proportion of pupils, those who do not stay on in the 6th form, would have been excluded from the sample.

CHAPTER 3. METHODOLOGY

53

Before GCSEs I surveyed children on: • parents’ occupation and educational level • family structure and siblings • parents’ cultural participation • pupils’ cultural participation • pupils’ cultural knowledge • pupils’ active and passive vocabulary • pupils’ expectations of exam grades • pupils’ intentions for future study • pupils’ occupational expectations and aspirations • self-evaluated intelligence • pupils’ attitude towards the intrinsic and instrumental value of education. Initially, I aimed to survey parents. This is because a theory of cultural reproduction cannot be tested using only information on pupils’ cultural capital. A correlation between pupils’ cultural capital and attainment does not support a theory of cultural reproduction unless there is also a correlation between pupils’ and parents’ cultural capital. However, I decided that surveying parents was not a realistic option, since few would have volunteered to participate. Therefore I have surveyed pupils on their parents’ characteristics. The cultural activities pupils were asked about regarding their parents include reading (and number of books in the home), newspapers taken, type of music and radio stations listened to, “formal culture” participation, and the subjects discussed by parents in the home. It would have been ideal to get information on parents directly from the parents themselves, as responses from pupils may be less accurate than responses from parents. For instance, it seems possible that pupils with high levels of cultural participation may over report their parents’ cultural participation. However, note that de Graaf et. al.(De Graaf et al., 2000) find that, in a survey of cultural practices in

CHAPTER 3. METHODOLOGY

54

the Netherlands, respondents own cultural practices have no effect on their reporting of their parents’ cultural practices. At any rate, the best available option was to get information on parents’ cultural activities from the pupils.

After GCSEs I have collected information on the actual GCSE results gained by the pupils. As far as possible, I have collected information on the pupils’ actual educational outcomes — whether they continued in education, and if so, what course of study they followed. However, the level of detail available on this question varied from school to school, and one school failed to provide the information altogether, despite making a commitment to do so at the start of the project, and despite repeated requests from me. Therefore, in my analyses of pupils’ decisions, I have used the pupils stated intentions of educational participation as an outcome variable. A table showing the associations between the key variables is shown in appendix A.

3.2.1

Pilot

The survey was piloted in order to assess the questionnaire. In particular, I needed to determine whether the questionnaire could be completed in the time available, even by pupils’ of relatively low ability, whether all the items on the questionnaire could be readily understood by pupils and whether the tests of cultural knowledge and active and passive vocabulary were pitched at approximately the right level. In general, I needed to confirm that all items that required a spread of responses in order to discriminate between individuals actually did so and that as far as possible, no item caused confusion or offence. I piloted my questionnaire at an Oxfordshire comprehensive, with a small sample (16 pupils), on July 18 1997. I used year 10 students rather than year 11 students. This was because year 11 students would have been too busy with their GCSE’s at this stage in the year to participate. I introduced the questionnaire, explaining that it asked them about various things, including what they did in their spare time, what they thought about studying, and what they wanted to do in the future. I stressed the confidentiality of the questionnaire. Since I did not intend to follow up on GCSE results for these

CHAPTER 3. METHODOLOGY

55

students, or what they did after their GCSEs, there was no need for them to identify themselves on the questionnaire, either by name or number, so the responses were completely anonymous. The pupils completed the questionnaire without much difficulty, and did not object to any of the questions. Only a few changes to the questionnaire were necessary. The only parts of the questionnaire that pupils had serious difficulties with were those intended to test active vocabulary. I had taken an open-ended sentence completion test from Denis Lawton’s Social Class, Language and Education (Lawton, 1968). This test involves completing sentences so that they make sense, using a specific number of words and without breaking into more than one sentence.

Example: The admiral went. . . . . . . . . . . . . . . . . . . . . . . . . . . .(add ten words) The admiral went to China six times when he was in the navy.

This test was originally used to test the ability to formulate sentences among 15-year-old boys, in order to compare the complexity of grammatical structures used by middle and working class boys. This test was too hard for the pupils in my sample — most of them managed at best a few correct responses. Even those pupils who gave correct responses to most of the items did not attempt to use complex grammatical structures — and, one may ask, why on earth should they, as they were not instructed to attempt to do this. It is possible that social class differences in performance on this test could reflect the ability to know what the assessor is going to reward in the absence of explicit instructions on this (i.e. a knowledge of the rules of the game) as much as superior writing skills. In addition, the results of this test would have been difficult and time consuming to analyse. Another test of active vocabulary from Lawton (Lawton, 1968) that I tried in the pilot was to ask pupils to write a short essay on one of the following subjects “home”, “school” or “my life in 10 years time”. Many pupils did not even attempt this, and those that did tended to write very short pieces indeed. Furthermore, the results of this test would have been perhaps even more cumbersome to analyse than those of the previous test. I have replaced these tests with my own test of active vocabulary, which is described in 3.3.1. Pupils have not found this test too difficult, and it has

CHAPTER 3. METHODOLOGY

56

the advantage that they can give up after thinking of as many synonyms as they can. In addition, this test can be scored as a simple mark out of five for each question. The other changes I made were minor. For instance, there were a few words that some pupils did not understand. Doing the pilot allowed me to identify words and phrases that were too difficult in order to change them for the main survey. This was important, as I could never have predicted some of the words pupils found difficult. However, this did not always work out as planned. For instance, some pupils did not understand the word shine as in “I would like a job that will allow me to shine”. I replaced the word “shine” with “excel”, only to find that several pupils did not understand that word either. The pilot school is predominantly working class, with low GCSE results, (in 1996, 21.5% of its pupils achieved 5 A–C grades). This made me feel reasonably confident that, if these pupils could complete the questionnaire easily, others should have no difficulty either. I found that the pupils’ scores on the passive vocabulary and cultural knowledge tests were approximately normally distributed, which was encouraging, although one cannot say too much on the basis of such a small sample.

3.3

Questionnaire Design and Coding

In constructing my questionnaire, I drew from various previous surveys of young people designed to measure various factors such as cultural participation, language ability, aspirations and attitudes towards school, (these include: PROJECT TALENT, British Social Attitudes: Young People’s Survey SCPR, 1994 and National Child Development Survey, ESRC Data Archive, 1993).

3.3.1

Cultural Capital Questions

I have surveyed pupils on a broad range of possible components of cultural capital. 1. Cultural knowledge

CHAPTER 3. METHODOLOGY

57

• Tested knowledge of famous cultural figures. 2. Language • Active and passive vocabulary test scores. 3. Activities • Reading: type and amount of books read, library use, newspapers read. • Television: type of TV programmes watched. • Music: type of music listened to, playing an instrument. • Participation in “public” or “formal” culture: art gallery, theatre and concert attendance. (Film attendance was included in the questionnaire, but excluded from the analysis, since only 7 respondents had watched films that could be categorised as in any sense “high-brow”.)

Cultural Knowledge The test of cultural knowledge is intended to measure pupils’ familiarity with the dominant culture. This test is of course not intended to reflect all aspects of a pupil’s cultural knowledge. However, it at least provides us with some indication of cultural knowledge, something that has been lacking in most previous research on cultural capital. The test consists of asking pupils to categorise 25 famous cultural figures according to whether these figures are associated with politics, music, novels, art or science. There were equal numbers of cultural figures in each of these categories. The figures were chosen to reflect both past and contemporary culture. The items were pitched at different levels of difficulty in order to give a scale that could differentiate well between individuals in the sample according to their level of cultural knowledge. Some figures were intended to be easily identifiable by practically anyone, while others were intended to be more difficult. Pupils were instructed not to guess if they did not know the answer, and in constructing a score, a deduction of 0.25 was made for false responses. (The figure of 0.25 simply reflects the fact that, for each question, there were five possible responses other than “don’t know”. Therefore, a pupil who knew none of the answers, but guessed consistently, could expect to get, on average, one in five correct. A deduction of 0.25 for each false response would balance this out, leaving the pupil with a score of zero.) Making a deduction for false

CHAPTER 3. METHODOLOGY

58

responses is standard practice in scoring multiple choice questions, and is important in the light of the possibility that boys may be more likely than girls to guess in multiple choice tests (Gipps and Murphy, 1994). The questions were introduced as follows:

Cultural Knowledge Each of the following names is a person you may have heard of. For each person listed, which do you associate him or her with most out of the following categories: politics, music, novels, art or science? If you do not know, do not guess, just tick “don’t know”.

The names of the people and the number of correct responses for each are summarised in Table 3.1. These results show that, in the case of novelists, the modern figures were far less well known than pre 20th Century ones such as Dickens and Austen. Similarly, among those associated with music, only Mozart was well known, and the least known figures were Gershwin and Miles Davis — those least likely, perhaps, to be categorised as ´elite rather than popular culture. The artists buck this trend in as much as Picasso is correctly identified as an artist by 72% of respondents. However, the most recent artist (Warhol) was the least known, despite his obvious association with popular culture, not least through the Velvet Underground. Only in the field of politics were contemporary and recent figures the best known, perhaps unsurprisingly given the role of the media, and in the case of JFK, an eponymous film that had been on at the cinema shortly before I administered my questionnaire. Only 25% could correctly categorise Marx. Of the scientists, Einstein stood head and shoulders above the rest as the only scientist that could be correctly identified by a majority of respondents (although it still seems a little shocking that one in ten failed to do so). The distribution of pupil’s scores is shown in Figure 3.1. This shows an approximately normal distribution with a wide spread of scores indicating that the test was pitched at about the right level of difficulty.

Passive vocabulary Passive vocabulary consists of those words that an individual can understand, but cannot necessarily actually use in their own writing or speech. The test of passive vocabulary used in the questionnaire was a conventional multiple

CHAPTER 3. METHODOLOGY

Albert Einstein Galileo Marie Curie Louis Pasteur Stephen Hawking Charles Dickens Jane Austen Virginia Woolf Graham Greene Martin Amis Bill Clinton John F. Kennedy Mahatma Gandhi Gordon Brown Karl Marx Wolfgang Mozart Rachmaninov Johannes Brahms Miles Davis George Gershwin Vincent van Gogh Pablo Picasso Claude Monet Rembrandt Andy Warhol

59

Percentage correct response 90 31 27 21 21 89 76 23 7 5 94 89 47 33 25 81 18 15 12 10 85 72 49 29 23

Table 3.1: Knowledge of cultural figures

CHAPTER 3. METHODOLOGY 100

60

frequency

90 80 70 60 50 40 30 20 10 0 0

2 4 6 8 10 12 14 16 18 20 22 24 26 Figure 3.1: Histogram of knowledge score

choice “sentence completion” test (see for instance (Levy and Goldstein, 1984)). This test is derived mainly from the National Child Development Study — SWEEP III (NCDS, 1993) (this test was administered to 16 year olds in 1974). 14 out of the 15 questions I used were derived from the 35 questions used by NCDS. I could not use the full NCDS test, as it would have taken too long. I found many of the questions originally used by NCDS ambiguous, i.e. in my view, there was more than one correct answer to many of the questions. For example:

In the days when people lived in caves and rough shelters there were no fields or farms; men got their food by hunting animals and by (finding, planting, gathering, growing, eating) berries and wild fruit.

I would guess that gathering was the required response, but in what sense is finding actually wrong, given that this word makes a perfectly good and sensible sentence? Not wishing my test to be an exercise in second-guessing the tester, I was careful to select only those questions which, in my view, had only one correct answer — an example is shown below. 15 such questions were asked and the scores were added to give each pupil a mark out of 15. The distribution of pupil’s marks for this test is shown in figure 3.2

CHAPTER 3. METHODOLOGY

61

Underline the word that correctly completes the sentence. If you do not know which word is correct, just move on to the next question. • Having had her expectations so much raised, it was very (realistic, discrediting, uplifting, disconcerting, discriminating) to have them suddenly shattered.

60

frequency

50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Figure 3.2: Histogram of passive vocabulary score — 41 were coded as missing

Active vocabulary Active vocabulary consists of those words that an individual can actually use, rather than simply understand. When piloted, tests of active vocabulary that have been used elsewhere proved unsatisfactory (see section 3.2.1 for details). Therefore, I developed a new test of active vocabulary. This test demands that pupils provide several synonyms for each of five words given.

CHAPTER 3. METHODOLOGY

62

Synonyms are words that mean the same or approximately the same as each other. For example happy and cheerful are synonyms. For each of the following questions think of as many synonyms as you can for each word (stop at five for each word). 1. Small 2. Stupid 3. Angry 4. Sad 5. Odd

Pupils were given a mark for each genuine synonym they provided thus giving a total mark out of 25 for active vocabulary. The results are summarised in Figure 3.3. Again there is a reasonable spread of results suggesting that this test was pitched at about the right level of difficulty. 80

frequency

70 60 50 40 30 20 10 0 0 2 4 6 8 10 12 14 16 18 20 22 24 Figure 3.3: Histogram of active vocabulary score — 91 were coded as missing

The passive and active vocabulary scores were summed to give an overall language score for each pupil. This composite score will be used in subsequent statistical models as a measure of pupil’s linguistic ability. The distribution is shown in Figure 3.4. This distribution is approximately normal.

CHAPTER 3. METHODOLOGY 60

63

frequency

50 40 30 20 10 0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 Figure 3.4: Histogram of language score — 34 were coded as missing

Reading Habits Previous surveys of cultural capital have generally asked individuals how much they read, and sometimes also whether they are a member of a public library, and how many books they have at home. I asked these questions, but also asked pupils what books and newspapers they read. The section of the questionnaire concerned with pupils’ reading habits asks pupils to state the titles and authors of any books they have read recently that are unconnected with their schoolwork, and to give the names of their favourite authors if they have any. Clearly, the responses to these questions must be categorised according to the “cultural capital” content of the books. The categorisation I used is shown in table 3.2. The starred categories are those for which credit for cultural capital was given. When unsure of the category a book fell into, I used the Book Review Digest database. This is a database of reviews from 100 English language journals such as the Times Literary Supplement and the New York Review of Books, from 1983 to the present. This allowed me to check out any title or author to see whether they had received reviews from the quality press, and also, to read those reviews in order to determine the genre of the book. Given the role of prestigious journals such as the Times Literary Supplement in conferring legitimacy on high culture, this seems like a reasonable way of determining the cultural status of contemporary books. The percentages of pupils who had recently read a book by an author in each category and who had a favourite book or author belonging to a given category is shown in Table 3.3. In deciding whether a book counts as high

CHAPTER 3. METHODOLOGY Category Children’s books Teenage books Factual low culture Adult low culture — including much horror, science fiction, fantasy and romance and some factual books, e.g. travel *Factual high culture books *Classic *Modern high culture — 20th century *Contemporary high culture — books of the sort that receive reviews in the quality press Other

64 Examples Roahl Dahl, Enid Blyton Judy Blume, Point Horror series Guinness book of records, books about cars, etc. Stephen King, Terry Pratchett, Danielle Steele

Science, humanities, some biographies, etc. Dickens, Austen Orwell, Graham Greene Salman Rushdie, Martin Amis

Table 3.2: Coding of Books

or low culture I have erred on the side of charity, since anything outside the lowest culture is quite rare. For instance, a fairly sophisticated science fiction author like Philip K. Dick would count as high culture. For those pupils that stated the authors and/or titles of any books, teenage books and adult low culture predominated. When coding the data, I saw that the same authors came up very regularly (Stephen King, Terry Pratchett, Danielle Steele). “High culture” books (the four starred categories) were included in the responses of only a small minority. Pupils were given a “reading score”, consisting of points awarded for: reading books unconnected with schoolwork once a month (1 point), reading such books once a fortnight or more (2 points), library membership, having any favourite author, having a “high brow” favourite author, “high brow” recent reading, and reading a broadsheet newspaper. This gives a score out of 7. The distribution of this score is shown in Figure 3.5.

Percentage with a favourite author in this category

Classic Contemporary high culture Modern classic Factual high culture Adult low culture Factual low culture Teenage Childrens Other None

65

Percentage who had read a book in this category

CHAPTER 3. METHODOLOGY

5.4 9.0 6.5 3.7 26.0 8.8 12.0 4.9 15.1 37.8

4.1 3.2 2.2 0.2 18.3 0.0 5.2 8.0 3.9 66.9

Table 3.3: Pupils’ reading habits

frequency 100 80 60 40 20 0 0 1 2 3 4 5 6 7 8 Figure 3.5: Histogram of reading cultural capital score

CHAPTER 3. METHODOLOGY

3.3.2

66

Viewing Habits

Television viewing was included in the questionnaire because this is an important part of the “cultural consumption” of most people. It also has the advantage (like books and unlike theatre, art galleries, etc.) of not being any more inaccessible to those who do not live in London than to those who do. Pupils were asked “Which TV programmes do you watch regularly? Name as many as you can.” Having asked pupils to list the television programmes they watched, I was faced with the task of categorising these programmes according to their cultural capital content. This is by its nature a somewhat subjective task, and one feels less confident in categorising TV programmes than categorising books, as the categories for TV programmes are not so well established and accepted. There is no debate about which fiction counts as “classic” (even if there is debate about whether a classic author is actually any good). Even modern fiction is put into pretty clear categories by such indicators as the publisher and the colour of the book’s spine, and whether it is reviewed in the quality papers. The initial judgements about these books (made by publishers, journalists, etc.) were perhaps no less subjective than those I am now making about TV programmes. However, I do not have to construct the categorisation of books myself. This categorisation is in a sense a social fact. Since TV programmes are not publicly categorised in the same way as books, there is more scope for disagreement with my categories. I looked through some of the questionnaires and watched a bit of any programme that I could not categorise initially. The coding I used, and the frequencies and proportions of pupils who had watched a programme in each of the categories, are shown in Table 3.4. I have divided the factual programmes according to their subject matter, “science”, “arts” and “politics/humanities”; and according to their level of sophistication. By “level of sophistication” I mean how in depth the programme is and how much attention it requires of the viewer. The judgements that I have made about factual programmes are, of course, open to question, but it is my judgements about non-factual programmes that I am more concerned about. I think that people are more likely to disagree in their judgements on non-factual programmes, as it is hard when considering these programmes not to be swayed by the question of whether or not one finds the programme enjoyable. The only non-factual category that I have divided into sophisticated/non-sophisticated is comedy. I have tried not to judge this distinction according to how funny I find particular

CHAPTER 3. METHODOLOGY Category *Science — sophisticated *Science — unsophisiticated *Arts — sophisticated

*Arts — unsophisticated *Politics, current affairs, society and humanities — sophisticated *Politics, current affairs, society and humanities — unsophisticated Other factual *Literary adaptations

*Comedy — sophisticated Comedy — unsophisticated

*Educational — sophisticated *Educational — unsophisticated Drama Science Fiction Soaps Teen/kids Sport

Examples QED, Horizon, Equinox Wildlife programmes South Bank Show, Late Review, Arena, Without Walls Film ’98 Bookworm Newsnight, Panorama, World in Action, Everyman The news (except Newsnight), Witness, The Big Story Have I got News for You, Rory Bremner Crimewatch, Watchdog, Food and Drink Tom Jones, Dance to the Music of Time, Pride and Prejudice Frasier, The Simpsons, King of the Hill Friends, Men Behaving Badly, Only Fools and Horses, Mr Bean Open University The learning zone Casualty, ER, The Bill Star Trek, X Files, Red Dwarf Brookside, Eastenders, Neighbours Hollyoaks, Blue Peter, The Fresh Prince of Bel Air football, athletics

Table 3.4: Coding of television programmes

67 Frequency 9 28

Percentage 2.0 6.2

3

0.7

4 11

0.9 14.7

66

14.7

99

22.0

1

0.2

133

29.6

228

50.8

4

0.9

3

0.6

177 126

39.4 28.1

307

68.4

313

69.7

130

29.0

CHAPTER 3. METHODOLOGY

68

programmes. Instead, I have tried to focus on the subtlety of the humour, the cultural references used, and the vocabulary used. To illustrate, let me take two programmes, Frasier and Shooting Stars. In Frasier, the humour is based on dialogue. Sophisticated vocabulary is used. In parodying the pretensions of the main characters, “high-brow” cultural references are made (to Freud, the opera, etc). Shooting Stars, on the other hand, relies on slapstick to a large degree. The vocabulary used is not sophisticated. The cultural references are popular (often to pop music). I considered the possibility of dividing drama into the categories of highbrow and popular. Clearly, a drama such as The Singing Detective would fit into the former category, with things like The Bill and Casualty in the latter. However, going through the TV schedules, I found that there were no high-brow drama series on at the time I was conducting the survey (if one excludes literary adaptations, which form their own category). As in the case of books, and for the same reason, when in doubt about whether to categorise a programme as low or high culture, I have erred on the side of high culture. For instance, I have categorised Have I got News for You as unsophisticated politics/current affairs, rather than as unsophisticated humour, thus giving credit for cultural capital. Having broken certain categories down into “sophisticated” and “unsophisticated”, I found that the frequencies of pupils watching these programmes were too low to justify this subdivision. For instance, only 3 pupils stated that they watched sophisticated arts programmes, and 4 that they watched unsophisticated arts programmes. So, in the final coding, these categories were collapsed. Pupils were given a “TV viewing” score, consisting of points awarded for watching any programme in each of the following categories: science, arts, politics, educational, literary adaptations, sophisticated comedy, i.e. a mark out of six.

3.3.3

Formal Culture and Music

Pupils were asked:

CHAPTER 3. METHODOLOGY

69

How often do you do each of these spare time activities (not organised by the school) — (Often, Sometimes, Hardly ever, Never). 1. Going to art galleries or museums 2. Going to see plays 3. Going to classical concerts 4. Playing an Instrument 5. Listening to classical music

As expected, none of these activities was very popular among my sample. Activity

Often

Art galleries and museums Plays Classical concerts Plays Instrument Classical listening

Hardly Ever 33.6

Never

0.9

Sometimes 13.8

4.9 0.4 14.5 3.9

17.0 4.5 14.0 9.6

38.5 16.9 13.2 16.8

39.6 78.1 58.3 69.7

51.7

Table 3.5: Formal culture and music The responses for all these items are heavily skewed towards the “never” end of the scale. This backs up my view that such questions are not a good way of differentiating between 16 year old respondents (if indeed they are a good tool for differentiating between levels of cultural participation of the adult population). Since more than half the respondents stated that they never went to art galleries or museums, any other response (often, sometimes, or hardly ever) was given one point for cultural capital. This pattern was followed for the other items. In the case of going to plays, more than half responded “hardly ever” or “never”, so a point was given for “often” or “sometimes”.

CHAPTER 3. METHODOLOGY

3.3.4

70

Parents’ Cultural Capital

Parents’ cultural capital was measured on a scale composed of items on: the number of books in the home, whether a broadsheet newspaper is read, subjects discussed in the home, participation in cultural activities, type of music listened to and radio stations listened to. The questions on subjects discussed in the home were as follows.

Which of the following have you heard your parents discuss? — (Often, Sometimes, Never). 1. Art 2. Politics 3. Books 4. Science 5. Current Affairs

A point for cultural capital was given if the response put the respondent in the top less than 50% of the sample. The questions on activities were as follows:

CHAPTER 3. METHODOLOGY

71

For each of these activities, do your parents do it often, sometimes, rarely or never? • Going to art galleries or museums • Reading novels • Reading non fiction • Going to see plays • Going to concerts • Playing a musical instrument • Evening or daytime classes

The same pattern was followed in giving points for cultural capital — for each item, a point was awarded if the level of participation put the respondent in a minority, at the higher end of the spectrum. Pupils’ were asked “Do your parents listen to the radio? If so, which station(s) do they listen to?” A mark for cultural capital was given for any one of the following stations: Radio 3, Radio 4, Classic FM, Jazz FM, and the World Service. Pupils were asked “If your parents listen to music, what kind(s) of music do they listen to?” Responses that fell within the categories of classical or jazz were given a point for cultural capital. Points for cultural capital were also given for having two or more cases of books in the home, and for taking a broadsheet newspaper. All the above items were summed to give an overall measure of parents’ cultural capital — a score out of 16.

3.3.5

Validity of Cultural Capital Questions

Criterion validity concerns whether a measure accurately reflects the concept that it is intended to reflect. Of course, one major problem with assessing

CHAPTER 3. METHODOLOGY

72

the validity of a measure in sociology is that the theoretical concepts one is dealing with are rarely precisely defined. This may be seen as a particular problem in attempting to construct a measure of cultural capital. In addition, the assessment of criterion validity requires an independent measure of the sociological construct in question. I certainly have no independent measure of the cultural capital of my sample of pupils. It could perfectly well be claimed that my operationalisation leaves out important elements of cultural capital, or that it includes items that do not really reflect cultural capital. The vagueness of the concept leaves plenty of scope for disagreement. However, I would argue that my aim is not simply to measure some received and theoretically pre-defined notion of cultural capital and then attempt to correlate it with achievement. Rather, I aim to explore the notion of cultural capital empirically. In order to examine whether the different cultural resources that I have distinguished belong together in the same concept, and deserve the name of capital one must ask whether each of them is associated with attainment. It will also be interesting to examine whether individuals who score highly on one element of cultural capital tend to also score highly on the other two. In other words, does possession of cultural capital usually mean possession of a general culture comprising knowledge, lifestyle and language (as Bourdieu’s writings would seem to suggest) or is possession of cultural capital more fragmentary than this? (These questions reflect construct validity rather than criterion validity). There are specific problems that may be anticipated with some of the questions I have asked to assess possession of cultural capital. Pupils may have a tendency to give the response that they think will be seen as desirable rather than a completely honest response. For instance, pupils may overestimate the frequency with which they participate in certain activities such as reading, listening to classical music, etc. When naming the TV programmes they watch regularly, they may favour more high-brow programmes, even if they do not really watch them regularly. (One girl exclaimed, “Panorama! I’ve seen that once” while filling in this question.) However, although it is possible that the level of these elements of cultural capital will be exaggerated by my survey, this does not mean that one must expect systematic distortion between social groups. E.g. It does not mean that one must expect girls to exaggerate more than boys, or middle class pupils to exaggerate more than working class pupils, or vice versa. Of course, I cannot rule out systematic distortion (perhaps because girls are more eager to please than boys, or because middle class pupils see high brow activities as more desirable than working class pupils do). Even if there is system-

CHAPTER 3. METHODOLOGY

73

atic distortion between social groups, this would not imply distortion in the patterns of association between variables. I computed scores for parental cultural participation and pupils’ cultural participation, giving a point for each item on the questionnaire that was designed to be an indicator of cultural capital. The reliability coefficient Cronbach’s α (Carmines and Zeller, 1979) for the parents’ (16 item) cultural participation scale is α = 0.84. For the pupils’ (11 item) scale α = 0.67. Note that the items these scales are composed of are not all on the same scale, so the reliability statistic can be seen as a baseline estimate. In the case of pupils, test scores for active and passive vocabulary and cultural knowledge were also calculated. For the cultural knowledge score (25 items), α = 0.87. The active and passive vocabulary scores were summed to create an overall vocabulary score (20 items), for which α = 0.84. All the above scales are approximately normally distributed. (Of course, these figures only reflect internal reliability, or the level of association between items on the scale. This is not to be confused with criterion validity).

3.3.6

Rational Choice Questions

In order to assess various rational choice theories of educational inequalities, it is crucial to examine the beliefs and motivations of pupils. Pupils’ beliefs about their own abilities were assessed using two questions, “What grades do you think you are likely to get in your GCSEs?”, and “In general, how do you rate your academic abilities as compared to other pupils at your school?”(these were both multiple choice questions). Pupils’ attitudes to education were assessed using a scale of 16 statements about studying. Pupils were asked to give a response ranging from “agree strongly” to “disagree strongly”. Items reflected both the labour market value of education, e.g. “The more qualifications you get, the better the job you are likely to get”, and the intrinsic or personal value of education e.g. “I enjoy studying”. Pupils were asked both “What sort of job would you like to be doing in 10 years time”, and “What sort of job do you think you will actually be doing in 10 years time”. These were open response questions. The coding scheme for this question was determined by the pupils’ responses. (See appendix D for the coding frame). These responses can only be allocated to social class

CHAPTER 3. METHODOLOGY

74

categories very approximately. Pupils were also asked “In choosing a job, what things about it do you think are important”. This was a multiple choice response question.

3.3.7

Reliability of Rational Choice Questions

The questions on pupils’ expectations and aspirations, both educational and occupational, would be invalid if pupils failed to put down what they truly expect and aspire to. For instance, false modesty might prevent those who expect to get mostly A grades at GCSE from saying so, or bravado might prevent those who expect to fail from saying so. Similarly, pupils might not be willing to admit that their aspirations are as high or as low as they really are. However, given the confidentiality of the questionnaire, there is no reason for pupils to present false expectations and aspirations. (Of course, the fact that there is no good reason for distorting one’s answers will not stop some people. For example, one girl said to me that she thought it would be “tempting fate” to say she expected to get Bs and Cs.) The questions about pupils’ attitudes towards education may be subject to the problem of pupils giving a response that they perceive to be desirable rather than a truthful response. Again, the confidentiality of the questionnaire should allay this fear to some extent. I have avoided the problem of “yea saying” by using a mixture of positive and negative statements. These questions were used to construct two scales for pupils’ evaluation of the intrinsic value of education, and of the instrumental value of education in the job market. The reliability for the whole 16-item scale is α = 0.80. The reliability for the 7-item sub-scale for attitudes to education as an intrinsic good is α = 0.73. The reliability for the 5-item sub-scale for attitudes to education as an instrumental good is less high, at α = 0.43. Nevertheless, I think that use of this sub-scale is justified in that it meets my theoretical purpose. Also, one must bear in mind that reliability analysis is based on an assumption that the responses for each item are normally distributed. This assumption is strongly violated in the case of this educational attitude scale.

CHAPTER 3. METHODOLOGY

3.3.8

75

Parents’ Social Class and Qualifications

Parents’ social class was determined from pupils’ responses to the following questions regarding each parent. • Please tell me about your parents’ jobs. If they are not working at the moment, please tell me about their most recent jobs. • What is the name of the job? • What kind of work do they do? • What sort of organisation do they work for? • Do they have their own business? The responses were coded using the Goldthorpe class schema and the Standard Occupational Classification (Office of Population Censuses and Surveys, 1991) class schema. The main reason for using the Goldthorpe schema was comparability with other quantitative work in the field of sociology of education, and this is the schema that I have used in most of the analysis. I collapsed the schema into six categories in order to avoid small cells, due to my small sample size.1 I took mother’s or father’s class, whichever was the higher, as determined by a simplified version of Erikson’s (Erikson, 1984) dominance schema2 . I tried using a composite measure such as that suggested by Heath and Britten (Heath and Britten, 1984), and found that the model fit was generally extremely similar to that gained using the dominance approach.3 The social class breakdown of the sample is given in Table 3.6. 1

• I = 1 (service class, higher) • II = 2 (service class, lower) • IIIa + IIIb = 3 (non-manual) • IVa+IVb + IVc = 4 (small proprietors) • V + VI = 5 (skilled manual) • VII = 6 (unskilled manual) 2

1 dominates 2, 2 dominates 4, 4 dominates 3, 3 dominates 5, 5 dominates 6. However, the results were harder to interpret because, given my small sample, It was not desirable to increase the number of class categories. Having used a four category schema to derive a ten category composite schema, and then having collapsed these cat3

CHAPTER 3. METHODOLOGY

76

Parents’ qualifications were determined as follows. Pupils were asked “What is the highest qualification that each of your parents has?” A set of multiple choice boxes followed for both “mother” and “father”. The options were: • No qualifications • CSE • Olevel • A level/Scottish higher • Vocational qualification (please specify) • Degree • Postgraduate qualification (masters, Ph.D. or D. Phil.) Vocational qualifications that were equivalent to a degree or postgraduate qualification were recoded as such. Mother’s or father’s qualifications were selected according to which was higher. The resulting schema was then collapsed into a four-category schema. 1. No qualifications or CSE, 2. O level, 3. Intermediate (A level or vocational), 4. Degree (undergraduate or postgraduate).

3.4

Sample

I surveyed 465 pupils. Clearly, from the point of view of minimising standard error, the larger the sample the better. However, I was clearly limited by time and other resources. There are two main issues here. Firstly, getting access to schools is difficult and takes time. Secondly, the data must be coded and inputted. I would have had difficulty both in getting a much larger sample, and in dealing with it once I had it. I have carried out my survey at two Oxfordshire comprehensives, and two London comprehensives. Both cultural reproduction theory and rational choice theory are concerned with general processes, which are not contingent on any particular school context. Therefore, these theories could be egories further to derive a six class schema, the meaning of the resulting classes was far from intuitive.

CHAPTER 3. METHODOLOGY

77

tested using any type of school. While a representative sample of the national year 11 population would be ideal, it would clearly be impossible for me to collect a sufficiently large sample to achieve this. What is really crucial for me is to have a good spread of pupils in terms of ability, social class, etc., rather than to have a sample that is representative. The schools I have chosen will inevitably be unrepresentative in some respects. However, it should be borne in mind that I am not attempting to make population estimates, but rather to examine processes that the theories under examination suggest should operate right across the educational system The reason for only using comprehensive schools is that, while I want my sample to contain a good spread in terms of characteristics such as social class and ability, I do not wish to examine school type effects, and therefore prefer to keep the variable “school type” constant as far as possible. Of course, the problem with this is that one cannot get a representative sample while excluding certain types of school. Pupils attending private schools in particular represent a significant section of the student population, and may be expected to differ from state school pupils in many relevant respects (e.g. the importance of cultural vs. economic capital in their academic attainment). However, given the necessarily small size of my sample, I felt that school type effects really did have to remain outside the scope of my thesis as far as possible. Furthermore, the comprehensive sector is extremely important in its own right, as a large majority of British secondary school pupils attend comprehensive schools. School 1 is a voluntary aided boys’ comprehensive (126 boys in year 11). School 2 is school 1’s “sister school”. It is a girls’ comprehensive which maintains a joint sixth form with school 2. School 3 is the pilot school. School 4 is a large comprehensive in Oxfordshire (213 pupils in year 11) that takes both boys and girls. Clearly, it is important for my sample to be reasonably mixed between boys and girls, given that some of my hypotheses concern gender. I have achieved this balance by using one boys’ and one girls’ school, as well as two co-ed schools. It might be argued that, ideally, I should have used only co-ed schools in order to avoid the complication of the possible “school effects” of single sex schools. However, the issue of possible differences in attainment between pupils at mixed and single-sex schools is highly contentious. Even if there were large differences in attainment, this would not imply that the relationships between the variables I am considering would vary between single sex and mixed schools. In addition, given the difficulties involved in getting access to schools, I think it would have been foolhardy to reject the option of using single-sex schools.

CHAPTER 3. METHODOLOGY

78

The social class breakdown of the schools is given in Table 3.6. School 1 and School 2 have a large proportion of service class families (44.8% and 42.0% respectively), compared to schools 3 and 4 where 14.7% and 30.8% per cent of families respectively are categorised as belonging to the service class. The proportion of families categorised as belonging to the skilled or unskilled manual classed is higher in schools School 3 and School 4(34.6% and 28.6% respectively) than in School 1 and School 2(10.5% and 11.0% respectively). School Parent’s class Missing Service class — higher Service class — lower Routine non-manual Petty bourgeois Skilled manual Manual

1

2

3

4

f 17 26

% 16.2 24.8

f 20 27

% 20.0 27.0

f 8 5

% 10.7 6.7

f 12 20

% 6.5 10.8

Total f % 57 12.3 78 16.8

21

20.0

15

15.0

6

8.0

37

20.0

79

17.0

13

12.4

14

14.0

26

34.7

44

23.8

97

20.9

17

16.2

13

13.0

4

5.3

19

10.3

53

11.4

5

4.8

4

4.0

13

17.3

28

15.1

50

10.8

6

5.7

7

7.0

13

17.3

25

13.5

51

11.0

Table 3.6: Social class breakdown for each school Another factor in choosing which schools to approach was location, since I do not have a car. Getting access to schools was difficult. Although I approached several schools by writing to the head, only School 4 was successfully approached in this way. The other schools were approached through personal contacts.

3.4.1

Administration of questionnaire

I administered a self-completion questionnaire to whole class groups of year 11 pupils. I decided to do this rather than interview pupils individually, since this approach allows one to get a larger sample. The questionnaire took pupils about 45 minutes to complete.

CHAPTER 3. METHODOLOGY

79

In all schools, the questionnaire was administered during a “PSE” (Personal and Social Education) period. Pupils were not allowed to confer while completing the questionnaire, but were able to ask either a teacher or me if they did not understand any of the questions. Pupils and schools were, of course, assured of the confidentiality of their responses. I introduced the questionnaire was there to answer questions while the students were completing the questionnaire. I introduced the questionnaire something along the lines of the following:

My name is Alice Sullivan and I am a research student from Oxford University. The questionnaire I would like you to fill in asks questions about things like your views about education, your plans for the future, and what you do in your spare time. There are also some questions about your family. The reason for these questions is that the circumstances in which people live, and the people they live with can have an important influence on how they do at school and on the decisions they make about what to do after leaving school. Everything in the questionnaire will be treated as confidential. I won’t pass any information about you on to the school or your parents. The only reason that you have to be identified on the questionnaire at all is so that I can follow up on what GCSEs you get and what you do after your GCSEs. Most of the questions are answered by ticking a box. If you have any problems with any of the questions, just put your hand up and ask.

The fact that I had to identify students in order to follow up on their GCSE results and on what they do after GCSEs posed a problem, as it meant that the questionnaires could not be anonymous. There are two alternative ways of dealing with this. One is simply to ask pupils to put their names on the questionnaire. The disadvantage of this is that pupils may be concerned that the questionnaire could be seen by people other than the researcher, who would then easily be able to identify individuals. The other alternative is to number the questionnaires and have a list attaching each pupil’s name to a number. This is the system which I agreed to use at School 4. At first I intended to hand out the questionnaires according to a pre-decided list of names and numbers. However, this did not work out because I could not get a list of which pupils would be there on any given day out of the school. On my first trip to the school, I ended up asking students to write their names on

CHAPTER 3. METHODOLOGY

80

the questionnaires. Thereafter, I numbered the questionnaires and asked the teachers present to make a note of who had been given each number as they handed the questionnaires out. However, even this was rather more timeconsuming than I had expected. I decided that the disadvantage of using named questionnaires was outweighed by the practical problems associated with numbering. Therefore, I was relieved to find that the heads of the other schools had no objection to pupils being asked to provide names.

3.5

Response Rates

In three out of the four schools, the entire year group was surveyed. In the remaining school, for time-tabling reasons, five out of seven forms were surveyed. Out of a potential sample of 557 pupils, 465 questionnaires were adequately completed, giving a response rate of 80%. The bulk of the nonresponse was due to absenteeism (17%), with the remaining 3% due to noncompletion or inadequate completion of the questionnaire. The level of missing data on social class is 12% (57 cases)4 . This is mainly because many students did not respond to the question on their parents’ occupations in sufficient detail for the responses to be categorised. In the case of parents’ qualifications, this problem is still more severe (122 missing cases). This is almost certainly due to pupils’ simply not knowing what qualifications their parents had. This level of missing data certainly reflects a problem with using children as a proxy source of information on their parents. Many children have no clearer idea of what their parents do for a living than “works in an office” for instance, and many children may never have talked to their parents about what qualifications they have. This lack of information may be informative in itself, perhaps revealing a lack of “social capital” within the home in Coleman’s sense (see for instance, (Coleman, 1990)). Therefore, I have included these missing cases within my analyses as separate categories. 4

This is a normal level of non-response in national surveys where people are asked about their parents’ occupations, such as the British Election Survey.

CHAPTER 3. METHODOLOGY

3.6

81

Ethics

The confidentiality of participating individuals and schools has been and will be preserved. No data relating to individuals has been or will be passed on to third parties. Pupils were not asked whether they wanted to complete the questionnaire. However, if any of them had objected to answering specific questions, they would not have had to do so. No one objected verbally, but some pupils did not fill in sections of the questionnaire. One pupil refused to fill in the questionnaire at all, and a few gave false names or no name. I told pupils that I would be looking at their GCSE results, and there were no objections to this. This is perhaps unsurprising given that GCSE results are not generally kept private within schools.

3.7

Conclusions

Responses on the cultural participation of parents and pupils and the linguistic abilities and cultural knowledge of pupils will allow me to assess: • Whether cultural capital is passed down from parents to children • Whether middle class people have higher levels of cultural participation than working class people, and whether male or female respondents have higher levels of cultural participation. • Whether cultural participation affects educational attainment • Which elements of cultural participation are most important in determining attainment • Whether the effect of cultural participation on educational attainment is direct, or mediated by cultural knowledge and linguistic ability Responses on the attitudes and aspirations of pupils will allow me to assess the impact of class and gender on decision making in educational careers. • By comparing pupils’ self-predicted GCSE grades and view of their own academic abilities to actual GCSE results I will assess whether working

CHAPTER 3. METHODOLOGY

82

class and female pupils underestimate their abilities as compared to middle class and male pupils. • By surveying pupils on their assessment of the intrinsic and instrumental value of education, I will be able to assess whether working class and middle class, and male and female pupils differ in their assessments of the worth of education both in the job-market and more generally. • I will be able to assess whether occupational aspirations vary by class and gender once GCSE performance has been controlled for • I will be able to assess whether educational plans vary by social class and gender once GCSE results are controlled for, and to examine possible mechanisms behind this, such as gender and class differences in subjectively perceived ability, attitudes to education, and occupational aspirations.

Chapter 4 Cultural Capital and Educational Attainment at GCSE 4.1

Introduction

This chapter will assess the merits of the cultural reproduction approach to the examination of class and gender differentials in educational attainment. According to Bourdieu’s theory of cultural reproduction, children from middle class families are advantaged in gaining educational credentials due to their possession of cultural capital. In order to assess this theory, I have developed a broad operationalisation of the concept of cultural capital, and have surveyed pupils on both their own and their parents’ cultural capital. In chapter 1 I argued that for Bourdieu’s theory to be backed empirically, he would need to show that: 1. parental cultural capital is inherited by children. 2. children’s cultural capital is converted into educational credentials. 3. educational credentials are a major mechanism of social reproduction in advanced capitalist societies. In this chapter, I will address the following questions: 83

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

84

• How cultural capital is distributed according to social class and educational level. • The extent to which cultural capital is passed down from parents to children. • Whether male and female pupils possess different levels of cultural capital. • What effect cultural capital has on GCSE attainment at age 16.

4.2

Analysis

I have used linear regression to examine the associations between pupils’ and parents’ cultural capital, and between cultural capital and GCSE attainment. I have used nested models in order to show the pattern of effects in greater detail, so that both direct and indirect effects of explanatory variables on the outcome variable are apparent. The order in which variables have been inserted into these nested models is to some degree arbitrary, but has been chosen in accordance with the theoretical framework I am using, and the questions I am trying to answer. For instance, social class is always included in the first model because I am interested in seeing whether the effect of social class can be explained by other variables, which are inserted in subsequent models.

4.2.1

Parental Cultural Capital

The first step in assessing the theory of cultural reproduction is to look at the distribution of cultural participation by social class. (The parental cultural participation variable has mean 4.78 and standard deviation 3.89). In figure 4.1 we can see a clear positive relationship between social class and parental cultural participation. Similarly, taking mother’s or father’s educational level, whichever is higher, there is a positive relationship between parents’ qualifications and their cultural participation — see figure 4.2. Parents with degrees have particularly high levels of cultural participation.

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

85

Parents’ cultural participation 10 8 6

........... .... ... ... ............................................ .... ... .... .... .... .... .... ... .... ... .... .... ........... .... ..... .... .... ... ... ............................................. .... .... .... ............................................ .... .... ... .... .. ... ................................................ ............ .... . ..... .... .... .... .... .... .... ... ... .. .... . .... ........................................... .... .... .... ... .... .... .... ... ............................................ .. . . . . . .......... ................................................. ............................................. . . . . ..... ..... ..... ............................................. ... .... .... ................................................ .... ..... .... ... .. . .... .... .... ... ............................................. .... ... .... ... .... ............................................ .... ... .... ... .... .... .... .... .. . . . . ........... ................................................ ........................................... . ..... ... .... .... .... .... ... .... .....................................................................................................................................................................................................................................................................................................................................................................................................................................

r

r

r

r

r

Skilled manual (50)

Manual (51)

Routine non-manual (97)

Petty bourgeois (53)

Service class — lower (79)

Missing (57)

2

r

r

Service class — higher (78)

4

Figure 4.1: Parents’ cultural participation by parents’ class (95% confidence intervals)

4.2.2

Pupils’ Cultural Capital

Having established an association between parental social class and cultural participation, we can move on to the question of whether cultural capital is transmitted within the home. To what extent is parental cultural participation associated with pupils’ cultural participation, controlling for background variables?

Pupils’ Cultural Activities I used linear regression to analyse the determinants of the activities component of pupils’ cultural capital. For each parameter, the B statistic is

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

Intercept Parents’ qualifications Missing Degree Intermediate O-level CSE or no qualifications Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Unskilled manual Sex Male Female School School 1 School 2 School 3 School 4 Parents’ cultural participation N η 2 for the model

Model 1 B(s.e) 4.16(0.42)*** ***

η 0.177 0.047

0.13(0.35) 1.71(0.42)*** 0.43(0.41) 0.53(0.36)

0.000 0.036 0.002 0.005

* 0.14(0.45) 1.35(0.46)**

2

Model 2 B(s.e) 3.33(0.38)***

86

η2 0.148 0.007

−0.12(0.30) 0.41(0.38) 0.24(0.36) 0.21(0.32)

0.000 0.003 0.000 0.001

0.035 0.000 0.019

0.16(0.39) 0.29(0.41)

0.013 0.000 0.001

0.48(0.43)

0.003

−0.88(0.38)

0.000

−0.18(0.40)

0.000

−0.37(0.35)

0.002

0.28(0.46) 0.28(0.46)

0.001 0.001

−0.17(0.40) 0.23(0.40)

0.000 0.001

−0.46(0.28)

0.006 0.006

−0.34(0.25)

0.004 0.004

* 0.45(0.33) 0.41(0.34) −0.66(0.32)*

0.023 0.004 0.003 0.009

0.29(0.29) −0.05(0.30) −0.24(0.28)

0.005 0.002 0.000 0.002

0.35(0.03)***

0.233

464 0.218

Table 4.1: Pupils’ cultural participation

464 0.400

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

87

Parents’ cultural participation 10

........... .... ... ... .... ..................................................................... .... ... .... .... .... .... ... .... .... ... ........... ... ..... ..................................................................... .... ... .... .... .... .... ... .. ............ ..... .... ... ..................................................................... .... ... .... .... .... .... ..................................................................... .... ... .... ... .... .. .. . . .... . .......... ..................................................................... . . .. . . ..................................................................... ..... ..... ..... . ... .... . . . .... . ... .................................................................... .................................................................. .... ... .... .... ..................................................................... .... .... ... .... ... .... .... . . .................................................................... ............. . ..... ... .... .... .... .... ... .... .....................................................................................................................................................................................................................................................................................................................................................................................................................................

r

8 6 4

r

r

r

Degree (105)

Intermediate (61)

O-level (109)

Missing (122)

2

None/CSE (68)

r

Figure 4.2: Parents’ cultural participation by parents’ educational level (95% confidence intervals) given, followed by the standard error. In addition, the η 2 statistic is shown. This describes the proportion of total variability in the dependent variable attributable to the variation in the independent variable. It is the ratio of the between groups sum of squares to the total sum of squares. Significance levels are denoted by * if p ≤ 0.05, ** if p ≤ 0.01, and *** if p ≤ 0.001. Table 4.1 shows two models. Model 1 shows the effects of parents’ qualifications, parents’ class, pupils’ gender and school attended on pupils’ cultural activities. All of these variables except gender have significant effects (at the 0.05 level) in this model. Having a parent who is a graduate and having a parent who is in the higher service class are significantly positively associated with pupils’ cultural activities. Parents’ cultural participation is introduced in Model 2. This shows that parents’ cultural participation (with an η 2 of 0.233) is by far the most important factor in accounting for the variation in

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

88

pupils’ cultural activities. (The η 2 statistic describes the proportion of total variability in the dependent variable attributable to the variation in the independent variable. It is the ratio of the between groups sum of squares to the total sum of squares). Neither social class nor educational credentials are significant once parental cultural participation has been included. This shows that the effect of these background variables on pupils’ cultural activities is entirely mediated by parents’ cultural participation. The effect of school attended is quite insignificant once parental cultural capital is taken into account. The absence of a school effect is important, as a crucial claim about cultural capital is that it is not transmitted by the school. (However, bear in mind the small number of schools in my sample, and that these are all comprehensive schools.) The Pearson correlation between parents’ cultural capital and pupils’ cultural activities is 0.617 (p ≤ 0.001). The strength of this relationship provides support for Bourdieu’s view that cultural resources are strongly transmitted from parents to children.

4.2.3

Language and Knowledge

Next, I modelled the pupils’ tested vocabulary and cultural knowledge scores. Parental cultural participation mediates the background variables to some extent, but not to the same extent as in the case of the activities component of cultural capital. This is unsurprising, as the parental cultural participation score is composed of similar items to the activities component of pupils’ cultural capital, whereas I have no direct measure of parental vocabulary or cultural knowledge. I modelled the pupils’ vocabulary score in stages (see table 4.2), first of all just including the background variables — parents’ qualifications, parents’ class, gender and school. Model 1 shows that gender and school are insignificant. These variables have overall significance values of 0.680 and 0.357 respectively. (By overall significance values I mean the significance value for the variable as a whole rather than for each parameter of the variable.) In Model 2, parental cultural capital is added to the model. This shows that the effects of parents’ social class and qualifications are partially mediated by parental cultural capital. The effect of parents’ qualifications is reduced from an η 2 of 0.041 in model 1 to 0.024 in model 2. The effect of social class

Intercept Parents’ qualifications Missing Degree Intermediate O-level CSE or no qualifications Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Unskilled manual Sex Male Female 0.97(1.55) −0.58(1.54)

−0.28(0.96)

0.055 0.003 0.025 0.006 0.000 0.002 0.000

0.000 0.000

*** −1.78(1.53) 5.34(1.58)**

2.37(1.49)

0.38(1.37)

1.53(1.56) −0.52(1.56)

−0.42(0.98)

0.000 0.000

0.001 0.000

0.000

0.003

0.035 0.003 0.014

0.001 0.013 0.016 0.008

η 0.178 0.023

2

−0.01(0.94)

1.11(1.52) −0.76(1.51)

0.43(1.33)

1.74(1.45)

* −1.89(1.48) 3.80(1.56)*

1.05(1.15) 3.24(1.44)* 3.71(1.37)** 2.13(1.21)

Model 3 B(s.e) 11.67(1.54)***

Table 4.2: Language score (continued overleaf. . . )

0.14(1.35)

1.67(1.48)

* −1.76(1.51) 4.03(1.60)*

0.95(1.17) 3.57(1.47)* 3.73(1.39)** 2.30(1.23)

0.002 0.029 0.020 0.010

1.26(1.19) 5.18(1.42)*** 4.24(1.41)** 2.70(1.24)*

Model 2 B(s.e) 14.32(1.45)*** *

η 0.200 0.038

2

Model 1 B(s.e) 15.35(1.45)*** **

0.000 0.000

0.001 0.001

0.000

0.003

0.035 0.004 0.013

0.002 0.011 0.016 0.007

η 0.113 0.021

2

−0.61(0.97)

0.68(1.56) −0.91(1.51)

0.05(1.33)

0.99(1.46)

−2.26(1.50) 2.90(1.59)

0.86(1.16) 3.06(1.46)* 3.60(1.37)** 2.04(1.21)

Model 4 B(s.e) 12.37(1.57)***

0.001 0.001

0.000 0.001

0.000

0.001

0.028 0.005 0.008

0.001 0.010 0.016 0.007

η2 0.127 0.021

School School 1 School 2 School 3 School 4 Parents’ cultural participation Pupils’ cultural participation Reading Formal cultural particpation Music TV N η 2 for the model

η 0.007 0.006 0.001 0.000

2

464 0.189

0.44(0.12)***

−2.05(1.10) −1.51(1.14) 0.43(1.08)

Model 2 B(s.e)

0.030

η 0.013 0.008 0.004 0.000

2

0.043

0.80(0.18)***

464 0.224

0.003

η 0.016 0.010 0.004 0.001

2

0.16(0.13)

−2.28(1.08)* −1.47(1.12) 0.62(1.05)

Model 3 B(s.e)

Table 4.2: Language score (. . . continued from overleaf)

464 0.164

−1.86(1.12) −0.94(1.15) −0.08(1.08)

Model 1 B(s.e)

0.019 0.000 0.000 0.040

−0.07(0.55) 2.24(0.53)*** 448 0.239

0.008

η2 0.020 0.011 0.005 0.002

0.71(0.24)** −0.01(0.49)

0.26(0.14)

Model 4 B(s.e) * −2.36(1.10)* −1.65(1.13) 0.88(1.07)

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

91

is reduced from an η 2 of 0.065 to 0.043. I am referring here to the η 2 statistic for the variables as a whole, i.e. the proportion of total variability in pupils’ vocabulary scores attributable to the variation in parents’ qualifications and social class. Model 3 shows that the effect of parental cultural capital is in turn mediated by the activities component of pupils’ cultural capital. This leaves parental social class and pupils’ cultural activities accounting for very similar proportions of the variation in pupils’ language score. The next step is to break the measure of pupils’ cultural activities down into its constituent parts in order to determine which cultural activities are associated with pupils’ vocabulary score. (The overall score is removed from the model before inserting the component scores). On Crook’s (Crook, 1997) view that public cultural participation serves to communicate status, whereas reading helps to develop abilities, reading should be positively associated with vocabulary and formal culture should not. And indeed, this is the case. But reading is not the only form of cultural participation that is positively and significantly associated with pupils’ vocabulary. In fact TV viewing habits account for a greater proportion of the variation in pupils’ vocabulary than does reading. The “music” variable however, (whether a pupil listens to classical music and/or plays an instrument) is not significant. Using the same procedure for pupils’ cultural knowledge (see table 4.3), Model 1 shows the background variables. In this model, gender is insignificant, but parents’ qualifications, parents’ social class and school attended are all highly significant. Higher service class backgrounds are significantly associated with cultural knowledge. Graduate parents are particularly strongly associated with cultural knowledge, but intermediate and O level qualifications are also significant. Model 2 shows that again, parental cultural participation partially mediates the background variables. The effect of a higher service class background is rendered insignificant in this model, and the overall effect of parents’ qualifications is reduced from η 2 = 0.073 to η 2 = 0.031. Model 3 shows that the effect of parental cultural participation is itself partially mediated by the activities component of pupils’ cultural capital. However, the direct effect of parents’ cultural participation is still highly significant in this case. Gender becomes significant, with a small advantage in favour of boys, once pupils’ cultural activities are included in the model. This means that all the variables in Model 3 have significant effects. Again, I broke down pupils’ cultural activities to see which elements of this measure are actually doing the work. I found the same pattern as for pupils’

Intercept Parents’ qualifications Missing Degree Intermediate O-level CSE or no qualifications Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Unskilled manual 0.008 0.067 0.014 0.015

0.065 0.012 0.015 0.006 0.004 0.003 0.000

1.15(0.62) 4.25(0.75)*** 1.86(0.74)* 1.71(0.65)**

*** −1.91(0.80)* 2.20(0.83)**

1.27(0.78)

0.94(0.72)

1.01(0.82) 0.04(0.82)

0.56(0.79) −0.01(0.79)

0.75(0.69)

0.70(0.76)

** −1.90(0.77)* 1.14(0.82)

0.90(0.60) 2.94(0.75)*** 1.45(0.71)* 1.38(0.63)*

Model 2 B(s.e) 6.27(0.74)*** **

0.001 0.000

0.003

0.002

0.043 0.013 0.004

0.005 0.033 0.009 0.011

η 0.137 0.034

2

0.62(0.78) −0.10(0.77)

0.89(0.68)

0.73(0.74)

** −1.96(0.76)** 1.02(0.80)

0.95(0.59) 2.79(0.74)*** 1.44(0.70)* 1.30(0.62)*

Model 3 B(s.e) 5.00(0.79)*** **

0.001 0.000

0.004

0.002

0.048 0.015 0.004

0.006 0.031 0.009 0.010

2

η 0.082 0.031

Table 4.3: Cultural knowledge scores (continued overleaf. . . )

η 0.164 0.072

2

Model 1 B(s.e) 7.11(0.76)*** ***

0.52(0.80) −0.10(0.78)

0.81(0.69)

0.62(0.75)

** −2.06(0.77)** 0.98(0.82)

0.98(0.60) 2.58(0.75)** 1.43(0.70)* 1.28(0.62)*

Model 4 B(s.e) 5.08(0.81)*** *

0.001 0.000

0.003

0.002

0.047 0.016 0.003

0.006 0.027 0.010 0.010

η2 0.085 0.027

Sex Male Female School School 1 School 2 School 3 School 4 Parents’ cultural participation Pupils’ cultural participation Reading Formal cultural particpation Music TV N η 2 for the model 0.045 0.000 0.015 0.020

η 0.004 0.004

2

0.074

0.36(0.06)***

464 0.338

0.025 0.000 0.008 0.012

η 0.006 0.006

2

** 0.01(0.56) 1.09(0.58) −1.29(0.55)*

0.80(0.49)

Model 2 B(s.e)

0.038

0.38(0.09)***

464 0.363

0.024

0.024 0.000 0.008 0.011

η 0.008 0.008

2

0.22(0.07)**

* −0.10(0.56) 1.11(0.57) −1.19(0.54)*

0.93(0.49)

Model 3 B(s.e)

Table 4.3: Cultural knowledge scores (. . . continued from overleaf)

464 0.286

*** 0.17(0.59) 1.56(0.60)** −1.70(0.57)**

0.68(0.51)

Model 1 B(s.e)

0.035 0.000 0.001 0.022

−0.15(0.28) 0.85(0.27)** 448 0.369

0.029

0.022 0.000 0.006 0.012

η2 0.005 0.005

0.49(0.12)*** −0.10(0.25)

0.25(0.07)***

* −0.26(0.57) 0.92(0.58) −1.25(0.55)*

0.70(0.50)

Model 4 B(s.e)

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

94

vocabulary. Reading has a significant association with pupils’ cultural knowledge. Participation in formal culture does not. The “music” variable is insignificant, whereas television viewing habits are significant. These findings support the view that participation in formal or public culture does not foster the intellectual resources that may give an advantage at school, and that reading does foster these resources. However, reading is not the only cultural activity that is associated with linguistic ability and cultural knowledge. Watching relatively sophisticated programmes on TV is also associated with these skills. Of course, these associations cannot tell us whether reading and watching sophisticated TV programmes foster knowledge or whether pupils’ reading and TV viewing habits simply reflect their level of intellectual ability. It seems highly likely that both of these processes occur. Ideally, one would control for measured ability at a given age (say 11 or younger) and then examine whether cultural participation has an effect on later performance in tests of ability and examinations controlling for the earlier ability score.

4.2.4

Gender

Gender does not account for a significant proportion of the variance in pupils’ activities or pupils’ vocabulary score, and only has a significant effect on pupils’ cultural knowledge once participation in cultural activities is controlled for. However, there are small differences in the average level of cultural capital of girls and boys. These differences generally favour girls as is shown in table 4.4. Male Female Total

µ σ µ σ µ σ Max

Reading 3.35 1.73 3.46 1.70 3.40 1.71 8

Other Activities 1.40 1.28 1.72 1.30 1.55 1.30 9

Knowledge 10.22 4.67 10.05 5.07 10.14 4.86 25

Language 18.10 8.56 18.86 8.87 18.46 8.71 40

Table 4.4: Gender differences in cultural capital (µ=mean, σ=standard deviation)

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

95

Girls have slightly more cultural capital than boys in terms of both reading and other activities, and score more highly on the language test. Boys, however, slightly outperform girls on the test of cultural knowledge. Of these differences, only the difference in cultural activities other than reading is actually significant at the 0.05 level.

4.2.5

GCSE attainment

Finally, what impact does cultural capital have on grades achieved in the GCSE examinations? I have modelled GCSE results using a point score for the total of GCSEs gained — giving 1 point for a G grade, 2 for an F etc.. This point score is approximately normally distributed. The effects of the background variables on pupils’ GCSE scores are shown by Model 1 (see table 4.5). Compared to unskilled manual backgrounds, all nonmanual backgrounds are associated with increased GCSE performance, with higher service class backgrounds providing the strongest advantage. Parents’ qualifications in the intermediate category (A level or vocational) and at degree level were significantly associated with GCSE scores. Family structure and the number of siblings a pupil has are also included in this model. (These variables have been excluded from previous analyses simply due to a desire not to overload the analyses with variables.) Being in an “intact” family (i.e. two original parents as opposed to a step- or single parent family) is quite significantly positively associated with GCSE performance. The number of siblings has a significantly negative effect (i.e. the greater the number of siblings, the lower the average GCSE performance). Model 2 shows that the effects of the background variables are mediated to an extent by parental cultural participation. For instance, the overall effect of parents’ class on pupils’ GCSE attainment is reduced from an η 2 of 0.094 in Model 1, to an η 2 of 0.067 in Model 2. The effect of having a graduate parent is rendered insignificant in this model, although intermediate qualifications are still significantly positive. Model 3 shows that the effect of parental cultural participaiton on pupils’ GCSE scores is partially mediated by the activities component of pupils’ cultural capital. In this model, all the variables are significant. Breaking down pupils’ cultural activities into formal, reading, music and TV, as before, we can see that the effect of reading is significant, and the effect of participation in formal culture is insignificant. TV viewing habits are also significant (though just barely at the 0.05 level), and music is not significant.

Intercept Parents’ qualifications Missing Degree Intermediate O-level CSE or no qualifications Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Unskilled manual Sex Male Female

η 0.145 0.057

0.006 0.012 0.016 0.002

0.094 0.000 0.060

0.036

0.023

0.023 0.003

0.017 0.017

−3.76(2.36) 6.40(2.85)* 7.48(2.81)** 2.22(2.47)

*** 0.18(3.08) 16.65(3.17)***

11.99(2.99)***

8.84(2.76)**

10.16(3.19)** 3.67(3.14)

** −5.41(1.95)**

2

Model 1 B(s.e) 27.42(3.21)*** ***

0.016 0.016

0.018 0.003

0.021

0.028

0.067 0.000 0.039

0.009 0.002 0.011 0.001

η 0.122 0.046

2

* −4.50(1.87)*

9.26(3.06)** 3.31(3.00)

8.76(2.64)**

10.54(2.88)***

*** −0.16(2.94) 13.03(3.10)***

−4.49(2.26)* 1.68(2.84) 5.85(2.70)* 1.10(2.37)

Model 3 B(s.e) 19.53(3.32)*** **

0.013 0.013

0.021 0.003

0.025

0.030

0.073 0.000 0.039

0.009 0.001 0.011 0.001

η 0.075 0.044

2

Table 4.5: GCSE results (continued overleaf. . . )

** −5.07(1.90)**

8.77(3.13)** 3.50(3.06)

8.15(2.69)**

10.29(2.94)**

*** −0.05(3.01) 13.33(3.16)***

−4.59(2.31)* 2.48(2.89) 6.04(2.76)* 1.37(2.42)

Model 2 B(s.e) 24.62(3.18)*** ***

** −5.26(1.94)**

8.77(3.19)** 3.29(3.05)

8.21(2.69)**

9.68(2.94)**

*** −0.06(3.02) 11.65(3.18)***

−4.60(2.31)* 1.92(2.90) 5.78(2.74)* 1.16(2.40)

Model 4 B(s.e) 20.58(3.40)*** **

0.018 0.018

0.018 0.003

0.022

0.026

0.060 0.000 0.031

0.010 0.001 0.011 0.001

η 0.081 0.045

2

*** −6.08(1.56)***

8.12(2.55)** 4.57(2.49)

7.97(2.20)***

9.18(2.39)***

** 3.22(2.46) 9.94(2.59)***

−6.01(1.88)** −2.90(2.38) 2.37(2.26) −1.33(1.97)

Model 5 B(s.e) 3.95(2.97) ***

0.034 0.034

0.023 0.008

0.030

0.033

0.051 0.004 0.033

0.023 0.003 0.003 0.001

η2 0.004 0.049

School School 1 School 2 School 3 School 4 Family structure Family intact Step or single parent family Sibling number Parents’ cultural participation Pupils’ cultural participation Reading Formal cultural particpation Music TV Cultural knowledge Language N η 2 for the model

0.043 0.043

0.015

*** 6.58(1.50)***

−1.08(0.43)*

449 0.364

η 0.050 0.008 0.000 0.030

2

Model 1 B(s.e) *** 4.11(2.24) −0.75(2.29) −7.93(2.15)***

0.009 0.052

0.041 0.041

η 0.038 0.006 0.002 0.022

2

0.044

1.56(0.35)***

449 0.423

0.011 0.011

0.042 0.042

η 0.034 0.005 0.002 0.021

2

−0.92(0.41)* 0.58(0.26)*

*** 6.19(1.43)***

Model 3 B(s.e) ** 3.07(2.14) −2.24(2.20) −6.24(2.07)**

0.000 0.013

0.27(1.08) 2.48(1.06)*

434 0.417

0.032 0.001

0.012 0.019

0.037 0.037

η 0.033 0.003 0.004 0.022

2

1.79(0.49)*** −0.46(0.97)

−0.95(0.42)* 0.79(0.28)**

*** 5.79(1.46)***

Model 4 B(s.e) ** 2.37(2.21) −2.95(2.26) −6.51(2.12)**

Table 4.5: GCSE results (. . . continued from overleaf)

449 0.397

−0.82(0.42) 1.14(0.23)***

*** 6.25(1.46)***

Model 2 B(s.e) ** 3.59(2.18) −2.26(2.25) −6.63(2.11)**

0.103 0.116

0.66(0.09)*** 449 0.604

0.007

0.000 0.005

0.051 0.051

η2 0.054 0.019 0.006 0.018

1.20(0.17)***

0.51(0.30)

0.04(0.35) 0.31(0.22)

*** 5.70(1.19)***

Model 5 B(s.e) *** 5.22(1.79)** −3.04(1.85) −4.82(1.74)**

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

98

This follows the pattern that was seen in modelling pupils’ linguistic ability and cultural knowledge. Previously, I stated that, if participation in cultural activities is linked to examination success, this may be due to the development of knowledge or a set of competencies. Including scores for vocabulary and cultural knowledge in the model, we can see that the effects of parents’ and pupils’ cultural particiption on GCSE attainment are indeed mediated in this way. In Model 3, parents’ cultural participation and pupils’ cultural participation both have highly significant effects. Once pupils’ vocabulary and cultural knowledge scores are included, in Model 4, both parents’ and pupils’ cultural activities become insignificant. So, Model 4 shows very strong effects for both vocabulary and cultural knowledge, leaving no significant direct effects for parents’ cultural participation or pupils’ culturalparticipation. That the effects of these variables are entirely mediated by cultural knowledge and language ability is striking given that this is not the case for parental social class, which remains highly significant after the knowledge and language variables are added to the model. This suggests that the mechanism through which cultural participation improves educational attainment is in fact the possession of knowledge or a set of competencies, whereas the effect of social class cannot be explained in this way. The effect of parents’ qualifications in the intermediate category is rendered insignificant by the inclusion of knowledge and language scores in the model. However, the negative effect of the missing category becomes highly significant in this model. It is possible that this reflects a lack of communication about and interest in education in homes where the children do not know what their parents’ educational level is as compared to those homes where the children know that their parents have no qualifications. (Although CSE and no qualifications are grouped together in this model, the same results emerged using only “no qualifications” as the comparison category.) The difference in GCSE scores in favour of girls cannot be explained by gender differences in cultural capital. The proportion of the variance explained by gender decreases by only a tiny amount (from η 2 0.020 to 0.018) when the activities component of pupils’ cultural capital is added to the model, and the gender effect actually increases once the knowledge and language scores are included.

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

99

Standard Occupational Classification class schema So far, I have used the Goldthorpe class schema in my analyses. One disadvantage of this schema from the point of view of assessing Bourdieu’s theory of cultural reproduction is that it does not have a separate professional class category. It may be that professionals, as opposed to owners and managers, have a distinctive dependence on cultural mechanisms for reproducing privilege. Using the SOC (Standard Occupational Classification) class schema, which has a separate professional class category, will allow me to assess this possibility. I have collapsed the SOC classes into 5 categories as follows. 1. Managers and employers: 1.1, 2.1, 1.2, 2.2, 13 2. Professionals: 3, 4 3. Non-manual and own account: 5, 6, 12, 14 4. Skilled manual: 8, 9 5. Semi- and Unskilled manual: 10, 11, 7, 15, 16 As before, I have determined parents’ social class according to whichever is “higher” — mother’s or father’s class. The order of dominance I have used is 1 dominates 2, 2 dominates 3, etc. (I appreciate that I could equally have categorised professionals as being “higher” than employers and managers. However, I chose to rank employers and managers above professionals as the former group are in positions of control over others.) Professional workers are defined as “persons engaged in work normally requiring qualifications of university degree standard”. Professional parents had a higher level of cultural participation (mean score 9.2, standard error 3.9) than parents who were employers or managers (mean score 5.4, standard error 4.0). In table 4.6 I model GCSE performance as before, but using SOC classes instead of Goldthorpe classes. Model 1 shows that that pupils’ whose parental class is defined as professional have a stronger advantage in GCSE performance than pupils’ whose parental class is defined as “manager or employer”. This is despite the fact that parental qualifications are controlled for in this model, and the fact that professionals are defined as being in work that generally requires a degree. Of course, many graduates are not in professional employment, and it may be that the classification of occupations generally

Intercept Parents’ qualifications Missing Degree Intermediate O-level CSE or no qualifications Parents’ SOC class Missing Employers and managers Professional Non-manual and own account Skilled manual Semi and unskilled manual Sex Male Female 0.006 0.010 0.021 0.003

0.091 0.001 0.008 0.054 0.028 0.000

0.020 0.020

−3.90(2.38) 6.11(2.88)* 8.44(2.80)** 2.94(2.46)

***

−2.29(2.87) 5.58(3.08)

16.94(3.41)*** 8.06(2.28)***

0.66(3.00)

** −5.80(1.95)**

** −5.36(1.90)**

0.45(2.92)

13.62(3.39)*** 6.84(2.23)**

−2.35(2.80) 4.47(3.00)

***

−4.70(2.32)* 1.99(2.93) 6.76(2.74)* 2.00(2.40)

Model 2 B(s.e) 27.12(2.97)*** ***

0.018 0.018

0.000

0.036 0.021

0.002 0.005

0.067

0.009 0.001 0.014 0.002

η 0.161 0.052

2

** −4.82(1.87)**

0.55(2.86)

12.70(3.33)*** 7.11(2.19)**

−2.56(2.74) 4.48(2.95)

***

−4.61(2.28)* 1.42(2.88) 6.58(2.69)* 1.77(2.35)

Model 3 B(s.e) 22.38(3.13)*** ***

0.015 0.015

0.000

0.033 0.024

0.002 0.005

0.068

0.009 0.001 0.014 0.001

η 0.106 0.051

2

Table 4.6: GCSE results using SOC class schema (continued overleaf. . . )

η 0.190 0.065

2

Model 1 B(s.e) 30.15(2.99)*** ***

*** −6.24(1.57)***

2.21(2.39)

8.03(2.80)** 5.72(1.83)**

0.98(2.31) 6.79(2.47)**

**

−6.00(1.90)** −2.88(2.42) 2.76(2.27) −0.77(1.97)

Model 4 B(s.e) 6.14(2.86)* ***

0.035 0.035

0.002

0.019 0.022

0.000 0.017

0.036

0.023 0.003 0.003 0.000

η2 0.011 0.051

School School 1 School 2 School 3 School 4 Family structure Family intact Step or single parent family Sibling number Parents’ cultural participation Pupils’ cultural participation Cultural knowledge Language N η 2 for the model 0.044 0.044

0.016

*** 6.75(1.50)***

−1.12(0.43)**

449 0.396

−0.85(0.42)* 1.15(0.23)***

*** 6.41(1.46)***

Model 2 B(s.e) *** 3.78(2.18) −2.87(2.25) −6.85(2.11)**

0.009 0.054

0.042 0.042

η 0.042 0.007 0.004 0.024

2

0.039

1.48(0.35)***

449 0.420

0.012 0.014

0.043 0.043

η 0.037 0.005 0.004 0.022

2

−0.93(0.41)* 0.64(0.26)*

*** 6.36(1.44)***

Model 3 B(s.e) ** 3.24(2.15) −2.82(2.21) −6.48(2.07)**

0.005 0.105 0.112

1.23(0.17)*** 0.65(0.09)*** 449 0.598

0.000 0.006

0.050 0.050

η2 0.056 0.021 0.006 0.018

0.45(0.30)

0.03(0.35) 0.35(0.22)

*** 5.73(1.20)***

Model 4 B(s.e) *** 5.47(1.81)** −3.12(1.86) −4.86(1.75)**

Table 4.6: GCSE results using SOC class schema (. . . continued from overleaf)

449 0.362

η 0.055 0.009 0.001 0.033

2

Model 1 B(s.e) *** 4.49(2.24)* −1.37(2.29) −8.24(2.14)***

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

102

demanding a degree used by SOC is out-dated. Today, many positions in management require a degree. In addition, many graduates must now take up clerical positions that would previously have been taken up by schoolleavers. Parental cultural participation is included in model 2. This somewhat reduces the effect of social class on GCSE performance, but social class in general, and the professional class in particular, remains highly significant. Model 3 shows the partial mediation of parents’ cultural participation by pupils’ cultural participation (the inclusion of pupils’ cultural participation has little impact on the social class effect). Pupils’ knowledge and language scores are included in model 4. This has the effect of reducing the advantage associated with the professional classes, but actually increasing the advantage of the employers and managers, so that, in this model, the proportion of the variance in GCSE results explained by each of these class categories is very similar. So, passing on cultural knowledge and linguistic ability seems to be an important part of the explanation of the academic performance of the children of professionals, but the advantage enjoyed by the children of employers and managers appears to be at its strongest once we control for these factors. This could suggest that the educational advantage enjoyed by the employer/manager group is largely due to material factors, whereas cultural factors play an important role for the professional classes.

Arts and Sciences Bourdieu never suggests that cultural capital is a less significant advantage in science subjects than it is in arts subjects. However, it might be suspected that cultural capital, especially the linguistic element of this, would in fact be a less valuable resource in science subjects. To examine this possibility, I will compare pupils’ grades in maths and English. The reason for choosing these subjects is that, unlike most GCSE subjects, they are not optional in any of the schools in my survey. The models for English GCSE (table 4.7) and maths GCSE (table 4.8) follow the same pattern as the models for total GCSE attainment. As in the case of pupils’ total GCSE score, the effects of both parents’ cultural activities and pupils’ cultural activities are mediated by pupils’ cultural knowledge and linguistic ability.

Intercept Parents’ qualifications Missing Degree Intermediate O-level CSE or no qualifications Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Unskilled manual Sex Male Female

η 0.172 0.029

0.000 0.010 0.015 0.008

0.076 0.002 0.038

0.017

0.009

0.008 0.000

0.043 0.043

−0.02(0.26) 0.65(0.31)* 0.80(0.31)* 0.52(0.27)

*** −0.32(0.34) 1.46(0.35)***

0.92(0.33)**

0.59(0.30)

0.68(0.35) 0.02(0.35)

*** −0.95(0.22)***

2

Model 1 B(s.e) 3.35(0.35)*** *

0.042 0.042

0.006 0.000

0.007

0.013

0.055 0.002 0.025

0.000 0.002 0.011 0.006

η 0.150 0.023

2

*** −0.86(0.21)***

0.62(0.34) −0.01(0.33)

0.61(0.29)*

0.80(0.32)*

*** −0.35(0.33) 1.14(0.34)**

−0.07(0.25) 0.23(0.32) 0.66(0.30)* 0.42(0.26)

Model 3 B(s.e) 2.51(0.37)*** *

0.038 0.038

0.007 0.000

0.010

0.014

0.060 0.003 0.025

0.000 0.001 0.011 0.006

η 0.097 0.022

2

*** −0.93(0.21)***

0.64(0.35) −0.03(0.34)

0.51(0.29)

0.66(0.32)*

** −0.39(0.33) 0.96(0.35)**

−0.12(0.25) 0.26(0.32) 0.64(0.30)* 0.40(0.27)

Model 4 B(s.e) 2.64(0.37)*** *

Table 4.7: English GCSE results (continued overleaf. . . )

*** −0.93(0.21)***

0.57(0.35) 0.02(0.34)

0.54(0.30)

0.78(0.33)*

*** −0.33(0.33) 1.19(0.35)**

−0.09(0.26) 0.32(0.32) 0.68(0.31)* 0.45(0.27)

Model 2 B(s.e) 3.09(0.35)*** *

0.043 0.043

0.008 0.000

0.007

0.010

0.049 0.003 0.018

0.001 0.002 0.011 0.005

η 0.107 0.023

2

*** −1.02(0.18)***

0.47(0.29) 0.09(0.29)

0.49(0.25)

0.64(0.28)*

* −0.03(0.28) 0.80(0.30)**

−0.24(0.22) −0.26(0.28) 0.29(0.26) 0.15(0.23)

Model 5 B(s.e) 0.94(0.34)**

0.068 0.068

0.006 0.000

0.009

0.012

0.031 0.000 0.016

0.003 0.002 0.003 0.001

η2 0.017 0.020

School School 1 School 2 School 3 School 4 Family structure Family intact Step or single parent family Sibling number Parents’ cultural participation Pupils’ cultural participation Reading Formal cultural particpation Music TV Cultural knowledge Language N η 2 for the model

0.008

−0.09(0.05)

453 0.301

−0.06(0.05) 0.10(0.03)***

** 0.44(0.16)**

Model 2 B(s.e) ** 0.90(0.24)*** −0.07(0.25) −0.09(0.24)

0.004 0.030

0.017 0.017

η 0.036 0.031 0.000 0.000

2

0.047

0.18(0.04)***

453 0.334

0.006 0.003

0.017 0.017

η 0.031 0.028 0.000 0.000

2

−0.08(0.05) 0.03(0.03)

** 0.44(0.16)**

Model 3 B(s.e) ** 0.84(0.24)*** −0.07(0.25) −0.05(0.23)

0.000 0.011

0.01(0.12) 0.25(0.12)*

438 0.329

0.039 0.000

0.006 0.007

0.013 0.013

η 0.026 0.022 0.000 0.000

2

0.22(0.05)*** −0.03(0.11)

−0.08(0.05) 0.05(0.03)

* 0.38(0.16)*

Model 4 B(s.e) * 0.75(0.24)** −0.10(0.25) −0.06(0.23)

Table 4.7: English GCSE results (. . . continued from overleaf)

0.018 0.018

** 0.47(0.17)**

453 0.280

η 0.040 0.032 0.000 0.002

2

Model 1 B(s.e) *** 0.94(0.25)*** 0.05(0.25) −0.20(0.24)

0.087 0.088

0.07(0.01)*** 453 0.508

0.010

0.001 0.000

0.018 0.018

η2 0.062 0.056 0.001 0.001

0.13(0.02)***

0.07(0.04)*

0.03(0.04) 0.00(0.03)

** 0.39(0.14)**

Model 5 B(s.e) *** 1.04(0.21)*** −0.16(0.22) 0.11(0.20)

Intercept Parents’ qualifications Missing Degree Intermediate O-level CSE or no qualifications Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Unskilled manual Sex Male Female 1.31(0.37)*** 0.52(0.36)

0.00(0.22)

0.091 0.000 0.063

0.019

0.022

0.032 0.005

0.000 0.000

*** 0.12(0.35) 1.97(0.36)***

1.01(0.34)**

0.98(0.32)**

1.40(0.37)*** 0.52(0.36)

−0.02(0.22)

0.000 0.000

0.029 0.005

0.020

0.015

0.072 0.000 0.048

0.008 0.001 0.016 0.000

η 0.061 0.050

2

0.05(0.22)

1.34(0.36)*** 0.50(0.35)

0.98(0.31)**

0.91(0.34)**

*** 0.09(0.35) 1.71(0.37)***

−0.51(0.27) 0.20(0.34) 0.84(0.32)** −0.02(0.28)

Model 3 B(s.e) 1.56(0.39)*** ***

0.000 0.000

0.031 0.005

0.022

0.016

0.075 0.000 0.048

0.008 0.001 0.015 0.000

η 0.035 0.049

2

−0.01(0.23)

1.25(0.38)** 0.52(0.36)

0.97(0.32)**

0.85(0.35)*

*** 0.11(0.36) 1.62(0.38)***

−0.51(0.27) 0.22(0.34) 0.85(0.33)** −0.04(0.29)

Model 4 B(s.e) 1.63(0.40)*** ***

Table 4.8: Maths GCSE results (continued overleaf. . . )

0.94(0.31)**

0.89(0.34)**

*** 0.11(0.35) 1.74(0.37)***

−0.52(0.27) 0.26(0.34) 0.85(0.32)** −0.00(0.28)

0.007 0.006 0.019 0.000

−0.47(0.27) 0.54(0.33) 0.95(0.32)** 0.06(0.28)

Model 2 B(s.e) 1.96(0.37)*** ***

η 0.075 0.057

2

Model 1 B(s.e) 2.17(0.37)*** ***

0.000 0.000

0.025 0.005

0.022

0.014

0.065 0.000 0.043

0.008 0.001 0.016 0.000

η 0.038 0.050

2

−0.10(0.20)

1.20(0.32)*** 0.60(0.31)

0.87(0.27)**

0.75(0.30)*

** 0.41(0.31) 1.36(0.32)***

−0.67(0.24)** −0.27(0.30) 0.46(0.28) −0.29(0.25)

Model 5 B(s.e) −0.01(0.37) ***

0.001 0.001

0.032 0.009

0.023

0.014

0.053 0.004 0.039

0.018 0.002 0.006 0.003

η2 0.000 0.050

School School 1 School 2 School 3 School 4 Family structure Family intact Step or single parent family Sibling number Parents’ cultural participation Pupils’ cultural participation Reading Formal cultural particpation Music TV Cultural knowledge Language N η 2 for the model 453 0.303

−0.13(0.05)* 0.08(0.03)**

** 0.46(0.17)**

0.57(0.25)* 0.05(0.26) 0.31(0.25)

Model 2 B(s.e)

0.015 0.020

0.017 0.017

η 0.013 0.011 0.000 0.004

2

0.020

0.12(0.04)**

453 0.317

0.017 0.004

0.016 0.016

η 0.012 0.010 0.000 0.004

2

−0.13(0.05)** 0.04(0.03)

** 0.46(0.17)**

0.53(0.25)* 0.06(0.26) 0.34(0.25)

Model 3 B(s.e)

Model 4 B(s.e)

0.000 0.007

0.00(0.13) 0.21(0.13)

438 0.310

0.015 0.000

0.017 0.006

0.013 0.013

η 0.010 0.007 0.000 0.004

2

0.14(0.06)* −0.04(0.12)

−0.13(0.05)** 0.05(0.03)

* 0.41(0.17)*

0.46(0.26) −0.02(0.27) 0.31(0.25)

Table 4.8: Maths GCSE results (. . . continued from overleaf)

0.019

−0.14(0.05)**

453 0.289

0.018 0.018

η 0.013 0.012 0.001 0.002

2

** 0.48(0.17)**

0.60(0.26)* 0.16(0.26) 0.21(0.25)

Model 1 B(s.e)

0.066 0.086

0.07(0.01)*** 453 0.475

0.000

0.001 0.000

0.017 0.017

η2 0.031 0.025 0.000 0.011

0.12(0.02)***

0.01(0.04)

−0.03(0.04) 0.01(0.03)

** 0.41(0.15)**

Model 5 B(s.e) ** 0.75(0.22)** −0.02(0.23) 0.48(0.22)*

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

107

One cannot use η 2 to directly compare effects in regressions with different independent variables, since the size of η 2 is affected by the total variance in the independent variable. (However, note that the variances for maths and English GCSE scores are not very different, being 4.18 and 3.72 respectively.) Therefore, the parameter estimates should be used to compare the role of cultural resources in the two models. The effects of parents’ cultural capital in Model 2 and of pupils’ cultural activities in Model 3 are larger in the case of English GCSE than in the case of maths GCSE. Since reading is the key component of cultural activities, it is unsurprising that this variable has a greater effect on performance in English than in maths. However, the effects of both reading and TV watched are significant in the case of English GCSE, whereas only reading is significant in the case of maths GCSE. The vocabulary and cultural knowledge scores have similar effects on performance in English and maths. So, the effect of cultural capital on examination attainment does not seem to be limited to arts subjects, although its impact may be strongest in arts subjects.

4.3

Conclusions

The concept of cultural capital has often been assimilated to the data available to researchers. By using data specifically designed to measure pupils’ and parents’ cultural capital, I have been able to provide a better test of Bourdieu’s theory. The first element of Bourdieu’s theory that I set out to test is the claim that cultural capital is transmitted by higher-class parents to their children. I broke this down into two questions, firstly, what is the social distribution of cultural capital, and secondly, to what extent is cultural capital transmitted from parents to their children. I found that parental cultural capital is strongly associated with parental social class and with parental qualifications. These associations back Bourdieu’s view that cultural capital is unequally distributed according to social class and education. The view that cultural capital is transmitted from parents to their children is strongly supported in the case of pupils’ cultural activities. This component of pupils’ cultural capital varies by social class, but this variation is entirely mediated by parental cultural participation. Further evidence to back the view that cultural capital is transmitted in the home is the lack of a school effect in determining this component of pupils’ cultural capital. The link

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

108

between parental cultural capital and pupils’ knowledge and language scores is weaker, but this is unsurprising given that my measure of parental cultural capital is a measure of activities. There is no school effect on the test of linguistic ability, and there is only a small school effect on cultural knowledge. This contrasts with a strong school effect on GCSE attainment, and suggests that linguistic ability and cultural knowledge are more strongly transmitted within the home than in the school. However, it must be borne in mind that my sample only contains four schools. Ideally one would collect a larger sample including different types of schools, as it is possible that school type might affect pupils’ cultural capital. For instance, it is possible that private schools may instil cultural capital in pupils. Pupils’ reading and the type of TV watched by pupils account for a significant proportion of the variance in linguistic ability and cultural knowledge, whereas participation in formal culture does not. This backs the view that reading develops the intellectual abilities of pupils, whereas participation in formal culture does not. This could be interpreted as supporting the views of Crook (Crook, 1997) and de Graaf et. al. (De Graaf et al., 2000) that public cultural participation serves to communicate status, whereas private cultural consumption is a means of intellectual self-development. Television watching is not an indicator of cultural capital that has been used by previous authors, but TV, in common with books, transmits information and may introduce an individual to new vocabulary and styles of expression. Note, however, that listening to classical music and playing an instrument are not associated with linguistic ability or cultural knowledge. Perhaps, then, the important distinction is not that of “public” or “formal” vs. “private” or “informal” cultural participation, but rather that of verbal or literary forms which use words to transmit information or content, vs. visual or musical forms which are not based on words and are therefore less likely to develop the skills that are rewarded within the school. Gender does not account for an important proportion of the variance in any component of pupils’ cultural capital. Although there are slight variations in cultural capital according to gender, these differences do not account for girls’ superior performance at GCSE level. I went on to examine whether cultural capital affects pupils’ attainment at GCSE level. The activities component of pupils’ cultural capital is a significant determinant of pupils’ GCSE score, as is parents’ cultural capital. Again, reading and watching TV are the only significant elements of pupils’ cultural participation. Of these, reading has by far the greater effect. These

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

109

effects are entirely mediated by pupils’ vocabulary and cultural knowledge. This firmly backs the view that the reason for the effect of cultural participation on academic attainment is that cultural participation is associated with intellectual resources which help pupils’ at school. This research gives no support to the view that teacher’s are prejudiced against working-class pupils because of their lack of cultural capital. (Note that Hurrel (Hurrel, 1995) has provided strong empirical evidence against the view that teachers are prejudiced against working-class pupils.) Furthermore, in the British context of an enormous decline in the status of the teaching profession, it increasingly seems odd to portray teachers as an ´elite (cultural or otherwise) who are prejudiced against non-´elite pupils. It may be argued that the association between cultural knowledge and GCSE attainment must be due to a bias towards high culture in the curriculum. However, it may be that pupils are rewarded highly in examinations and assessed coursework for demonstrating precisely that knowledge which they are unlikely to have gained within the school. This would be consistent with Bourdieu’s claim that the school fails to give explicitly to everyone that which it implicitly demands of everyone. In this case, pupils from backgrounds poor in cultural capital may suffer most from a curriculum that is designed to avoid content and styles that are associated with the dominant culture. Furthermore, it seems unlikely that it would be possible, and certain that it would be undesirable, to introduce a form of assessment that would not reward linguistic ability or cultural knowledge, broadly defined. Parents’ social class retains a large and significant direct effect on GCSE attainment, controlling for the cultural capital variables. Therefore, it seems that cultural capital is one mechanism through which higher-class families ensure educational advantage for their children, but it leaves most of the social class differential in attainment unexplained. Other mechanisms, such as class differentials in material resources and educational aspirations must account for the remaining differential in educational attainment. However, using the SOC (Standard Occupational Classification) class schema, I have shown that cultural resources do seem to be an important part of the explanation for the educational advantage enjoyed by the children of professional parents, whereas this is not the case for the children of employers and managers. In conclusion, I have tried to give a fair test of Bourdieu’s theory of cultural reproduction, and have found that, although it provides some useful insights, and helps to explain class differentials in educational attainment, it does not

CHAPTER 4. EDUCATIONAL ATTAINMENT AT GCSE

110

provide a complete account of these differentials. Cultural capital is associated with social class, and is transmitted from parents to children. The possession of cultural capital does have a significant effect on GCSE attainment. However, this gives us only a partial explanation of class differentials in GCSE attainment. I stated previously that one cannot say which cultural activities should be seen as “capital” without an analysis into which cultural activities are associated with educational success. Reading and TV viewing habits are associated with GCSE attainment and with cultural knowledge and linguistic ability (which in turn are associated with GCSE success). This is evidence that it is reasonable to see these activities as cultural capital. There is no evidence here, on the other hand, that musical habits (listening and playing) or participation in formal culture constitute capital. In sum, this work vindicates the usefulness of “cultural capital” as an explanatory concept, but does not support the grand theory of “cultural reproduction”.

Chapter 5 Beliefs and Desires 5.1

Introduction

It is well documented that, even controlling for initial educational attainment or ability, working-class pupils have often been less likely to remain in post-compulsory education and to pursue prestigious educational options than middle-class pupils. (See for instance, Shavit and Blossfeld (Shavit and Blossfeld, 1993), and Micklewright (Micklewright, 1989)). Rational choice theorists have focused on these “secondary effects of stratification” rather than on initial inequalities in educational attainment. According to rational choice theory, actors maximise the fulfilment of their desires on the basis of their beliefs about the situation (or at least, in general they behave as if this was what they were doing). This means that rational choice theorists assume that if any agent x wants d and believes that action a is the best means to attain d, then x does a (all other things being equal). Or, as Elster puts it: “The action should be the best way of satisfying the agent’s desires given his beliefs.” (Elster, 1990, p. 19) So, actions follow rationally from the beliefs and desires of the agent. (Or, in the language of economics, the expectations and preferences of the agent.) But, any action can be interpreted as rational in the light of some set of beliefs and desires. For example, if I clap my hands together 100 times every 111

CHAPTER 5. BELIEFS AND DESIRES

112

day at sunset, this may be explained by my belief that in doing so I am summoning evil spirits, and by my desire to destroy the universe. So, the assumption of rationality alone does not rule out any possible action. It is clear that we cannot generate any hypotheses about action simply on the basis that it is rational without making any claims about the beliefs and desires held by the actor. One response to this problem on the part of rational choice theorists has been the attempt to give a more substantive definition of rationality, one that specifies what counts as a rational belief or a rational desire. This is inherently problematic, as in general we do not choose our beliefs and desires, or at least we do not choose them in the same sense that we choose our actions, so the rationality of beliefs and desires cannot be defined instrumentally, as the rationality of actions is. Attempts to define “rational” desires have been unconvincing, failing to overcome the basic point that one cannot argue with taste. It may seem more reasonable to attempt to categorise beliefs as rational or irrational. For instance, Elster (Elster, 1987) states that for a belief to be rational, it must be based on the available evidence, caused by the evidence, and caused by the evidence in the right way. This condition is violated if an agent’s beliefs are affected by their interests or desires, or if cognitive errors (i.e. illogical thinking) play a part in the formation of the belief. This seems perfectly sensible as a description of how we ought to form beliefs. Unfortunately, strong evidence has been provided by social psychologists that this is not the way that most of us actually form beliefs, and that cognitive errors are in fact widespread and systematic (Kahneman et al., 1982). For Elster, this is perhaps unimportant, as he sees rational choice primarily as a normative theory. However, if rational choice is to work as an explanatory theory, then it should not be founded on a false view of belief formation. Boudon’s (Boudon, 1994) response to evidence of the importance of cognitive biases in belief formation has been to adapt his concept of rationality to allow for these errors. This leaves us nearly back where we started, with a version of rational choice theory that rules very little out. Since what we mean by “rationality” when talking about beliefs, desires, and actions are three entirely different things, why should a sociologist who makes rational choice assumptions about the relation between an agents beliefs and desires and their actions assume that the agents’ beliefs and desires are also “rational”? The point that an agent’s desires may not be best explained by any kind of rationality assumption, even if the step from desires to actions is, is made by Brennan (Brennan, 1990). I see no reason why this principle cannot also be extended to apply to the beliefs of the agent. So, rather than

CHAPTER 5. BELIEFS AND DESIRES

113

theorising about what might count as a rational motivation for an action, it may be more useful to analyse agents’ motivations empirically. Rational choice theorists analysing class differentials in educational participation have made different assumptions about students’ beliefs and desires. To date, these claims have not been assessed empirically. Yet it is crucial to examine the motivations of student decision-makers directly, as their actions alone cannot tell us what their motivations are.

5.2

Rational Choice Theories

As I stated in chapter 2, several rational choice theorists have addressed the issue of social class differences in educational participation, and the claims they make regarding students’ beliefs and desires vary considerably. Murphy (Murphy, 1990) takes issue with the view that class differentials in educational attainment have anything to do with inequality of opportunity. He states that the equation of class differentials in educational attainment with class inequality is due to a failure to take class differences in educational aspiration seriously. In Murphy’s view, working class youth simply demand lower levels of education than do middle class youth. This difference in the levels of demand for education is not due to inequality of opportunity, but is simply a matter of taste. Murphy treats preferences as given, and not in need of explanation. His analysis is problematic in that the supposed social class difference in tastes is inferred from the differential in educational participation. Yet no direct evidence is given to show that the difference in educational participation is actually due simply to a relative distaste for education on the part of the working classes, as opposed to, for instance, class differences in resources or in students’ perceptions of their own ability. Gambetta (Gambetta, 1987) also explains the class differential in educational participation in terms of preferences, but introduces an additional mechanism in the form of social norms. “. . . it could either be that relatively more subjects in the middle class feel a greater normative pressure to resist the temptation to abandon school after a failure or, on the other hand, that relatively more subjects in the working class do not attach as high a value to education.” (Gambetta, 1987, p. 173)

CHAPTER 5. BELIEFS AND DESIRES

114

Surprisingly, there does not seem to be a huge gulf between these rational choice accounts of class differentials in educational participation, and culturalist accounts such as Willis’ (Willis, 1977) explanation of the behaviour of his ‘lads’ in terms of a working-class anti-school culture. (Note though, that only some of the working-class pupils Willis spoke to were actually part of this ‘lads” culture. Many of the others were keen to do well at school.) The common feature of these accounts is the view that working class people have relatively negative attitudes to education or schooling in comparison with middle class people. However, other rational choice theorists reject the view that values, norms and beliefs about education vary according to social class. Boudon (Boudon, 1974) states that class differentials in educational attainment are explained by the difference in the costs and benefits that are associated with different educational options for students from different social classes. For Boudon, the benefits associated with each educational option vary with social class because ambition is relative to the social starting point of an individual. So, a working class child who wants to be a lawyer must be more ambitious than a middle class child who wants to be a lawyer. Therefore, prestigious educational options may be essential in avoiding social demotion for middle class pupils, whereas working class pupils can avoid social demotion without pursuing such options. This leads to middle class pupils being more likely to pursue prestigious educational options than working class pupils at any given level of ability. This view is developed by Breen and Goldthorpe (Breen and Goldthorpe, 1997). They make explicit Boudon’s implicit assumptions, 1) that people’s priority is to avoid social demotion, rather than to pursue social mobility, and 2) that failure in a high prestige option is believed to be more likely to lead to social demotion than not attempting to pursue such an option. Breen and Goldthorpe explicitly state that values, norms and beliefs regarding education do not vary by social class, and that classes differ in terms of two factors only, average ability and resources. In sum, rational choice theorists have made various assumptions about the attitudes and beliefs of individuals regarding education. The question of whether preferences vary by social class is the subject of much disagreement. But no attempt has been made to ground these claims about students’ beliefs and desires in empirical evidence.

CHAPTER 5. BELIEFS AND DESIRES

5.3

115

Research Questions

I have stated that some rational choice theorists strongly reject the idea that attitudes to education vary by social class, while others accept it. It certainly seems possible that attitudes towards education may vary by social class. Education may be valued as a good in itself and/or as a means to success in the labour market. There could be a social class difference in the evaluation of the worth of education on the labour market, for instance if working class pupils saw educational credentials as less essential to occupational success than did middle class pupils. Another possibility is that middle class students may place a higher intrinsic value on education than working class pupils, perhaps being more likely to see education as enjoyable, or as a tool for self-development. A further possibility, perhaps applying especially strongly to the intrinsic value of education, is that the subjective evaluation of education may be especially positive among those families with high levels of “cultural capital”. In households rich in cultural capital, familiarity with the dominant culture is important for leisure activities, and this may lead to an increase in the value placed on education as an intrinsic good. There may be a gender difference in attitudes to education. Girls are often seen as having a more positive or “mature” attitude towards education than do boys, and therefore being better behaved and harder working in school. This could be a possible explanation for girls’ superior performance at GCSE. If attitudes to education do vary by social class and gender, then this may help to explain differentials in educational participation. If, on the other hand, attitudes to education do not vary by social class, then the idea that the class differential in educational participation simply reveals a difference in preferences for education between the social classes must be abandoned. Beliefs about ability may not be entirely determined by actual ability, and it is possible that an individual’s social background and gender may affect these beliefs. Breen and Goldthorpe’s formulation of rational choice theory does not assume that beliefs are necessarily formed entirely rationally, on the basis of the relevant evidence. What it does assume is that there is no systematic distortion of beliefs by social category. So, if it is the case that people in general overestimate their abilities, this would not contradict Breen and Goldthorpe’s view. However, if, for example, men overestimate their abilities in comparison with women, this would violate Breen and Goldthorpe’s

CHAPTER 5. BELIEFS AND DESIRES

116

assumption. This assumption is in need of empirical test. It seems quite possible that pupils in different social categories may form systematically distorted views of their abilities, due perhaps to different evaluations of their abilities from parents and teachers. If it is the case that working class pupils underestimate their abilities in comparison to middle class pupils, this might help to provide an explanation for the comparatively low rate of participation of working class pupils in post-compulsory education, and particularly in more prestigious educational courses. So, in this chapter, I will assess the posssibility that beliefs and attitudes regarding education may be socially structured in the ways outlined above.

5.4 5.4.1

Analysis Attitudes to Education

The pupils in my sample generally responded positively to the questions on their attitudes towards education. For instance, 81 per cent either agreed or strongly agreed that ‘The more qualifications you get, the better the job you are likely to get’, and 69 per cent either agreed or strongly agreed that ‘Studying is worthwhile for its own sake’. See appendix C for the educational attitudes scale and a reliability analysis of this scale. Table 5.1 shows pupils’ mean scores in the educational attitude scale, according to gender, social class and parental educational level. The maximum possible score (for a pupil who responded ‘agree strongly’ to all the positively phrased questions, and ‘disagree strongly’ to all the negatively phrased ones) would have been 80. Female pupils, pupils from the service class and pupils who’s parents are graduates all have relatively high mean scores on the scale of educational attitudes, but only slightly so. An analysis of variance will show whether the differences in educational attitudes between these social categories are significant. It is often claimed that there is a difference between the social classes in their attitudes to education. As I stated previously, Murphy puts social class differentials in educational attainment and participation entirely down to the relative distaste for education of the working classes, and several theorists

Female

Service class

Non service class

Degree

No Degree

Mean Std. Deviation

117

Male

CHAPTER 5. BELIEFS AND DESIRES

56.7 7.62

58.3 6.64

58.8 6.90

56.9 6.93

58.5 6.43

57.6 7.29

Table 5.1: Pupils’ attitudes to education — mean scores state that there are social class differences in norms and attitudes regarding education. Boudon and Breen and Goldthorpe on the other hand, explicitly exclude the possibility of social class differences in attitudes or tastes for education. And it is this latter view that seems to be borne out by table 5.2, which shows no significant association between social class and attitudes to education for the pupils’ in my sample, even before GCSE performance is controlled for. This is true both of attitudes to the intrinsic value of education and attitudes to the value of education in the job market. (There is, however, a significant association between social class and those items that fell into neither of these categories.) This lack of association between social class and attitudes to education is all the more striking given that pupils from working class backgrounds tend to receive markedly worse GCSE grades than pupils’ from middle class backgrounds, and GCSE performance is significantly associated with attitudes to education, as table 5.2 shows. However, it may be argued that the lack of a significant association is due to the use of a 6-category measure of social class. Perhaps if a bivariate measure is used, simply comparing the service class to the non service-class, a significant association may be revealed. Table 5.3 shows the results of this analysis. It shows a small but significant association between service class status and educational attitudes. However, breaking down the scale into its constituent parts shows that there is no significant association between service class status and attitudes to education either as an intrinsic good or as an instrumental good, but only with those items that fit into neither of these categories. This is difficult to interpret, but one possibility is that the items that fit into neither the ‘instrumental’ nor the ‘intrinsic’ category reflect behaviour rather than attitudes, (the item ‘I muck about in lessons’ would be an example of this). Table 5.4 shows the results controlling for GCSE attainment, in order to

Pupils’ Cultural Capital

GCSE score

Neither

Parents’ Cultural Capital

Intrinsic

Qualifications

Instrumental

Class

Educational Attitudes

118

Sex

CHAPTER 5. BELIEFS AND DESIRES

η2

0.012

0.025

0.026

0.048

0.057

0.052

p η2 p η2 p η2 p

0.018 0.000 0.865 0.014 0.010 0.012 0.014

0.068 0.014 0.353 0.020 0.148 0.029 0.038

0.104 0.024 0.139 0.021 0.200 0.021 0.194

0.000 0.015 0.008 0.032 0.000 0.049 0.000

0.000 0.007 0.071 0.048 0.000 0.063 0.000

0.000 0.012 0.018 0.013 0.016 0.132 0.000

Instrumental Intrinsic Neither

Qualifications

Educational Attitudes

Class

Table 5.2: Pupils’ attitudes to education — analysis of variance

η2

0.017

0.012

p η2 p η2 p η2 p

0.018 0.011 0.081 0.013 0.054 0.017 0.018

0.061 0.011 0.077 0.007 0.194 0.013 0.050

Table 5.3: Pupils’ attitudes to education using bivariate parental class and education

CHAPTER 5. BELIEFS AND DESIRES

119

Pupils’ Cultural Capital

Neither

Parents’ Cultural Capital

Intrinsic

Qualifications

Instrumental

Class

Educational Attitudes

Sex

show the variation in the attitudes of pupils at the same level of educational attainment. The reason for controlling for GCSE attainmment is that only those differences in pupils’ attitudes which remain apparent once GCSE performance has been controlled for will be useful in explaining any differences in the tendency to remain in further education of pupils from different social groups with similar levels of attainment at GCSE. In order to be charitable to the social class variable, I have used the bivariate version of this variable. Controlling for GCSE results, there is a clear lack of association between social class and attitudes to education. This is important, as, if the lower staying on rates of working class pupils compared to middle class pupils at the same level of initial examination performance were to be explained by social class differences in attitudes to education, then these differences would have to remain after controlling for GCSE performance.

η2

0.009

0.003

0.001

0.019

0.028

p η2 p η2 p η2 p

0.045 0.000 0.966 0.012 0.018 0.008 0.052

0.533 0.005 0.311 0.007 0.181 0.001 0.836

0.795 0.006 0.264 0.003 0.481 0.001 0.726

0.003 0.008 0.050 0.022 0.001 0.005 0.122

0.000 0.002 0.285 0.038 0.000 0.015 0.008

Table 5.4: Pupils’ attitudes to education controlling for GCSE results

One might guess that, even if there is no relationship between social class and attitudes to education, there might be a relationship between parents’ qualifications and pupils’ attitudes to education. After all, it is quite plausible that parents who have benefited from education themselves will pass on a positive view of education to their children. In fact, table 5.2 shows no significant association between parents’ qualifications and pupils’ attitudes to education. Table 5.3 shows that this lack of association holds when a bivariate categorisation of parental qualifications (degree or no degree) is used. This lack of

CHAPTER 5. BELIEFS AND DESIRES

120

association holds both for attitudes to education as an intrinsic good and for attitudes to education as providing an advantage in the labour market. (The association with those items that fell into neither category is just significant at the 0.05 level.) Again, it is striking that, despite the strong association between parents’ educational level and pupils’ educational performance, the lack of association between parents’ qualifications and pupils’ attitudes to education is apparent even before controlling for pupils’ GCSE attainment. Table 5.4 shows that the insignificance of association becomes even clearer once GCSE results are controlled for. (Again, a bivariate categorisation of parental qualifications is used.) Given the absence of a clear association between attitudes to education and social class, it is striking that there is a highly significant association between parents’ cultural participation and pupils’ attitudes to education, which remains after controlling for performance at GCSE. This is most significant in the case of pupils’ view of the intrinsic worth of education, (η 2 = 0.022, p ≤ 0.001). The association between parents’ cultural participation and pupils’ attitudes to education as an instrumental good is just barely significant at the 0.05 level. A slightly stronger association holds between pupils’ own cultural participation and attitudes to education than between parents’ cultural participation and pupils’ attitudes to education. Again, this association is at its strongest in the case of pupils’ view of the intrinsic worth of education, (η 2 = 0.38, p ≤ 0.000). There is no significant association between pupils’ cultural participation and the extent to which pupils value education as a labour market good. The high value placed on the intrinsic worth of education by families rich in cultural capital may be due to the fact that the leisure activities valued by these families, such as reading, may demand and/or develop intellectual abilities which also provide an educational advantage. Since these abilities are important to social life, (where social life includes cultural participation), as well as to work, education is more likely to be seen as crucial to self-development in families that have high levels of participation in the dominant culture. Girls have significantly more positive attitudes to education than do boys, but this association is limited to attitudes to the intrinsic value of education, and is not very large. The association between gender and the evaluation of education as an intrinsic good, controlling for GCSE attainment, is η 2 = 0.012, p ≤ 0.018. There is no significant difference in boys’ and girls’ evaluations of the value of education in the job market, suggesting that this aspect of educational attainment is equally important to both boys and girls.

CHAPTER 5. BELIEFS AND DESIRES

121

So, attitudes to education vary only very slightly by social class, and this association is mediated by pupils’ GCSE performance. However, cultural participation and gender are significantly associated with attitudes to education as an intrinsic good.

5.4.2

Beliefs About Ability

I have stated that it is possible that individuals’ beliefs about their own abilities may not be entirely determined by their actual abilities, and that the truth or falsity of this claim may be determined by empirical test. I asked pupils in their final GCSE year what grades they thought they were likely to get in their GCSEs. I also asked them to rate their academic abilities in comparison to other pupils at their school. I then compared these responses to the actual grades achieved by the pupils. Firstly, how accurate are the pupils’ beliefs about their own abilities in general? It seems that, overall, pupils’ overestimate themselves. Table 5.5 shows pupils’ responses to the question ‘In general, how do you rate your academic abilities as compared to other pupils at your school?’. Strikingly, only 4 per cent of pupils rated themselves as either below average or poor. Pupils did not show a strong tendency to rate themselves as ‘excellent’, (only 5.4% did so). The great majority of pupils, (90.9%) rated themselves as either average or above average. Table 5.6 shows pupils responses to the question ‘What grades do you think you are most likely to get in your GCSEs?’. No pupil placed themselves in the lowest category, ‘mostly Gs and ungraded’, although in fact 16 pupils in my sample failed to get any GCSEs at all.

Excellent Above average Average Below average Poor

f 25 164 256 15 2

% 5.4 35.5 55.4 3.2 0.4

Table 5.5: Pupils’ self-assessed ability — frequencies

I gave pupils a score for their actual GCSE results, giving one point for a G grade, 2 for an F etc., up to 8 points for an A*. I then divided this score by the number of GCSEs each pupil said they were studying for at the time

CHAPTER 5. BELIEFS AND DESIRES

Mostly Mostly Mostly Mostly Mostly Mostly Mostly Mostly

As As and Bs Bs and Cs Cs and Ds Ds and Es Es and Fs Fs and Gs Gs and ungraded

122 f 13 69 160 162 45 12 1 0

% 2.8 14.9 34.6 35.1 9.7 2.6 0.2 0.0

Table 5.6: Pupils’ self-predicted grades — frequencies

of the survey, to give a mean score. Table 5.7 shows a comparison of pupils’ self-predicted grades to the actual mean scores pupils’ achieved. This table shows a general tendency for pupils to be overly optimistic in predicting their results. For instance, I gave 4 points for a D grade and 5 for a C. Therefore, it seems fair to say that a pupil who expects to get ‘mostly Cs and Ds’ is expecting to achieve an average score of around 4.5. Table 5.7 shows that, in fact, pupils’ who expected to achieve mostly Cs and Ds gained on average a mean score of 3.4. This means that these pupils’ mean grade scores were on average approximately one grade lower than they predicted.

Mostly Mostly Mostly Mostly Mostly Mostly Mostly Mostly

As As and Bs Bs and Cs Cs and Ds Ds and Es Es and Fs Fs and Gs Gs and ungraded

Average score should be 7.0 6.5 5.5 4.5 3.5 2.5 1.5 0.5

Actual mean score 6.3 5.9 4.7 3.4 2.8 1.8 2.3 −

Table 5.7: Pupils’ self-predicted grades and actual grades

However, this data only contradicts the strong view that rational actors must form accurate beliefs. It has no bearing on Breen and Goldthorpe’s view that

CHAPTER 5. BELIEFS AND DESIRES

123

beliefs are not systematically distorted by social category. If the view that there is no systematic social belief distortion is to be supported, there should be no relationship between social class, gender, etc. and pupils’ estimation of their own abilities once actual academic performance is taken into account. The claim that there is no systematic distortion of beliefs about ability is perhaps at its least plausible when applied to gender, as it has often been claimed that girls underestimate themselves in comparison to boys. And indeed, this view is borne out by this study. Table 5.8 shows pupils’ estimation of their own academic abilities broken down by gender. We can see that girls have a lower estimation of their abilities than do boys. In particular, boys are considerably more likely to rate themselves as being above average than girls are (42.1% compared to 28.2%). Girls are much more likely to rate themselves as average than boys are, (63.2% compared to 48.3%). Table 5.9 shows pupils’ self-predicted grades broken down by gender. Girls are more likely than boys to have placed themselves in the bottom half of the table (mostly Ds and Es or less). 15.7% of girls are in the bottom four categories, as compared to 10% of boys. This is in contrast to the fact that the girls achieved on average superior GCSE results to the boys, and in particular, to the fact that fewer girls than boys failed to achieve 5 Cs or more. 53.8% of girls failed to achieve this, as compared to 63.9% of boys. So, the unrealistic optimism of some less able pupils seems to be more extreme in the case of boys, although the low proportion of pupils of either sex who place themselves in the bottom four categories may be seen as more striking than the gender difference.

Excellent Above average Average Below average Poor

Male f % 15 6.2 102 42.1 117 48.3 6 2.5 2 0.8

Female f % 10 4.5 62 28.2 139 63.2 9 4.1 0 0

Table 5.8: Self-assessed ability by gender

Table 5.10 shows that boys predicted significantly higher grades for themselves than did girls, controlling for actual GCSE grades (η 2 = 0.022, p ≤ 0.001). Boys also rated their own academic abilities significantly more highly than girls rated their own abilities, controlling for both actual GCSE grades

CHAPTER 5. BELIEFS AND DESIRES

Mostly Mostly Mostly Mostly Mostly Mostly Mostly Mostly

124

Male f % 6 2.5 35 14.6 88 36.8 87 36.4 18 7.5 5 2.1 0 0.0 0 0.0

As As and Bs Bs and Cs Cs and Ds Ds and Es Es and Fs Fs and Gs Gs and ungraded

Female f % 7 3.1 34 15.2 72 32.3 75 33.6 27 12.1 7 3.1 1 0.4 0 0.0

Table 5.9: Self-predicted grades by gender

Qualifications

Parents’ Cultural Capital

Pupils’ Cultural Capital

Ability

η2 p η2 p

Class

Grades

Sex

and school attended (η 2 = 0.019, p ≤ 0.003).

0.022 0.001 0.019 0.003

0.041 0.004 0.036 0.011

0.062 0.000 0.025 0.129

0.037 0.000 0.015 0.010

0.052 0.000 0.025 0.001

Table 5.10: Pupils’ self-predicted grades and estimation of own abilities

If pupils’ perceptions of their own abilities are affected by social class, then this is a serious problem for Breen and Goldthorpe. On their model, pupils’ subjective probabilities of success affect their decisions regarding post-compulsory education, but these subjective probabilities are not distorted by social class, and the class difference in educational participation rates is explained entirely by differences in average ability and resources. Table 5.10 shows that there is a highly significant association between social class and pupils’ estimation of their own abilities, controlling for actual academic achievement. (This is using the six-category class categorisation.) The proportion of the variance in pupils’ self-predicted grades that is explained by social class, controlling for pupils’ actual GCSE grades, is 0.041 (p ≤ 0.004).

CHAPTER 5. BELIEFS AND DESIRES

125

In addition, pupils from the higher social class categories rated their academic abilities significantly more highly than other pupils, controlling for actual GCSE performance and school attended (η 2 = 0.036, p ≤ 0.011). Another factor that may influence pupils’ perceptions of their own abilities is the educational level of their parents. Table 5.10 shows that parents’ qualifications have a significant effect on pupils’ self-predicted grades, controlling for actual grades, with the children of more highly qualified parents overestimating their grades in comparison with other pupils (η 2 = 0.062, p ≤ 0.000). This is using 7 categories for parental qualifications. The effect of parents’ qualifications on pupils’ estimation of their academic abilities compared to others in their school is not significant, however. The level of cultural participation of pupils and their parents might also be expected to have an influence on pupils’ estimations of their own academic ability. Table 5.10 shows that parents’ cultural participation has a significant positive association with both pupils’ self-predicted grades, controlling for actual grades, and with pupils’ evaluation of their general academic abilities, controlling for GCSE grades and school attended. In both cases, the associations with pupils’ own cultural participation are stronger still. Pupils’ cultural capital seems to have a particularly strong association with pupils’ self predicted grades (η 2 = 0.052, p ≤ 0.000). Perhaps cultural participation affects pupils’ self-image, giving them exaggerated confidence in their own abilities. (Recall that pupils’ cultural participation is strongly associated with parents’ cultural participation — Pearson correlation = 0.617, p ≤ 0.000.)

5.5

Conclusions

I have stated that no determinate predictions can be derived from the decision rule of rational choice alone, i.e. we can say that people will behave rationally in pursuit of their desires, on the basis of their beliefs about the situation, but this tells us nothing if we do not know anything about those desires and beliefs. I have described some of the assumptions made by rational choice theorists about the beliefs and desires of individuals regarding education. Rational choice theorists have generally not used subjective data on individuals, preferring to focus on actual behaviour. Perhaps data on outcomes is seen as more reliable than data on attitudes and beliefs. However, I would contend that, in this case, actions do not speak louder than

CHAPTER 5. BELIEFS AND DESIRES

126

words, since we cannot make confident inferences about either the beliefs and desires that form the basis for a decision or the rationality of the decision making process from the action alone. Therefore, an understanding of social class and gender differences in educational participation demands an empirical analysis of the beliefs and desires relevant to the decision of whether to stay on in education or training. The view that attitudes to education vary according to social class gains little support from this research. In order to reveal any significant association between social class and attitudes to education it was necessary to use a bivariate class classification. The association between social class and attitudes to education is small and is limited to those items within the educational attitudes scale that could not be categorised as applying to either the value of education as a positional good, of value in the labour market, or to its worth as an intrinsic good. Furthermore, this small effect disappeared once GCSE performance was controlled for. This is consistent with the view that any social class difference in attitudes to education is actually caused by the comparatively low levels of educational success of working class pupils, rather than vice versa. It could be that previous claims that working class youth disliked or were hostile to education were false. Alternatively, my findings could reflect a real social change, as working class pupils recognise the need for educational credentials in the current labour market. Whereas Willis’ lads could be confident of finding unskilled manual work, young people today have no certainty of getting a job, and certainly cannot expect job security. One might also point to the process of “credential inflation” whereby, since absolute levels of educational attainment are increasing, it is all the more crucial for each individual to gain at least some qualifications in order to keep up. Schools may also play a role, as they are under increasing pressure to improve examination performance, and to reduce the number of pupils who leave school with no qualifications. This could lead teachers to attempt to instil a belief in the value of education in all their pupils, rather than just those who seem likely to do well. Perhaps more surprisingly, there is no relationship between parents’ qualifications and their children’s attitudes to education. In contrast, both parents’ and pupils’ cultural capital have a significant effect on pupils’ attitudes to the intrinsic value of education. This ties in with the idea of the cultured family promoting the value of education in the development of an individual, rather than just its value in the labour market.

CHAPTER 5. BELIEFS AND DESIRES

127

Girls’ are significantly more positive in their attitudes to education than are boys. However, this effect is not large and is limited to the evaluation of education as an intrinsic good. The greater tendency of girls, as compared to boys, to see education as having some intrinsic worth, may improve girls’ levels of motivation in school, as it is not always easy to work hard in pursuit of a distant goal. Boys and girls do not seem to differ in the value they place on education as a good in the labour market. The view put forward by Breen and Goldthorpe (1997) that beliefs are not systematically distorted by social category is not supported by this research. Boys significantly overestimate themselves compared to girls, both in predicting their GCSE results and in evaluating their general academic abilities compared to others at their school. Pupils from higher social class categories significantly overestimate their general academic abilities and their GCSE grades as compared to pupils from lower social class categories. Pupils whose parents are relatively well qualified overestimate their GCSE performance significantly compared to pupils whose parents are less well qualified. Pupils whose parents have relatively high levels of cultural participation significantly overestimate both their GCSE grades and their general academic abilities. The association with pupils own cultural participation is even stronger, suggesting that participation in ‘cultured’ activities gives pupils a high estimation of their own abilities. In sum, the view that the social class differences in educational participation that remain when taking account of initial examination performance can be explained by social class differences in attitudes towards education cannot be supported, since there is no significant association between social class and pupils’ attitudes to education once GCSE performance is controlled for. However, neither can the view that attitudes to education do not vary by any social category be supported, since both gender, and parents’ and pupils’ cultural capital have some impact on pupils’ attitudes to education. Systematic belief distortion may help to explain differentials in educational participation, since it seems that gender and social background do systematically distort pupils’ beliefs about their own abilities. A possible explanation for this is that parents and/or teachers evaluate children differently according to their gender and social background, and this affects pupils’ perceptions of their own abilities. Parents and teachers also have a role in encouraging or discouraging pupils from pursuing further and higher education. Therefore, their beliefs about pupils’ abilities are likely to have a direct impact on pupils’ educational decisions. Of course, the fact that systematic distortion by social class, gender, etc. of pupils’ beliefs about their own academic abil-

CHAPTER 5. BELIEFS AND DESIRES

128

ities occurs does not in itself show that this is even part of the explanation for class and gender differentials in educational participation rates. However, it does suggest that analysis of this possibility is worthwhile. This analysis will be presented in the next chapter.

Chapter 6 Pupils’ Choices 6.1

Introduction

It has been noted that educational participation is associated with social class, and that this association remains once academic ability is controlled for (Shavit and Blossfeld, 1993). Micklewright (Micklewright, 1989) provides evidence that this has also been the case in Britain. Note however that Micklewright’s analysis is based on the NCDS, and the study children reached the minimum school-leaving age of 16 in 1974. This chapter will address the question of social class and gender differences in rates of educational participation post-16. There are three important decisions that pupils make at this stage. 1. Whether or not to stay on in further education, 2. What curriculum to follow (e.g. the academic option of A levels, or the vocational option of a GNVQ), 3. What subjects to take. Previous chapters have outlined the positions taken by various rational choice theorists on the question of educational decision-making. Boudon (Boudon, 1974) distinguishes between the ‘primary’ effects of stratification, which are inequalities in educational attainment occurring before any important educational branching point (such as the decision to stay on in post-compulsory 129

CHAPTER 6. PUPILS’ CHOICES

130

education) has been reached, and the secondary effects of stratification, which are differences in educational participation which remain once the primary effects have been controlled for. In the British context, class differences in GCSE results can be seen as the primary effects of stratification, whereas class differences in rates of participation in further education that remain once GCSE performance has been controlled for can be seen as the secondary effects of stratification. Working within Boudon’s framework, Breen and Goldthorpe (Breen and Goldthorpe, 1997) claim that class differentials in educational survival rates are due to the greater value placed on obtaining service class positions by pupils from service class origins as compared to pupils from non-service class origins, due to the priority attached by all pupils to maintaining their current class status, rather than pursuing upward mobility. (This view is termed relative risk aversion by Breen and Goldthorpe, but, as I have argued, it is better understood as loss aversion). This view implies that service class pupils will attach a higher value to the qualifications that are potentially a gateway to service class positions than do non-service class pupils. Other possible explanations of social class differences in educational survival rates have been suggested, such as the fact that working class pupils evaluate their own abilities less favourably than do service class pupils at the same level of academic performance (see 5). This could mean that working class pupils underestimate their chances of success in further education (by which I mean 16-18 education or training) as compared to similarly able service class pupils. Girls have higher rates of staying on in further education than do boys, although girls and boys opt for the prestigious option of A levels at similar rates. One possible explanation for girls’ relatively high rate of staying on would be if there were greater labour market incentives for girls to stay on than for boys. This would be the case if there were greater job opportunities for male early school leavers than for female early school leavers, for instance. Another possibility is that girls’ higher estimation of the intrinsic value of education may play a role in making them want to stay on in further education, or that the jobs girls aspire to are more likely to require further educational qualifications than are the jobs boys aspire to. Subject choices may also vary by social class and gender. Since some subjects are perceived as more difficult than others (e.g. maths and science are often perceived as particularly hard) perceptions of ability may play a role here. Occupational aspirations can be expected to play a role in determining subject choice, as qualifications in different subject areas are useful in entering

CHAPTER 6. PUPILS’ CHOICES

131

different occupations. So, social class and gender differences in occupational aspirations may help to explain social class and gender differences in subject choice. Since pupils’ occupational aspirations are likely to affect their educational decisions, it is important to examine how these aspirations differ by class and gender, and how these differences may be explained. In chapter 2 I discussed Breen and Goldthorpe’s view that an individual’s current class status is crucial in determining their aspirations, and that individuals are concerned primarily to preserve their current class status rather than to achieve upward mobility. So, how strong is the link between current class status and aspirations? To what extent does the link remain once GCSE performance is taken into account? And is the link between current class and occupational aspirations due entirely to what Breen and Goldthorpe term “relative risk aversion” or could other factors play a role? Also, since in the case of gender differences in aspirations, current class status cannot be the explanation, what other possible explanations are there? For instance, it could be that there are social class and gender differences in values regarding work, effectively leading to different subjective hierarchies of occupations. Working class and female pupils’ lower estimation of their own abilities compared to middle class and male pupils could be another factor in explaining social class and gender differences in aspirations. Social networks are also a potential explanatory factor behind social class differences in occupational aspirations. Having adult “role models” who do a particular type of job may bring that type of job into a pupil’s field of possibilities, and may increase the pupils’ knowledge about how to get into this type of employment. Since parents are likely to have friends who are predominantly of their own class, middle class pupils will know a larger number of adults doing middle class jobs, and will therefore be likely to have a greater awareness of a larger number of middle class jobs than will working class pupils. Cultural participation may play a part in many of the causal mechanisms suggested above, as it is associated with pupils’ self-perceived ability and with their view of the intrinsic value of education. It is also possible that parents’ cultural participation is associated with the family’s social networks.

6.2

Research Questions

1. How are occupational aspirations distributed by social class, gender, parental education and cultural participation? Do pupils’ values about

CHAPTER 6. PUPILS’ CHOICES

132

what is important in a job vary according to social category? 2. What causal mechanisms underlie the differences in occupational aspirations that remain once GCSE performance is controlled for? To what extent are: (a) perceived ability, (b) attitudes to desirable factors in a job, and (c) social networks useful in explaining social class and gender differences in occupational aspirations. Is there an effect of cultural participation, and is it mediated by the above factors? 3. How do choices regarding participation in further education break down by social class and gender? 4. What causal mechanisms underlie the differences in educational participation rates that remain once GCSE performance is controlled for? To what extent are: (a) attitudes to education, (b) perceived ability, and (c) occupational aspirations useful in explaining these differences. Does cultural participation have an effect, and if so, to what extent is it mediated by the above three factors? 5. What explains gender differences in subject choice?

6.3

Occupational Aspirations

Pupils were asked what job they would like to be doing in 10 years time and what job they thought they would actually be doing in 10 years time. Table 6.1 gives their responses to what they would like to be doing in 10 years time according to gender. Boys are more likely to aspire to professional positions, and girls to “associate professional” positions such as teaching, nursing, and social work. No girls

CHAPTER 6. PUPILS’ CHOICES

owner or manager professional associate professional skilled manual unskilled manual armed forces routine non-manual protective services travel and tourism hair and beauty shop work or waiting catering computing animals children/carer arts and media sport fantasy other/don?t know total

133

Male f % 13 5.5

Female f % 11 5.1

Total f % 24 5.4

41 5

17 2.2

28 22

13.0 10.2

69 27

15.4 6.0

27 5

11.6 2.2

0 0

0.0 0.0

27 5

6.0 1.1

8 5

3.4 2.2

1 18

0.5 8.4

9 23

2.0 5.1

9

3.9

2

0.9

11

2.5

1

0.4

10

4.7

11

2.5

3 0

1.3 0.0

9 1

4.2 0.5

12 1

2.7 0.2

6 12 0 0 40 25 2 30

2.6 5.2 0.0 0.0 17.2 10.8 0.9 12.9

0 4 10 27 39 3 1 29

0.0 1.9 4.7 12.6 18.1 1.4 0.5 13.5

6 16 10 27 79 28 3 59

1.3 3.6 2.2 6.0 17.7 6.3 0.7 13.2

232

100.0

215

100.0

447

100.0

Table 6.1: Occupational aspirations by gender

CHAPTER 6. PUPILS’ CHOICES

134

aspire to manual positions, and girls are more likely than boys to aspire to routine non-manual positions. A large proportion of girls aspire to work with children or animals, whereas no boys aspire to this. About 10 per cent of boys aspire to be involved in sport in some way, whereas this is an unusual choice for girls. Table 6.2 shows the same information broken down by social class. Unsurprisingly, service class pupils are more likely to aspire to professional occupations and occupations in the arts and the media than are non-service class pupils. More surprisingly, non-service class pupils are more likely to aspire to ownership or management positions than are service class pupils. The data for ‘what job do you think you will actually be doing in 10 years time’ (not shown in detail here) shows much the same class and gender variations as the data for what pupils would like to be doing, but unsurprisingly, pupils in general are less likely to give the more ambitious or desirable options in response to this question.

6.4

Attitudes to Jobs

One factor that could affect pupils’ ambitions is their view of what are the characteristics of a good job. Pupils were asked “In choosing a job, what things about it do you think are important? ”. They were given a list of options (see questionnaire in appendix B), and asked to tick as many of them as they liked, but then to choose which was most important, second most important and third most important. Table 6.3 shows pupils’ choices of the most important characteristic of a job broken down by sex. By far the most common choice for boys is “It should be well paid”, whereas girls are most likely to choose “It should give me the opportunity of helping others”. Surprisingly, boys are almost twice as likely as girls to say that not getting in the way of their family responsibilities is the most important thing in choosing a job. Perhaps boys see their family responsibilities as being to provide for the family financially, and therefore would see only badly paid or insecure jobs as incompatible with this. Alternatively, perhaps girls see their future family responsibilities as more demanding and therefore accept that a job will necessarily interfere with these, whereas boys, seeing their family

CHAPTER 6. PUPILS’ CHOICES

owner or manager professional associate professional skilled manual unskilled manual armed forces routine non-manual protective services travel and tourism hair and beauty shop work or waiting catering computing animals children/carer arts and media sport fantasy other/don’t know total

135

Service class f % 6 3.9

Non service class f % 16 6.6

37 15

24.2 9.8

29 10

12.0 4.1

4 2

2.6 1.3

18 3

7.4 1.2

2 1

1.3 0.7

7 15

2.9 6.2

1

0.7

9

3.7

2

1.3

8

3.3

2 0

1.3 0.0

9 0

3.7 0.0

0 2 2 5 43 11 1 17

0.0 1.3 1.3 3.3 28.1 7.2 0.7 11.1

6 14 7 13 30 13 2 33

2.5 5.8 2.9 5.4 12.4 5.4 0.8 13.6

153

100.0

242

100.0

Table 6.2: Occupational aspirations by parental class

CHAPTER 6. PUPILS’ CHOICES

136

Male hands respect outdoor thought excel family control variety pay creative convenient people clean promotion own boss secure not responsible helping

f 7 9 6 8 7 29 1 7 93 15 1 5 3 6 1 21 1 19

Female % 2.9 3.8 2.5 3.3 2.9 12.1 0.4 2.9 38.9 6.3 0.4 2.1 1.3 2.5 0.4 8.8 0.4 7.9

f 4 6 1 12 14 15 2 10 32 22 3 18 0 6 2 28 0 45

% 1.8 2.7 0.5 5.5 6.4 6.8 0.9 4.5 14.5 10.0 1.4 8.2 0.0 2.7 0.9 12.7 0.0 20.5

Table 6.3: Occupational values by gender

responsibilities as smaller, may think that a job that interfered with these would be overly demanding. The above table includes many small cells, and it is easier to see patterns in the responses if the items are grouped together into categories. Below is a categorisation, designed to group together items reflecting similar values. Control and respect It should make others respect me It should give me control over others The job should let me be my own boss Cronbach’s α for this subscale = 0.40

CHAPTER 6. PUPILS’ CHOICES

137

Intellectual stimulation It should involve using my head and need thought and concentration It should allow me to excel It should allow me to be creative It should involve variety α = 0.45 Material Factors: Money, security, prospects It should be well paid It should be a secure job It should offer chances of promotion α = 0.50 Manual The job should involve working with my hands It should be an outdoor job α = 0.38 People It should give me the chance of being with other people It should give me the opportunity of helping others α = 0.36 Not too arduous It should not get in the way of my family responsibilities It should have convenient hours and conditions It should be a clean job It should not have too much responsibility α = 0.46 A factor analysis is shown in table 6.4. This shows that component 1 loads highly on those items intended to reflect material values, but also on some of those items designed to reflect “control and respect” and a job that is “not too arduous”. Nonetheless, I think that I am justified in separating these three categories of values. For example, although a Cronbach’s α of 0.21 between “It should be well paid” and “It should have convenient hours and conditions” shows a fair likelihood that an individual who selected one of these items also selected the other, in practice these values may be in conflict in actually choosing a job. Therefore, in examining which items respondents

CHAPTER 6. PUPILS’ CHOICES

Hands Respect Outdoor Thought Excel Family Control Variety Pay Creative Convenient Being with people Clean Promotion Own boss Secure Not responsibility Helping others

138

1 0.18 0.56 0.06 0.28 0.33 0.39 0.43 0.25 0.46 0.26 0.44 0.28

2 0.50 0.21 0.58 −0.08 0.06 −0.31 0.31 −0.14 −0.06 0.33 −0.21 −0.03

Component 3 4 −0.22 0.33 −0.03 0.04 −0.17 0.18 0.45 0.34 0.64 −0.12 −0.31 0.07 −0.08 0.07 0.50 0.08 −0.07 −0.33 0.36 0.18 −0.34 −0.25 −0.05 0.36

0.48 0.55 0.23 0.46 0.39

−0.07 −0.29 0.49 −0.32 0.25

−0.08 0.18 0.05 −0.15 −0.21

−0.26 −0.04 −0.30 0.15 −0.31

−0.17 0.03 0.19 0.27 0.09

0.26 −0.26 0.17 −0.24 0.14

0.23

−0.16

−0.26

0.69

0.00

0.07

5 0.39 −0.33 0.15 0.06 −0.01 0.39 −0.42 0.31 −0.03 0.13 0.09 −0.55

6 −0.01 −0.12 −0.37 −0.25 −0.14 0.15 −0.02 0.35 −0.42 0.39 0.25 0.27

Table 6.4: Factor analysis of occupational values

selected as most important (rather than those simply selected as important) it seems fair to put these items into different categories. Component 2 loads most highly on the two items reflecting manual work. Component 3 scores most highly on the four items reflecting a desire for ‘intellectual stimulation’. Component 4 loads most highly on the two items reflecting being with and helping others. Components 5 and 6 seem to just reflect statistical noise. Table 6.5 shows that the material aspects of a job (money, security and prospects) are valued as the most important factor in choosing a job by nearly 50% of the boys, as compared to nearly 30% of the girls. The three categories that were most popular with the girls were material factors, “people” (being with others or helping others) and intellectual stimulation. Boys were considerably less likely than girls to choose people or intellectual stimulation as the most important factor in choosing a job. Boys were more

CHAPTER 6. PUPILS’ CHOICES

Control Intellectual Material Manual People Not arduous Total

Male f % 11 4.5 37 15.3 120 49.6 13 5.4 24 9.9 34 14.0 242 100.0

139 Female f % 10 4.5 58 26.0 66 29.6 5 2.2 63 28.3 18 8.1 223 100.0

Table 6.5: Occupational value categories by gender

likely than girls to choose one of the items in the “not too arduous” category than were girls, and, unsurprisingly, boys were more likely to say that being outside or working with their hands was the most important factor. The proportion of pupils valuing control and respect most highly was surprisingly small, and equally so for boys and girls.

Control Intellectual Material Manual People Not arduous Total

Service class f % 5 4.8 40 29.9 105 37.6 2 5.2 28 17.1 16 13.1 157 100.0

Non service class f % 12 3.2 47 15.9 59 41.1 13 1.3 43 17.8 33 10.2 251 100.0

Table 6.6: Occupational values by parental class

Table 6.6 shows that both service class and non-service class pupils were most likely to choose money, security or prospects as the most important factor in choosing a job. The main difference between the service class and non-service class is in the likelihood of choosing intellectual stimulation as the most important factor in choosing a job. Service class pupils were nearly twice as likely to give an item in this category as their first choice compared to other pupils. (This social class difference may well be expected to disappear once educational attainment at GCSE is controlled for). Unsurprisingly, service class pupils are less likely than others to choose working outdoors or

CHAPTER 6. PUPILS’ CHOICES

140

with their hands as the most important factor in choosing a job. Overall though, pupils from the service class and non-service class seem quite similar in their choices of the most important factor in choosing a job.

6.5

Occupational Aspirations and Values

Breen and Goldthorpe put forward a model of social class differences in occupational aspirations which they describe as “relative risk aversion”, although, as I have argued, it is better understood as ‘loss aversion’. On this model, individuals aim primarily to avoid social demotion. The idea that people are keener to avoid losses than to make gains has been found useful in explaining the behaviour of subjects in certain experiments in social psychology (Kahneman et al., 1982). However, it is very difficult to see how the view that “loss aversion” is the mechanism behind social class differences in occupational aspirations could be tested directly. Perhaps the most we can say is that if “loss aversion” is the mechanism behind the lower aspirations of working class pupils, there should be a direct association between social class and aspirations. The association should be unmediated, especially by cultural factors. One alternative to the “loss aversion” mechanism is the possibility that cultural factors, such as attitudes to what constitutes a desirable attribute in an occupation, may mediate the effect of social class. Another possibility is that knowledge of jobs at the higher end of the social spectrum may be associated with social class, and this may limit the aspirations of lower class pupils. Both attitudes towards and knowledge of employment may be associated with cultural participation. It seems likely that aspirations are adjusted in response to perceived possibilities. One would expect academic performance, and perhaps also subjective perceptions of ability to affect individuals’ occupational aspirations. If no association between social class and aspirations remains once these factors are controlled for, then there will be no need for an additional explanation in terms of “loss aversion”. As discussed in chapter 2, the idea of “loss aversion” does not apply readily to gender differences in educational and occupational aspirations and outcomes. An analysis in terms of the alternative mechanisms suggested above may shed more light on gender differences.

CHAPTER 6. PUPILS’ CHOICES

6.5.1

141

Occupational values

Gender and social class differences in responses to the question of what is the most important factor in choosing a job have been described. I will now model these responses in order to show whether these gender and social class effects are mediated by other factors. Table 6.7 shows the model for whether pupils’ chose an item reflecting intellectual stimulation as most important. Model 1 shows that social class is significant, as pupils from the upper service class are more likely to have chosen intellectual stimulation as the most important factor in choosing a job. Parents’ education is insignificant, and the only significant parameter for this variable is the missing category. Therefore, this variable is removed from the analysis (bear in mind though that parents’ education may have been significant if there had been less missing data). Boys are significantly less likely than girls to have chosen intellectual stimulation as the most important factor in choosing a job. Model 2 shows the effects of parental cultural participation, and parents’ social contacts. Parents’ social contacts are insignificant. Parents’ cultural participation is significantly positively associated with pupils’ valuing intellectual stimulation most highly. This mediates the social class effect, rendering it insignificant. Model 3 shows that the effect of parents’ cultural participation is mediated by pupils’ own cultural participation. Model 4 shows that the effect of pupils’ cultural participation is itself mediated by pupils’ GCSE grades. Model 5 shows that the effect of GCSE grades becomes insignificant once pupils’ cultural knowledge is included in the model. So, pupils with high levels of cultural knowledge are more likely to say they rate intellectual stimulation most highly in choosing a job. Boys are less likely than girls are to value this most highly. Table 6.8 shows the model for whether pupils selected an item reflecting material values as the most important in choosing a job. Model 1 shows that neither social class nor parental education is significant as a whole. Boys are significantly more likely to value material factors most highly in choosing a job, and school attended is also significant. Parents’ cultural participation is negatively associated with valuing material factors

0.09(0.60) 0.90(0.54) 0.36(0.54) 0.48(0.52) 0.37(0.58) 0.14(0.63)

−0.93(0.37)* 0.49(0.40) −0.07(0.35) −0.31(0.44)

0.04(0.60) 1.13(0.55)* 0.42(0.55) 0.52(0.53) 0.51(0.57) 0.28(0.63)

−0.65(0.43) 0.20(0.45) 0.15(0.45) −0.35(0.43) −0.97(0.37)** 0.58(0.40) 0.09(0.35) −0.42(0.44)

0.44(0.40) −0.05(0.36) −0.34(0.44)

−0.87(0.37)*

0.40(0.58) 0.10(0.63)

0.49(0.52)

0.50(0.41) 0.08(0.36) −0.14(0.45)

−0.81(0.38)*

0.41(0.62) 0.24(0.66)

0.50(0.56)

0.23(0.59)

0.19(0.63) 0.79(0.60)

−0.01(0.60) 0.84(0.55) 0.31(0.54)

Model 4 B(s.e) −3.30(1.06)**

Model 3 B(s.e) −2.37(0.53)***

Table 6.7: Models of desire for intellectual stimulation (continued overleaf. . . )

Constant Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Parents’ qualifications Missing Degree Intermediate O-level Male School School 1 School 2 School 3

Model 2 B(s.e) −2.12(0.52)***

Model 1 B(s.e) −1.38(0.53)**

0.51(0.42) −0.14(0.37) −0.14(0.45)

−1.04(0.39)**

0.45(0.62) 0.31(0.66)

0.52(0.57)

0.28(0.59)

0.34(0.64) 0.75(0.61)

Model 5 B(s.e) −3.37(0.64)***

40.8 465

43.8 465

0.07(0.06)

Model 2 B(s.e) 0.09(0.04)*

47.5 465

0.13(0.06)*

Model 3 B(s.e) 0.05(0.04)

52.1 458

0.01(0.01)

0.02(0.01)* 0.10(0.23)

0.09(0.04)* 0.01(0.02) 61.6 461

0.09(0.06)

Model 5 B(s.e) 0.01(0.05)

0.11(0.06)

Model 4 B(s.e) 0.03(0.04)

Table 6.7: Models of desire for intellectual stimulation (. . . continued from overleaf)

Parents’ cultural participation Parents’ social contacts Pupils’ cultural participation GCSEs Self-assessed ability Cultural knowledge Language χ2 N

Model 1 B(s.e)

Class Missing Service class higher Service class lower Routine non manual Petty bougeois Skilled manual Parents’ qualifications Missing Degree A-level Vocational O-level CSE Sex (M) School School 1 School 2 School 3 Parents’ cultural participation 0.81(0.27) ∗ ∗ ∗∗ 0.61(0.28)∗ 0.56(0.32) 0.98(0.30) ∗ ∗∗ −0.05(0.03)

Model 2 B(s.e)

0.78(0.27) ∗ ∗ ∗∗ 0.64(0.29)∗ 0.56(0.33) 1.02(0.29) ∗ ∗∗ −0.02(0.03)

Model 3 B(s.e)

Table 6.8: Material values (continued overleaf. . . )

−0.25(0.36) 0.36(0.48) −0.38(0.49) −0.15(0.35) −2.15(1.08)∗ −0.40(0.42) 0.87(0.28) ∗ ∗ ∗∗ 0.50(0.30) 0.60(0.34) 1.02(0.30) ∗ ∗∗ −0.08(0.03)∗

0.43(0.39) 0.47(0.47)

0.34(0.44)

0.51(0.43)

0.07(0.43) 0.23(0.46)

Model 1 B(s.e)

0.83(0.28) ∗ ∗ ∗∗ 0.56(0.29) 0.55(0.33) 1.14(0.30) ∗ ∗∗ −0.02(0.04)

Model 4 B(s.e)

Parents’ social contacts Pupils’ cultural participation GCSEs Cultural Knowledge Vocabulary χ2 N 43.31 ∗ ∗∗ 465

52.99 ∗ ∗∗ 465

47.29 ∗ ∗∗ 465

−0.12(0.05)∗

Model 3 B(s.e)

Table 6.8: Material values (. . . continued from overleaf)

Model 2 B(s.e) −0.07(0.04)

Model 1 B(s.e)

−0.03(0.02) 50.80 ∗ ∗∗ 461

0.01(0.01) 0.00(0.03)

−0.12(0.04)∗

Model 4 B(s.e)

CHAPTER 6. PUPILS’ CHOICES

146

most highly. Model 2 shows that parents’ social contacts are not significantly associated with material values, but the inclusion of this variable renders parental cultural participation insignificant. There seems to be a problem of collinearity with these two variables, in that each is significant in the absence of the other, but neither is significant in the presence of the other. It is unsurprising that these two items should be highly correlated, as people are likely to choose friends who share their interests. Model 3 shows that pupils’ cultural participation is negatively associated with material values, and mediates the effect of parents’ cultural participation. Model 4 shows that GCSE performance, cultural knowledge scores, and vocabulary scores have no significant association with material values. In sum, boys are more likely than girls to say they value material factors most highly in choosing a job. It could be that boys are still disposed to see themselves as potential breadwinners for a family. Cultural participation is negatively associated with the choice of material factors as being most important in choosing a job. It could be that this is because those engaged in cultural participation value intellectual stimulation more highly, or that the choice of material factors as most important conflicts with a cultivated self-image. Table 6.9 shows the model for valuing control and respect most highly in choosing a job. Model 1 shows that neither social class, parental qualifications, sex, school, parental cultural participation, nor parental social contacts are significant. Model 2 shows that pupils’ cultural participation is negatively associated with choosing an item reflecting respect or control as most important in choosing a job. In model 3, this effect is mediated by GCSE grades. Model 4 shows that neither cultural knowledge nor vocabulary scores are significant. The only variable associated with choosing being with or helping others as the most important factor in choosing a job is gender. In the full model, the coefficient for males as compared to females is −0.28 (0.36), significance level 0.0003. No variable except school attended is associated significantly with the choice of items reflecting a non-arduous job as most important, so it is unnecessary to show a table for this item. Working outside or with one’s hands was chosen as the most important factor in choosing a job by only 18 people, so providing a model for this variable

0.19(0.73) 0.55(0.72) −0.28(0.71)

7.44(15.91) 6.99(15.94) 8.73(15.91) 7.92(15.90) 7.95(15.94) 7.96(15.91) 0.26(0.62)

−0.33(0.70) −1.53(1.15)

−0.06(0.80) −0.33(0.95) −0.48(0.84) −1.69(1.17)

Model 2 B(s.e)

Model 3 B(s.e)

Table 6.9: Control and respect (continued overleaf. . . )

Class Missing Service-class higher Service-class lower Routine non manual Petty Bourgeois Skilled manual Parents’ qualifications Missing Degree A-level Vocational O-level CSE Sex (M) School School 1 School 2 School 3

Model 1 B(s.e)

Model 4 B(s.e)

20.469 465

0.02(0.12)

6.218∗ 465

−0.25(0.11)∗

Model 2 B(s.e)

11.578 ∗ ∗ 461

−0.03(0.02) −0.02(0.07)

−0.03(0.01)∗

0.01(0.07) 11.723∗ 461

−0.17(0.12)

Model 4 B(s.e)

−0.18(0.11)

Model 3 B(s.e)

Table 6.9: Control and respect (. . . continued from overleaf)

Parents’ cultural participation Parents’ social contacts Pupils’ cultural participation GCSEs Cultural Knowledge Vocabulary χ2 N

Model 1 B(s.e) 0.03(0.08)

CHAPTER 6. PUPILS’ CHOICES

149

would be rather pointless, although there is a significant negative association between GCSE grades and choosing this item. So, there are no direct social class effects on pupils’ responses as to what the most important factor in choosing a job is. Pupils’ from service class backgrounds are more likely to choose intellectual stimulation as being most important, but this effect is mediated by cultural participation, which in turn is mediated by GCSE grades and cultural knowledge. Cultural participation is negatively associated with valuing material factors and control and respect most highly. Boys are less likely than are girls to value intellectual stimulation or being with or helping others most highly, and are more likely to value material factors as most important in choosing a job.

6.5.2

Occupational Aspirations

In order to model pupils’ occupational aspirations as a bivariate logistic regression, pupils’ aspirations have to be categorised simply as “high” or “low”. I have categorised positions of ownership and control, professional and associate professional positions, and positions in the arts and media as “high”, and all others as “low”. I think the positions I have categorised as “high” approximately correspond to service class positions, but I realise the imprecise and crude nature of this measure. Ultimately, one cannot quite get round the fact that most people do not respond to questions about their aspirations with answers such as “manager in a company with 50+ employees”. Table 6.10 shows the model for pupils’ aspirations. Model 1 shows that pupils from higher service class backgrounds had significantly increased aspirations. Boys have significantly lower aspirations than do girls on this model, but bear in mind that the more detailed breakdown of aspirations given earlier showed girls to be more likely than boys to aspire to associate professional positions, but less likely to aspire to professional positions. Model 2 shows that there is a positive association between parents’ cultural participation and pupils’ aspirations, but parents’ social contacts are not significant. This is perhaps surprising, as one might have thought that parents’ social contacts would affect the pupils’ knowledge of different types of occupations, and therefore affect their aspirations. However, it could be that cultural participation has a greater effect than social contacts on people’s knowledge of the occupational structure, and, in particular, of the prestigious positions within it. Model 3 shows that pupils’ cultural participation is highly significantly as-

Constant Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Parents’ qualifications Missing Degree Intermediate O-level Male School School 1 School 2 School 3 Parents’ cultural participation 0.00(0.39) 0.42(0.44) −0.65(0.49)

−0.05(0.40) 0.46(0.43) −0.63(0.49)

−0.76(0.31)* ** 1.03(0.34)** −0.42(0.36) −0.46(0.36) −0.04(0.04)

0.61(0.47) −0.69(0.51)

−0.02(0.41)

0.54(0.43)

Model 3 B(s.e) −1.72(0.47)*** ** −0.64(0.49) 1.18(0.47)*

−0.77(0.33)* ** 1.10(0.36)** −0.37(0.37) −0.17(0.38) −0.06(0.04)

0.43(0.50) −0.69(0.54)

−0.26(0.44)

0.20(0.47)

Model 4 B(s.e) −2.01(0.97)* * −0.55(0.52) 0.80(0.50)

Table 6.10: Job aspirations (continued overleaf. . . )

−0.83(0.29)** ** 1.07(0.32)** −0.38(0.33) −0.49(0.35) 0.08(0.03)*

0.57(0.41)

0.49(0.42)

−0.08(0.35) 0.69(0.41) 0.32(0.40) 0.30(0.37) −0.90(0.29)** *** 1.08(0.32)** −0.37(0.33) −0.59(0.34)

Model 2 B(s.e) −0.64(0.39) *** −0.56(0.47) 1.28(0.45)**

Model 1 B(s.e) −0.38(0.41) ** −0.58(0.46) 1.28(0.46)**

−0.90(0.34)** ** 1.21(0.37)** −0.50(0.38) −0.12(0.39) −0.09(0.04)*

0.50(0.50) −0.67(0.54)

−0.18(0.45)

0.26(0.47)

Model 5 B(s.e) −3.17(0.57)*** * −0.41(0.52) 0.89(0.51)

Parents’ social contacts Pupils’ cultural participation Occupational values Control and respect Intellectual stimulation Material Manual Non arduous GCSEs Self-assessed ability Cultural knowledge Language χ2 N 92.0 447

90.7 447

0.61(0.37) 0.32(0.33) −2.00(1.18) 0.48(0.44) 0.03(0.01)*** 0.19(0.22)

0.78(0.36)* 0.38(0.32) −1.98(1.13) 0.36(0.42)

154.2 435

0.00(0.65)

−0.24(0.62)

131.9 441

0.25(0.06)***

Model 4 B(s.e)

0.29(0.06)***

Model 3 B(s.e)

Table 6.10: Job aspirations (. . . continued from overleaf)

Model 2 B(s.e) 0.02(0.05)

Model 1 B(s.e)

0.09(0.04)* 0.01(0.02) 163.3 437

0.34(0.34) −1.90(1.20) 0.51(0.45) 0.03(0.01)**

0.51(0.38)

0.09(0.65)

0.24(0.06)***

Model 5 B(s.e)

CHAPTER 6. PUPILS’ CHOICES

152

sociated with pupils’ aspirations, and this mediates the effect of parents’ cultural participation. Model 4 shows that, overall, the effect of occupational values is insignificant. Only the “intellectual stimulation” value is significantly associated with pupils’ aspirations (the comparison value here is “non-arduous”). Model 5 shows that the effect of GCSE performance is significant, but selfassessed ability has no effect controlling for this. The inclusion of GCSE performance renders the social class effect insignificant. This would seem to show that the social class effect is due to the revision of aspirations in line with chances of success. Crucially, this leaves no need for an explanation in terms of “relative risk aversion”. Model 6 shows that cultural knowledge has a significant effect on pupils’ aspirations, but vocabulary scores do not. The effect of cultural knowledge partially mediates the effect of GCSE performance. Cultural knowledge may be associated with knowledge of prestigious occupational positions — both of what these positions entail, and how people get into these positions. It may also be that pupils with high levels of cultural knowledge perceive themselves (quite possibly correctly so) as being more likely than others to achieve these positions.

6.6

Educational Choices

This section will describe the breakdown of pupils’ intentions regarding further education, before giving an analysis in the form of logistic regressions.

6.6.1

Choosing to stay on in different courses

Table 6.11 shows that A levels were the most common option, with nearly half the pupils in my sample intending to pursue this option. Boys and girls were about equally likely to say they intended to study for A levels. However, girls seemed to be considerably more likely to intend to take GNVQs than boys were, and correspondingly less likely to say they intend not to pursue any further education at all. Table 6.12 shows pupils’ stated intentions according to their social class status. Service class students were very likely to intend to do A levels, with

CHAPTER 6. PUPILS’ CHOICES

A levels GNVQ GCSE retakes Leave Total

Male f % 113 47.7 70 29.5 6 2.5 48 20.3 237 100.0

153 Female f % 105 47.5 89 40.3 4 1.8 23 10.4 221 100.0

Total f % 218 47.6 159 34.7 10 2.2 71 15.5 458 10.0

Table 6.11: Educational intentions and gender

A levels GNVQ GCSE retakes Leave Total

Service class f % 119 75.8 23 14.6 3 1.9 12 7.6 157 100.0

Non service class f % 82 33.6 110 45.1 4 1.6 48 19.7 244 100.0

Missing f % 17 29.8 26 45.6 3 5.3 11 19.3 57 100.0

Table 6.12: Educational intentions and social class around three quarters of service class pupils giving this response, as compared with only about a third of non service class pupils. Non service class pupils were considerably more likely both to opt for GNVQs and to reject further education than were service class pupils. Since the amount of missing data on parents’ jobs is not small, this category has been included in the table. It seems clear that those pupils who did not give an occupation (current or previous) for either parent approximate the non service class rather than the mean. This makes sense, as pupils who’s parents have high status occupations may be both more willing to respond to this item on the questionnaire, and less likely to give answers that are too unclear to be coded. In sum, pupils from lower social classes were more likely to intend to leave education altogether after their GCSEs, and more likely to opt for GNVQs rather than A levels if they intended to stay on. Male and female pupils were similarly likely to opt for A levels, but female pupils were considerably more likely to intend to stay on and take a GNVQ, rather than leave.

CHAPTER 6. PUPILS’ CHOICES

154

Reliability of Intentions Unfortunately, two of the schools in my sample were unable to give me adequate information on the actual post-GCSE destinations of their pupils, and therefore, I am using the pupils stated intentions as the outcome variable in the analyses to follow. However, the reliability of the pupils’ stated intentions can be tested using data from the two schools who were able to provide this information.

Stated Intention

Not A levels A levels Total

Outcome Not A levels A levels f % f % 58 71.6 39 30.5 23 81

28.4 100.0

89 128

69.5 100.0

Table 6.13: Reliability of A level choices

Table 6.13 shows that about 72% of those who stated that they would not do A levels did not, and about 70% of those pupils who stated that they would do A levels actually embarked on an A level course.

Stated Intention

Not FE FE Total

Outcome Not further education Further education f % f % 22 40.0 13 8.4 33 60.0 141 91.6 55 100.0 154 100.0

Table 6.14: Reliability of further education choices

Table 6.14 shows that about 92% of those who said that they intended to pursue further education actually did so. However, only 40% of those who did not intend to continue in education or training actually left. This low level of reliability may be accounted for by the very low number (only 55) who are confirmed as not having pursued further education.

−0.33(0.62) −0.76(0.52)

−0.07(0.49)

0.00(0.59)

−0.44(0.53) −0.39(0.63)

−0.73(0.33)* ** 1.13(0.41)** 1.26(0.56)* 0.42(0.36)

−0.23(0.41) 0.51(0.61) 0.20(0.57) −0.02(0.46) −0.73(0.34)* * 1.10(0.43)** 1.38(0.58)* 0.34(0.36) −0.70(0.33)* ** 1.16(0.41)** 1.32(0.56)* 0.44(0.36)

Model 3 B(s.e) −1.08(0.49)* 0.04(0.01)***

−0.81(0.35)* * 1.12(0.42)** 1.10(0.56) 0.38(0.37)

Model 4 B(s.e) −1.61(1.64) 0.04(0.01)***

Table 6.15: Pupils’ intention to puruse further education (continued overleaf. . . )

Constant GCSEs Parents’ qualifications Missing Degree Intermediate O-level Male School School 1 School 2 School 3 Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual

Model 2 B(s.e) −0.68(0.45) 0.04(0.01)***

Model 1 B(s.e) −0.17(0.62) 0.04(0.01)***

−0.76(0.35)* * 0.96(0.44)* 1.19(0.56)* 0.53(0.37)

Model 5 B(s.e) −0.52(0.52) 0.03(0.01)**

91.4 443

87.1 443

−0.01(0.07)

0.08(0.05)

Model 2 B(s.e) 0.86(0.28)**

89.4 443

0.15(0.07)*

Model 3 B(s.e) 0.91(0.29)**

95.0 440

0.04(0.02)

0.41(0.27)

0.12(0.07)

Model 4 B(s.e) 0.82(0.29)**

*

0.07(0.08)

Model 5 B(s.e) 0.70(0.30)*

2.49(1.04)* 98.1 425

−0.69(0.65) 1.84(0.76)* 0.69(0.80)

Table 6.15: Pupils’ intention to pursue further education (. . . continued from overleaf)

Family intact Sibling number Parents’ cultural participation Parents’ social contacts Pupils’ cultural participation Self-assessed ability Educational attitudes Occupational aspirations Owner or manager Professional Associate professional Arts and media χ2 N

Model 1 B(s.e) 0.92(0.29)** 0.02(0.08)

CHAPTER 6. PUPILS’ CHOICES

6.6.2

157

Further Qualifications

I will model pupils’ decision as to whether to stay on in further education (table 6.15). However, only 81 of the pupils in my sample intended not to pursue any form of further education, and, perhaps as a result of this, the overall model fit is not good. Model 1 shows that, unsurprisingly, GCSE results have a strong positive association with the decision to stay on rather than leave. Controlling for this, there is no significant social class effect, and no significant effect of parents’ qualifications. The effect of gender is (just) significant at the 0.05 level, with boys being less likely to decide to stay on as compared to girls. School attended was also significant. The number of siblings has no significant effect. Pupils from households with an “intact” family structure (i.e. living with both mother and father) are significantly more likely than those from “non-intact” families to stay on in further education. One can break down this effect to examine the effect of stepfamilies and single parent families. The effect of being in a step-parent family (B = −1.16(0.36), p > 0.002) seems to be more significantly negative than the effect of being in a single parent family (B = 0.84(0.36), p > 0.02). This is surprising in the sense that stepfamilies are generally not economically disadvantaged to the extent that single parent families are. Interestingly, Kiernan (Kiernan, 1992) finds, in her analysis of NCDS data, that young people from stepfamilies formed after death or divorce were most likely to leave home early, and for reasons of friction. It could be that stepparents are reluctant to support young people financially once compulsory education has come to an end. (However, in the case of the NCDS cohort, remarriage did not appear to have a negative effect on educational outcomes for children (Elliott and Richards, 1991). It may well be that there is a difference in outcomes between children whose stepparents are in a stable relationship with their natural parents, and children whose stepparents are not in a stable relationship with their natural parents. Some of the latter may have to put up with a series of “boyfriends” or “girlfriends” of their natural parent). Model 2 shows that neither parents’ cultural participation nor parents’ social contacts have any significant effect. Model 3 shows that pupils’ cultural participation is significantly positively associated with deciding to stay on in further education. Model 4 shows that pupils’ self-assessed ability and attitudes to education are not significant. Model 5 shows the effect of pupils’ occupational aspirations. Those who

CHAPTER 6. PUPILS’ CHOICES

158

aspire to professional and arts and media positions are significantly more likely to decide to stay on in further education, controlling for GCSE results, etc. So, GCSE performance is an important factor in determining intentions to pursue further qualifications. The overall model fit is relatively poor, and it may be that the small number of pupils who did not intend to go into further education accounts for this, and (possibly) for the lack of either a social class effect or a parental education effect. Given the lack of these effects, the effect of family structure is striking. The gender effect on staying on is not mediated by any of the other variables, and remains significant in the final model. Since this model does not seem to explain the gender effect, one must look for alternative explanations. One possibility would be that 16 year old boys could simply be better able than girls to find relatively attractive employment, perhaps with some element of on the job training.

6.6.3

A Levels

Model 1 (table 6.16) shows that, unsurprisingly, the effect of GCSE attainment on the decision to do A levels is highly significant. There is no significant sex effect. There is a significant school effect, which could be due to a peer group effect, a school policy effect, or an effect of uncontrolled for parental inputs. Having a parent who is a graduate is highly positively associated with the decision to do A levels. Since A levels are the traditional route to university, it may be that the children of graduates are simply brought up assuming that they will do A levels. (Note that the “missing” category seems highly similar to the “no degree” category.) It is probably safe to assume that respondents who were unable or unwilling to state their parents’ educational level were rarely the children of graduates. Social class has no effect once parents’ qualifications are included in the model. (If we remove parents’ qualifications from the model, a significant advantage for the upper service class is revealed.) The effects of parents’ social contacts and cultural participation are insignificant (model 2). Model 3 shows that pupils’ own cultural participation has a significant positive association with the decision to do A levels. Model 4 shows that the effect of pupils’ cultural participation is mediated by the effects of pupils’ self-assessed ability and pupils’ attitudes to education. Cul-

Constant GCSEs Parents’ class Missing Service class higher Service class lower Routine non manual Petty bourgeois Skilled manual Parents’ qualifications Missing Degree Intermediate O-level Family intact Sibling number Male *** −0.15(0.42) 1.82(0.52)** 0.17(0.46) −0.22(0.42)

−0.18(0.36)

−0.19(0.42) 1.79(0.54)** 0.10(0.46) −0.23(0.41)

−0.18(0.36)

Model 3 B(s.e.) −4.37(0.60)*** 0.09(0.01)***

**

Model 2 B(s.e.) −3.97(0.58)*** 0.09(0.01)***

−0.18(0.36)

−0.15(0.42) 1.82(0.52)** 0.17(0.46) −0.22(0.42)

***

Model 4 B(s.e.) −4.37(0.60)*** 0.09(0.01)***

−0.41(0.39)

−0.16(0.44) 1.93(0.53)*** −0.01(0.49) −0.23(0.44)

***

Model 5 B(s.e.) −7.01(1.75)*** 0.08(0.01)***

Table 6.16: Pupils’ intention to pursue A-levels (continued overleaf. . . )

0.03(0.43) 1.84(0.56)** 0.26(0.48) −0.17(0.43) 0.04(0.29) 0.05(0.08) −0.09(0.37)

−0.61(0.58) −0.78(0.61) **

0.05(0.50)

0.29(0.56)

−0.48(0.58) 0.43(0.63)

Model 1 B(s.e.) −4.05(0.69)*** 0.09(0.01)***

−0.67(0.43)

−0.12(0.47) 2.01(0.57)*** 0.25(0.53) −0.20(0.48)

***

Model 6 B(s.e.) −6.81(1.85)*** 0.07(0.01)***

School School 1 School 2 School 3 Parents’ cultural participation Parents’ social contacts Pupils’ cultural participation Self-assessed ability Educational attitudes Occupational aspirations Owner or manager Professional Associate professional Arts and media χ2 N 272.8 454

−0.02(0.06)

Model 2 B(s.e) ** 1.27(0.39)** 0.06(0.41) −0.82(0.45) 0.08(0.05)

275.8 454

0.15(0.06)*

Model 3 B(s.e) ** 1.25(0.40)** 0.06(0.40) −0.81(0.44)

275.8 454

0.15(0.06)*

Model 4 B(s.e) ** 1.25(0.40)** 0.06(0.40) −0.81(0.44)

0.09(0.02)***

0.08(0.02)***

297.5 451

0.64(0.28)*

0.58(0.26)*

1.03(0.42)* 310.6 433

−1.57(0.93) 1.83(0.51)*** 0.23(0.57)

***

0.07(0.08)

Model 6 B(s.e) ** 1.25(0.46)** −0.53(0.47) −1.34(0.53)*

0.11(0.07)

Model 5 B(s.e) ** 1.29(0.42)** −0.32(0.43) −1.08(0.48)*

Table 6.16: Pupils’ intention to pursue A-levels (. . . continued from overleaf)

272.3 443

Model 1 B(s.e) *** 1.40(0.41)** 0.34(0.42) −0.82(0.45)

CHAPTER 6. PUPILS’ CHOICES

161

tural participation is associated with confidence in one’s own abilities and a positive attitude towards education, and this seems to explain the effect of pupils’ cultural participation. Model 5 shows the effect of pupils’ occupational aspirations. Aspirations to work in professional or arts and media positions are strongly positively associated with the decision to do A levels. In sum, GCSE performance is extremely important in determining pupils’ decisions to do A levels. This is unsurprising, as schools and colleges generally require a minimum of 4 or 5 C grades at GCSE before allowing a student to enrol for an A level course. In addition, students’ perceptions of their chances of success in an A level course are likely to be largely determined by their knowledge of their own abilities. (Note that students’ subjective perceptions of their own abilities have a significant impact, controlling for actual GCSE performance as measured by GCSE success.) Parents’ education has a significant impact on A level plans, mediating the effect of social class. This could suggest that it is important to parents with degrees, and to their children, that they are not downwardly mobile educationally. This may be more important than the fear of downward occupational mobility, contrary to the model put forward by Breen and Goldthorpe (1997).

6.6.4

Subject Choices

This section will outline the distribution of subject choices according to social class and gender.

Arts Science Humanities Languages Social science

Male f % 59 52.2 71 62.8 40 35.4 8 7.1 10 8.8

Female f % 61 55.0 45 40.5 51 45.9 17 15.3 37 33.3

Table 6.17: A level subject choice by gender

Table 6.17 shows the proportion of the 113 male and 111 female students

CHAPTER 6. PUPILS’ CHOICES

162

intending to do A levels who stated that they intended to take at least one subject from each of the following subject groups: arts, sciences, languages, humanities, and social sciences. (The reason for the greater number of pupils shown as having chosen A level subjects than pupils shown as having chosen to do A levels at all is that some pupils ticked more than one box in the question of what they intended to do next, in which case only their first response is used to calculate the number of pupils taking a particular option, but all responses on subject choice are used.) Table 6.17presents an unsurprising pattern. Boys are considerably more likely than girls to opt for at least one science subject. Girls are somewhat more likely than boys to opt for at least one humanity, and considerably more likely to opt for a social science or a language. Male and female students are about equally likely to choose at least one arts subject.

Arts Science Humanities Languages Social science

Service f 70 60 60 17 21

class % 58.3 50.0 50.0 14.2 25.8

Non Service class f % 40 46.5 48 55.8 26 30.2 6 7.0 12 14

Table 6.18: A level subject choice by class

Table 6.18 presents the same information broken down by social class. Non service class pupils are slightly more likely to take at least one science, but less likely to take at least one of any of the other subject categories. In the case of GNVQs, it is more difficult to categorise the subjects according to type. Table 6.19 shows the five most popular GNVQ choices broken down by gender. Since I am dealing with individual subjects rather than groups of subjects here, the numbers are quite small, and one must be especially cautious in drawing conclusions from them. The clearest gender difference is girls’ greater likelihood of choosing health and social care. Boys seem to be more likely to choose art and design and information technology, whereas for the girls in this sample, business studies and leisure and tourism are more popular options than they are for the boys. There is no point in breaking this data

CHAPTER 6. PUPILS’ CHOICES

Art and Design Business Health Social IT Leisure Tourism

Male f % 13 22.4 9 15.5 2 3.4 11 19.0 5 8.6

163 Female f % 9 12.2 16 21.6 28 37.8 7 9.5 8 10.8

Table 6.19: GNVQ subject choice by gender

down by social class as the numbers of service class pupils choosing each GNVQ subject are very small indeed.

6.6.5

Gender and Subject Choice

Girls are less likely than boys to choose science subjects at A level and at degree level. This is the subject of concern because subject choice has been shown to have consequences in the labour market. For instance, Cheung (Cheung, 1997) finds that a substantial proportion of the labour market disadvantage suffered by female graduate respondents to the NCDS survey was accounted for by the concentration of women graduates in the humanities and education. I will model the subject choices of the students who opted to do A levels. I will examine the decision to take at least one “hard science” (maths, physics or chemistry), since biology is often seen as the “feminine” science, and indeed, male and female respondents to my survey had extremely similar likelihoods of choosing biology. I will then go on to model the number of hard sciences taken as a proportion of the total number of A levels. The results are shown in table 6.20. Model 1 shows that boys are significantly more likely than girls to have opted for “hard science” A levels, controlling for both overall GCSE results and maths results. Maths GCSE results are significantly positively associated with the decision to do hard science A levels. For the pupils in my survey, there is a slight difference in favour of boys in maths GCSE results (although this is no longer true nationally). The average score for boys is 3.6, and for girls 3.4. (Recall that 3 points are allocated for an ‘E’, 4 points represent a ‘D’, so this difference represents only a fraction of a grade). Overall GCSE

CHAPTER 6. PUPILS’ CHOICES

Constant GCSEs Maths GCSE result School School 1 School 2 School 3 Male Self-assessed ability Occupational aspirations Owner or manager Professional Associate professional Arts and media χ2 N

164

Model 1 B(s.e.) −1.85(0.65)** −0.01(0.02) 0.34(0.17)*

Model 2 B(s.e.) 0.42(1.03) −0.01(0.02) 0.30(0.17)

Model 3 B(s.e.) −0.41(1.17) 0.01(0.02) 0.12(0.18)

−0.96(0.43)* −0.02(0.44) −0.69(0.80) 1.03(0.50)*

−1.17(0.44)** −0.12(0.45) −0.81(0.79) 0.90(0.51) 0.69(0.26)**

−0.91(0.51) 0.00(0.54) −0.52(0.90) 0.99(0.62) 0.63(0.29)* *** 0.25(0.98) 1.56(0.41)*** 0.70(0.63)

19.6 222

26.4 220

−1.32(0.51)** 66.0 209

Table 6.20: Pupils’ choice of any hard science A levels grades actually have a slightly negative, though insignificant, association with the decision to take science A levels, once Maths grades have been controlled for. This raises the interesting possibility that girls’ overall superiority at GCSE could actually make them less likely to take science A levels, unless their grades in maths and science keep up. For example, if a girl gets a B grade in maths, but the rest of her grades are As, she is probably less likely to take maths A level than a boy who gets a B in maths, and the rest of his grades are Cs. Model 2 shows the effect of self-assessed ability. A high subjective estimation of a pupil’s own ability is associated with a greater likelihood of choosing

CHAPTER 6. PUPILS’ CHOICES

165

maths, physics or chemistry at A level. This mediates the sex effect, making it insignificant. This suggests that self-confidence is an important factor in choosing these subjects, since they are perceived as particularly difficult. (Note that actual maths performance is rendered insignificant in this model, suggesting that self-evaluated ability is more important than actual ability.) Model 3 shows the effect of occupational aspirations. Aspiring to professional positions is significantly positively associated with opting for hard sciences, whereas aspiring to arts and media positions is negatively associated with opting for hard sciences. We have seen that boys are more likely to aspire to professional positions than are girls. In fact, the variable of “occupational aspirations” is sufficient to mediate the sex effect, removing self-assessed ability. Once again, self-confidence is the key to this effect. There is a strongly significant association between self-assessed ability and professional aspirations, controlling for GCSE results. (B = 0.63, S.E. = 0.24, p > .008.) Table 6.21 shows models for the proportion of A levels taken that are in hard science subjects. The statistic given is η 2 . Model 1 shows that gender is significant, and accounts for a similar proportion of the variation explained as maths GCSE results do. As in the previous regression, GCSE results have a negative effect, though not significantly so, once maths GCSE grades have been controlled for. Model 2 shows that self-assessed ability has a significant effect. This has a very slight mediating effect on the gender effect. Model 3 shows the effect of occupational aspirations. Aspirations for a professional position are significantly positively associated with the proportion of A level choices that are hard sciences, whereas aspirations towards arts and media positions are negatively associated with the proportion of hard sciences chosen. The effects of both maths GCSE grades and self-assessed ability become insignificant in this model. In conclusion, the fact that girls are less likely than boys to opt for hard science A levels seems to have more to do with girls’ self-image than with their actual abilities. It is perhaps surprising that girls continue to underestimate themselves in comparison to boys given girls’ superior academic performance at GCSE. Given the current data, it is not possible to determine why girls should have lower self-perceptions than boys. One obvious possibility is that parents and/or teachers may not evaluate girls as positively as boys of similar ability.

Intercept Sex Male Female School School 1 School 2 School 3 School 4 GCSEs Maths GCSE result Self-assessed ability Occupational aspirations Owner or manager Professional Associate professional Arts and media Other N η 2 for the model 219 0.118

Table 6.21: Proportion of hard sciences

221 0.097

−0.00(0.00) 0.05(0.02)* 0.024

0.005 0.029

−0.13(0.05)* −0.00(0.05) −0.12(0.10)

0.040 0.034 0.000 0.009

0.07(0.03)*

0.031 0.026 0.000 0.007

η 0.024 0.023 0.023

0.008 0.023

* −0.15(0.05)** −0.01(0.06) −0.13(0.10)

η 0.002 0.028 0.028

2

−0.00(0.00) 0.04(0.02)*

Model 2 B(s.e) 0.29(0.13)* * 0.14(0.06)*

2

Model 1 B(s.e) 0.05(0.08) * 0.16(0.06)*

0.187 0.002 0.088 0.006 0.026

*** −0.08(0.12) 0.19(0.04)*** 0.08(0.07) −0.11(0.05)* 208 0.279

0.018

0.001 0.008

0.025 0.020 0.000 0.005

η2 0.013 0.021 0.021

0.06(0.03)

−0.00(0.00) 0.02(0.02)

−0.10(0.05)* 0.00(0.05) −0.10(0.09)

Model 3 B(s.e) 0.20(0.12) * 0.12(0.06)*

Constant GCSEs Male School = 1 School = 2 School = 3 School = 4 Self-assessed ability Occupational aspirations Owner or manager Professional Associate professional Arts and media Other Carer

Model 2 B(s.e) 2.22(1.58) −0.04(0.02)∗ −3.26(0.85)∗∗∗ −6.64(27.21) −1.41(0.61)∗ −1.12(0.64) 0 −0.22(0.43)

−9.50(37.52) 0

−1.01(1.17) −7.05(67.59) 11.45(65.29)

Model 3 B(s.e) 2.59(0.97) ∗ ∗ −0.06(0.02)∗ −3.54(0.98)∗∗∗ −7.57(42.38) −2.01(0.73) ∗ ∗ −1.83(0.80)∗ 0

3.43(0.87) ∗ ∗∗

Model 4 B(s.e) 3.59(1.10) ∗ ∗ −0.02(0.02) −2.33(0.86) ∗ ∗ −6.62(27.58) −1.58(0.73)∗ −1.28(0.77) 0

Table 6.22: GNVQ subject choices (continued overleaf. . . )

Model 1 B(s.e) 1.52(0.76)∗ −0.04(0.02)∗ −3.20(0.84)∗∗∗ −6.57(27.27) −1.35(0.59)∗ −1.09(0.64) 0

3.55(1.34) ∗ ∗∗

Model 5 B(s.e) 3.63(1.28) ∗ ∗ −0.01(0.03) −2.66(0.94) ∗ ∗ −7.67(45.59) −1.91(0.81)∗ −1.62(0.79) 0

Occupational values Control Intellectual Material Hands Easy People χ2 N 36.474 ∗ ∗∗ 130

Model 2 B(s.e)

60.773 ∗ ∗∗ 129

Model 3 B(s.e)

61.873 ∗ ∗∗ 129

Model 4 B(s.e)

Table 6.22: GNVQ subject choice (. . . continued from overleaf)

36.214 ∗ ∗∗ 130

Model 1 B(s.e)

0.50(1.34) −1.33(1.31) 0.54(0.76) 2.20(1.34) −8.36(40.89) 0 70.548 ∗ ∗∗ 126

Model 5 B(s.e)

CHAPTER 6. PUPILS’ CHOICES

6.6.6

169

GNVQs

The most striking gender difference in GNVQ choices is the tendency of girls to choose “health and social care”. Table 6.22 shows a logistic regression examining this gender difference. Model 1 shows a highly significant gender effect. GCSE performance is also significantly negatively associated with choosing health and social care as a GNVQ subject. However, model 2 shows that self assessed ability is not significant once GCSE performance has been controlled for. Model 3 shows no significant effect of occupational aspirations. However, in model 4, I coded this variable differently, in order to look specifically at whether the aspiration to work with children or as a carer affected the choice of health and social care GNVQ. Unsurprisingly, this is highly significant. The gender effect is partially mediated in this model, but still remains highly significant. Occupational values are included in the final model, as one might suspect that the fact that girls are less likely to choose material factors as most important and more likely to choose being with or helping people as most important might help to explain the gender differential in choosing this GNVQ. However, occupational values in fact have no significant effect on this decision. So, although A GNVQ in health and social care is perhaps unlikely to lead to a well paid job, girls do not choose this option because of a relative lack of concern for material circumstances. Perhaps girls simply see entering the “caring” labour market as a good way to avoid unemployment. Boys may be unwilling to enter this labour market due to their own gender stereotypes, but, alternatively, they may suspect that this labour market is relatively closed to males. In the case of working with children, they might well be right in this view, as men who wish to work with young children are often viewed with suspicion.

6.7

Conclusions

Social class appears to have little effect on what factors are deemed important in choosing a job. There are some gender differences, with girls valuing intellectual stimulation and being with and helping others more than boys do. However, these values were not significant in determining the level of pupils’ occupational aspirations. Girls were more likely to aspire to associate

CHAPTER 6. PUPILS’ CHOICES

170

professional positions than were boys, but less likely to aspire to professional occupations. There was no social class effect on aspirations once GCSE results were controlled for. This suggests that the lower aspirations of working class pupils simply reflect an adjustment to their perceived chances of success. The vast majority of the pupils in my sample intended to stay on in school, and there was no social class effect on this decision. Boys were less likely to choose to stay on than girls were, and there was a significant effect of family structure, with pupils from step and single parent families considerably less likely to stay on. There was a social class effect on the decision to do A levels, controlling for GCSE results. However, this effect was mediated by parents’ educational qualifications. An explanation of the social class effect in terms of “relative risk aversion” cannot be supported, as there is no direct social class effect. Rather, it seems that pupils whose parents have degrees wish to pursue A levels, which represent the traditional route to university. It could be that these pupils do not wish to be educationally downwardly mobile. Alternatively, they may have a greater knowledge of the value of higher qualifications on the labour market. Girls are less likely than boys to choose hard sciences, and this seems to be partly due to their lower estimation of their own abilities compared to boys, and partly due to the lower proportion of girls who aspired to professional positions. Girls choice of health and social care as a GNVQ subject is partly explained by their greater tendency to aspire to “caring” occupations, such as working with children. However, this tendency was not explained by the value girls put on being with and helping others. It could be that boys and girls are responding to the perceived openness of this labour market to each sex.

Chapter 7 Conclusions The aim of this thesis has been to explain social class and gender differentials in educational outcomes, and, in particular, to evaluate the usefulness of “cultural reproduction theory” and “rational choice theory” in examining these questions. Very often in contemporary sociology, theory and empirical work are separated, to the detriment of both. A basic aim of this thesis has been to produce a genuinely theory-driven piece of empirical research, avoiding the dangers of vacuous and irrefutable theory on the one hand and tedious number crunching on the other I have criticised Bourdieu for the vagueness of his theory. Such vagueness is a shield against refutation, and the “heavily articulated sentences” (Bourdieu, 1990) in which the theory is expressed provide further protection. However, the impossibility of giving an “authentic” operationalisation of Bourdieu’s theory should not serve as an excuse for failing to provide anything better than crude proxy measures. This thesis has examined the concept of cultural capital empirically, and determined which elements of the dominant culture are actually associated with educational success (and can therefore be termed capital, in that they provide a return). Thus, I have shown how the originally rather nebulous concept of cultural capital can be transformed into a concrete and meaningful concept with genuine applications to empirical research. I have critically analysed the rational choice approach, and have argued that the principle of “revealed preference” relied upon (explicitly or implicitly) by rational choice theorists makes their theories irrefutable. Rejecting the doctrine of “revealed preference” I have examined the beliefs and attitudes of students directly. This has allowed me to evaluate competing hypothe-

171

CHAPTER 7. CONCLUSIONS

172

ses as to the reasons for social class and gender differences in educational participation post-16 empirically.

7.1

Cultural Reproduction

According to Bourdieu’s theory of cultural reproduction, participation in the dominant culture is transmitted within the middle and upper class home, and enables the children of these classes to succeed in gaining high levels of educational credentials. It is unclear from Bourdieu’s formulation of his theory whether the educational advantage bestowed by participation in the dominant culture is due to the communication of status or to the acquisition of useful skills, such as reading ability, or to some combination of the two. I have used as broad as possible an operationalisation of cultural participation, including reading habits, TV viewing, music listened to, playing an instrument, and attendance at art galleries and museums. I have also tested pupils’ vocabularies (ability to use “educated” language being a crucial part of Bourdieu’s concept of cultural capital) and pupils’ cultural knowledge. I have found that, in line with Bourdieu’s theory, there is a strong association between parents’ and pupils’ cultural participation, suggesting that cultural capital is transmitted within the home. Cultural participation is associated with GCSE attainment, but only partially explains the social class difference in attainment. Service class pupils retain a large advantage over their non-service class peers once cultural participation and even vocabulary and cultural knowledge have been controlled for. Analysis using the SOC class schema shows that vocabulary and cultural knowledge partially explain the advantage of pupils from professional backgrounds, but not the advantage enjoyed by pupils from employer or managerial backgrounds. Not all elements of participation in the “dominant” culture are associated with GCSE success (controlling for background variables). Participation in “formal” culture (going to art galleries, museums, concerts) is not associated with GCSE attainment, and neither is playing a musical instrument or listening to “high-brow” music (classical or jazz). Therefore, there are no grounds for calling these activities cultural capital. Reading behaviour and watching relatively “high-brow” programmes on TV, on the other hand, are associated with GCSE success. This would seem to back the view, put forward by Crook

CHAPTER 7. CONCLUSIONS

173

(Crook, 1997) and de Graaf et. al. (De Graaf et al., 2000), that cultural participation is associated with educational performance because certain forms of cultural participation (such as reading) help to develop cognitive skills which help at school. This is in contrast to the view that cultural reproduction works through the communication of status (to teachers for instance). Further evidence to back the former rather than the latter view is given by the fact that the effect of pupils’ cultural participation on GCSE attainment is entirely mediated by their cultural knowledge and vocabulary test scores. Both parents’ and pupils’ cultural participation had a significant effect on pupils’ attitudes to the intrinsic value of education. This suggests that families rich in “cultural capital” promote the value of education in the development of an individual rather than just its value in the labour market. Cultural participation is also strongly associated with self-assessed ability and self-predicted grades, controlling for actual GCSE grades. This suggests that participation in “cultured” activities gives pupils a high estimation of their own abilities. Cultural participation had a small effect on the decision to do A levels (controlling for GCSE results, social class, etc.), and this was mediated by the effects of pupils’ attitudes to education and beliefs about their own abilities. So, certain forms of cultural participation (reading and watching relatively “high-brow” TV) are associated with cultural knowledge and linguistic ability. These abilities in turn are associated with educational attainment at GCSE. In addition, cultural participation is associated with positive attitudes towards education and with self-confidence (albeit over-inflated selfconfidence) in one’s own abilities, and both of these characteristics are associated with the decision to pursue the ambitious educational option of A levels.

7.2

Rational Choice

I have discussed various versions of rational choice theory, and the pitfalls of both “strong” and “weak” assumptions regarding human motivations in general, and motivations for educational choices in particular. I have argued that there is no reason for a theory that assumes the link between beliefs and desires on the one hand and actions on the other to be rational to assume that people’s beliefs and desires are also “rational”. (The meaning of the

CHAPTER 7. CONCLUSIONS

174

term rationality is at any rate quite different when applied to each of these areas.) An agent’s motivations cannot be inferred from their actions, as any given action may be consistent with an infinite number of combinations of desires and beliefs about the situation. Therefore, it is crucial to gather direct evidence on people’s beliefs and desires, attitudes and aspirations. It is often assumed that attitudes to education vary according to social class. However, I have found that the variation in pupils’ stated attitudes towards education varied only slightly by social class, and this variation was entirely mediated by success at school, as measured by GCSE results. Neither was there any association between parents’ qualifications and pupils’ attitudes to education. These results are surprising given the widespread assumption that working class people have a more negative attitude to education than middle class people do. As I stated in chapter 5, it may be that this view was never justified, or it may be that a change has taken place, as qualifications have become essential in the job-market, regardless of social class. Pupils whose parents are in the higher occupational categories and have higher levels of education overestimate themselves compared to others. Generally, those in dominant positions, whether in terms of gender, class, education, or culture, perhaps justify those positions by developing an exaggerated belief in their own abilities. Breen and Goldthorpe (Breen and Goldthorpe, 1997) suggest that working class pupils’ attach a lower value to middle class occupations than do middle class pupils due to “relative risk aversion”. On this view, the prime aim of individuals from all classes is to avoid downward social mobility rather than to achieve upward social mobility. As I pointed out in chapter 2, the term “relative risk aversion” is misleading. The term “loss aversion” would more accurately describe this process which is driven by fear of losing one’s current class position. I assessed the possibility that pupils’ occupations are driven by loss aversion by surveying pupils on their occupational expectations and aspirations, and also on the factors they considered important in choosing a job, since any difference in aspirations could be value driven rather than loss aversion driven. I found that occupational values did not vary by social class, and, in addition, no significant class difference in aspirations remained once GCSE results had been controlled for, suggesting that pupils’ simply adjust their expectations in line with their likely opportunities. The great majority of pupils in my sample intended to say on in education

CHAPTER 7. CONCLUSIONS

175

post-16, and there was no social class effect on this decision once GCSE performance had been controlled for. Pupils from single or step-parent families were less likely to choose to stay on than pupils in “intact” families. This effect is striking given the absence of a social class effect on staying on in further education. The social class effect on the decision to do A levels was mediated by the effect of parents’ qualifications. This finding is crucial, as the lack of a direct social class effect on this decision is inconsistent with Breen and Goldthorpe’s explanation in terms of “relative risk aversion”. Perhaps students’ decisions are driven more by a fear of educational downward mobility than of occupational downward mobility. Students’ self assessed abilities, attitudes to education, and occupational aspirations have some effect on the intention to do A levels, but these variables do not mediate the effect of parental education. Overall, it seems that perhaps the stress that rational choice theorists put on the “secondary effects” of stratification is misplaced, at least in the British context. The social class effect on attainment at the compulsory level (GCSE) is strong, and is only partially explained by other variables. It is striking that this is not the case for educational participation beyond the compulsory level. The crucial determinant of participation in further education is performance at GCSE. The reason for Breen and Goldthorpe’s assumption that there are important social class effects on choices at 16+, controlling for ability, may be that this was the case for the NCDS children (Micklewright, 1989). (Note that Micklewright controls for tested ability in maths and comprehension, but not for examination performance. For evidence that social class affected the likelihood of getting a degree or higher vocational qualification, controlling for both tested ability and examination performance at 16, see Sullivan and Cheung (Sullivan and Cheung, 2000)). However, the British educational system has altered dramatically since these children, born in 1958, were at school. The NCDS cohort grew up in an era of major educational reforms. They entered secondary school aged 11 in 1969. The majority of the NCDS cohort members went to comprehensive schools which were reorganised from the three different types of schools under the so-called “tripartite system”. During the years of the comprehensive reorganisation, some NCDS cohort members still went to these grammar, technical and secondary modern schools before particular schools were re-organised. Entry to these schools was primarily based on the results of the “11 plus” examination, which pupils sat at age 11. Pupils who passed the 11 plus examination generally went to gram-

CHAPTER 7. CONCLUSIONS

176

mar schools, whereas pupils who failed the examination went to either the more vocationally oriented technical schools or to secondary modern schools. (Private schools were largely untouched by government reforms). During this period, there was a two-tier system of examinations. At age 16, most grammar school pupils took the General Certificate of Education at Ordinary level (GCE O level), whereas the examination aimed at secondary modern pupils was the Certificate of Secondary Education (CSE) which was set at a lower level. Comprehensive reorganisation and the raising of the school leaving age to 16 in 1974 gave a boost to the numbers taking the CSE exams, and the majority of pupils subsequently left school with some kind of certificate. Subsequently, comprehensivisation has continued, and only a small number of grammar schools remain. The introduction of the GCSE (in 1988) means that all pupils now follow a similar curriculum until age 16. It seems likely that, by pushing the first key educational transition point back to age 16, this reform has reduced the importance of the secondary effects of stratification. Despite this, Breen and Goldthorpe have simply assumed that the secondary effects of stratification not only continue to exist, but are crucial in explaining social class differences in outcomes. In fact, I would suggest that, in the British context at least, it is the primary effects of stratification, i.e. social class effects on attainment at GCSE that are crucial in explaining social class differences in educational outcomes. The reason for the neglect of these primary effects by rational choice theorists is that they do not fit readily into a rational choice framework.

7.3

Gender

Whereas social class differentials in educational attainment and participation have remained fairly stable, gender differences have diminished rapidly in most advanced industrialised societies, and in some instances, such as that of attainment at GCSE, these differences have been reversed. Despite this, strong differences in subject choice remain. The girls in my sample were less likely to choose hard sciences at A level than were boys. This effect seemed to be partly explained by girls’ lower estimation of their own abilities, and partly by their lower tendency to aspire to professional occupations as compared to boys. Girls’ tendency to reject science subjects is important in the light of evidence that women in Britain still earn considerably less than similarly qualified men, even when

CHAPTER 7. CONCLUSIONS

177

they are childless, and evidence that subject choice was an important factor in explaining sex differences in occupational outcomes for the NCDS cohort (Cheung, 1997). Boys were less likely to choose to stay on in further education than girls, but there was no significant gender difference in the tendency to choose A levels. Despite girls’ relatively high levels of attainment at GCSE, girls still seem to lack confidence in their abilities as compared to boys. Boys significantly overestimated their academic abilities as compared to girls. It could be that boys are still given more positive messages about their abilities (perhaps by parents or teachers) than girls are given, even though the boys’ actual levels of academic performance are lower. Girls were slightly (but significantly) more positive than boys in their evaluation of education as an intrinsic good. There was no significant difference between boys and girls in their evaluation of education as a labour market good. There were significant gender differences in occupational values, with girls being more likely to choose intellectual stimulation and being with and helping people, as opposed to material factors, as being most important in choosing a job. It could be that boys still see themselves as future breadwinners for a family, and therefore feel the need to prioritise income more than girls do. Girls were more likely to aspire to associate professional jobs and less likely to aspire to professional jobs than were boys. This is perhaps unsurprising, as associate professional jobs, such as teaching, nursing, and social work, are female dominated. It could be that girls still feel the need to choose jobs that are easily compatible with being the primary carer for children. It is often claimed that girls’ cultural participation (notably reading) is higher than that of boys. Therefore, it seemed plausible that the “cultural reproduction” approach might shed light on girls’ superiority at GCSE. In fact, for this sample, gender differences in cultural participation proved to be slight, and certainly could not provide an explanation for girls’ superior GCSE performance as compared to boys. The gender difference in GCSE results is quite substantial for the pupils in my sample (as it is nationally). The average GCSE points score (recall that 1 point is awarded for a G, 2 for and F, etc.) for a boy is 35 (equivalent to 5 Ds and 3 Cs) as compared to 39 for a girl (equivalent to 7 Cs and a D). This is, of course, not as dramatic as the social class difference in GCSE results

CHAPTER 7. CONCLUSIONS

178

- the average service class pupil in my sample gained 51 points (equivalent to 5 Bs and 3 As) whereas the average non service class pupil gained 34 points (6 Ds and 2 Cs). (The mean GCSE score for those whose parental occupation was recorded as “missing” was even lower at 28). However, the gender gap is substantial, and it is unsurprising that it should be the focus of much interest. On the other hand, one should not be tempted into drawing a parallel between the gender differential in GCSE performance and the social class differential. When we concern ourselves with differentials in educational achievement, rather than absolute levels of achievement, this is because education is a positional good - i.e. a good whose value depends on how much we have compared to others. Since men are still strongly advantaged in the labour market, their relatively poor performance at GCSE should not be an object of great concern. However, that does not mean that gender differences at GCSE should not be an object of great interest. Girls’ superior performance is interesting precisely because it breaks the rule that socially disadvantaged groups tend to under-perform educationally too.

7.4

Family Size and Structure

Family size (i.e. the number of siblings a pupil has) and family structure (intact two parent family, step-parent or single parent family) were primarily included in the analyses simply as “control variables”. However, both of these variables had significant associations with GCSE attainment. The negative effect of having a higher number of siblings was mediated by the pupils’ cultural knowledge and language scores. This is consistent with previous research suggesting that relatively thinly-spread parental attention in larger families affects children’s intellectual development (Nisbet, 1953). However, the negative effect on GCSE attainment of being in a step- or single-parent family is not mediated in this way. Furthermore, controlling for GCSE performance, pupils’ from step- and single parent families were less likely to plan to stay on in post-compulsory education or training after their GCSEs. The effect of being in a step-family actually seemed to be more negative than the effect of being in a single-parent family, which is surprising in that one would expect step-parent families to be considerably better off financially than single-parent families. It seems likely, however, that the level of economic resources that are available to a child are affected by factors other than total household income. For instance, a step-parent may be less willing to support a child beyond compulsory education than a “real” parent would be.

CHAPTER 7. CONCLUSIONS

7.5

179

Suggestions for future research

I have argued for the superiority of the approach of collecting data which is specifically designed to address the research questions at hand over the approach of using whatever data is to hand to address one’s research questions. However, I cannot deny the enormous disadvantage of collecting one’s own data with the limited resources available to a student. My data set is modest in size, and unrepresentative. The fact that my data-set does not contain the private sector is a particular problem for the assessment of cultural reproduction theory. The mechanisms of social reproduction may vary between the state and private sectors. Private schools may be more likely to encourage cultural activities and to promote knowledge of the dominant culture. Private schools may function as an alternative route to educational credentials for families who have economic capital but lack cultural capital. A comparison between the state and private sectors would also be relevant to rational choice theory. Staying on rates are thought to be higher in the private sector than in the state sector once attainment at the compulsory level has been controlled for. If this is true once family characteristics such as social class have been controlled for, is this due to the efforts of such schools to promote staying-on as worthwhile, or to a peer-group effect, or perhaps to private school pupils having relatively high estimations of their own abilities? Claims about educational attainment are often made on the basis of outdated evidence, such as NCDS. The claim that social class affects rates of educational participation in Britain, controlling for educational attainment at the compulsory level, is often made in the present tense, although it relates to information on children born in the 1950’s. There is clearly a need for an up to date representative data-set dealing with educational attainment and participation. Enormous changes in the educational system, and in the wider society, have taken place since the NCDS children left school in the 1970’s, and we cannot simply assume that these have made no difference to pupils’ experiences. Claims regarding the effect of family structure on children’s educational experiences are also made on the basis of NCDS data, despite the huge changes that have taken place in family life. My data suggests that family structure is still very important, but much more detailed information on family structure would be needed in order to analyse the reasons (emotional, social, economic) behind this effect. Another limitation

CHAPTER 7. CONCLUSIONS

180

of my data set is the absence of information on ethnicity. My data set is too small too have provided reasonable sized sub-samples of minority ethnic groups. A larger survey would be able to examine whether there are true ethnic differences in educational attainment (controlling for social class) and what the reasons are for the differing educational performance of different minority ethnic groups. A regrettable omission from my research in terms of theory is that I have not drawn on Coleman’s concept of “social capital”. For Coleman, social capital can exist within the home, in the form of the level of support given to pupils by parents. Social capital also exists in the form of social norms and networks, in the community surrounding a school where parents know each other (and possibly also the teachers). “A child’s friends and associates in school are sons and daughters of friends and associates of the child’s parents.” (Coleman, 1990, p. 318). This affects the relationship of parents to the school. Parents are able to get a lot more information about what is going on in the school, and about the behaviour of their own child. Parents who talk to each other can establish strong norms of behaviour for their children. Coleman argues that the community around Catholic schools in the US, which is created by the church, creates and enforces strong norms against dropping out of high school, and that this explains the lower drop out rate in Catholic schools than in other private schools. Coleman also uses the concept of social capital within the school. Coleman, in “The Adolescent Society” (Coleman, 1961) discusses the effects of peer groups on pupils. He asks why academic prowess does not lead to popularity in school, whereas sporting prowess does. For Coleman, the answer is that sporting competitions take place between schools, whereas academic competition takes place between individuals. Therefore, individuals who do well academically are chided for “rate-busting”, whereas pupils who excel in sports gain popularity and status. Whatever the explanation for the social norms prevailing in schools, it seems likely that these norms may have an important effect on educational attainment, and, more broadly, on students’ well-being while in school. Smith and Tomlinson (Smith and Tomlinson, 1989) find that having a high proportion of high achievers in a school improves the achievement of pupils in the school as a whole. (This could of course work through teachers’ expectations as well as peer group norms). Power et. al. (Power et al., 1998) present an interesting finding that, whereas academically able pupils at a grammar school were likely to worry about not being able to keep up with the work, academically

CHAPTER 7. CONCLUSIONS

181

able pupils at a comprehensive were much more likely to worry about other pupils thinking they were too clever. The sample size of Whitty et al’s study are far too small to generalise from. However, I think that the norms imposed by peers may well have as much impact on attainment as the rules imposed by the school, yet studies focusing on attainment have not generally looked at these norms. In addition, the desire for a peer group that will have a positive effect on the child seems to be an important factor in school choice, (Gewirtz et al., 1995; Walford, 1994) and may well be an explanatory factor in any “private school effect”. For all these reasons, I think that if we are to gain an understanding of the processes underlying inequalities in educational attainment and participation within the contemporary British education system, we need a large and representative data-set, containing both state and private schools.

7.6

General Theoretical Conclusions

Throughout this thesis, I have stressed the need to integrate theory with empirical research. The need to operationalise a theory in order to test it against empirical evidence almost inevitably leads to the abandonment of “grand theory” (in C. Wright Mills’ phrase (Mills, 1959)) in favour of something more precise. The empirical work will also be unlikely to sustain the theory in whole. Some concepts will prove fruitful, others unhelpful or unfounded. The grand theory of cultural reproduction is unsupported by this thesis, as by other work that has sought to unpack the theory in order to give an empirical analysis (DiMaggio, 1982; DiMaggio and Mohr, 1985; DiMaggio and Mohr, 1995; Crook, 1997; De Graaf et al., 2000). Instead, we return to the much earlier idea that middle class families have cultural resources, as well as economic and social resources, which put them at an educational advantage. Furthermore, cultural resources seem to account for only a modest proportion of the educational advantage enjoyed by pupils from the higher social class categories. In addition, participation in the “formal culture” of art galleries, museums, etc. has no association with educational attainment at GCSE, controlling for social class and other background variables. So, rather than the grand theory of cultural reproduction we are left with a much more modest story — reading and TV viewing habits are associated with abilities (linguistic and cultural knowledge) which aid academic performance.

CHAPTER 7. CONCLUSIONS

182

Rational choice theorists have been torn between strong (and obviously false) and weak (and vacuous) versions of rational choice theory. I have pointed out the need for data on agents’ motivations — something that rational choice theorists have generally shied away from. Rational choice theorists seem to hold a rather macho conception of the nature of science, according to which it is characterised by the rejection of “soft” data, rather than by the testing of hypotheses. In the sociology of education, rational choice theorists have focused on the “secondary effects” of stratification since these effects can readily be seen as due to choices. In my view, this has led to an exaggeration of the importance of the secondary effects of stratification, and a neglect of the primary effects. Only Murphy (Murphy, 1990) attempts to characterise social class differences in educational performance prior to any key decision point in terms of choices. As I have argued, this attempt is highly problematic. Not all of social life can be readily or usefully characterised in terms of decision-making. This is a complex area. It is clear, for instance, that desires, attitudes and beliefs are not straightforwardly chosen, and yet it would be wrong to suggest that human will and agency can be can be entirely absent from shaping an individuals’ desires, attitudes and beliefs. However, I would suggest that, to the extent that a sociological question cannot usefully be seen in terms of “decision-making”, rational choice theory will not be a useful tool for tackling the question. Overall, I would argue that sociologists, regardless of theoretical perspective, should strive firstly for clarity. Vague and obscure theories have no use except to inflate the author’s reputation among the gullible. (It is a sad and perplexing fact that the humanities and social sciences are full of people who are willing to be awed by anything incomprehensible). This criticism must be most strongly directed against French and French-inspired sociology (for illustration of this point with authors other than Bourdieu, see Sokal and Bricmont (Sokal and Bricmont, 1998)). However, it should be remembered that formality is not the same thing as clarity, and one can hide behind formulae as well as words. Secondly, it is crucial to abandon the distinction between theoretical and empirical work. Speculation about the social world without any real engagement with it ranks alongside speculation as to the existence of an external world for sheer futility. The existence of people styling themselves sociological theorists, or “social theorists” without ever engaging in empirical work is absurd. Such people should be banished to departments of philosophy, or if

CHAPTER 7. CONCLUSIONS

183

they won’t have them (as seems probable) to departments of cultural studies.

Appendix A Correlations

184

Occupational aspirations

Language

Graduate parent

GCSEs

FE

Cultural knowledge Educational attitudes

A levels

σ p σ p σ p σ p σ p σ p σ p σ p

Alevels

0.48 0.00 0.28 0.00 0.44 0.00 0.60 0.00 0.45 0.00 0.39 0.00 0.46 0.00

Cultural knowledge

0.16 0.00 0.21 0.00 0.63 0.00 0.43 0.00 0.57 0.00 0.39 0.00

0.48 0.00

Educational attitudes 0.23 0.00 0.23 0.00 0.06 0.24 0.10 0.03 0.19 0.00

0.28 0.00 0.16 0.00

FE 0.33 0.00 0.14 0.01 0.16 0.00 0.28 0.00

0.44 0.00 0.21 0.00 0.23 0.00

GCSEs 0.30 0.00 0.58 0.00 0.44 0.00

0.60 0.00 0.63 0.00 0.23 0.00 0.33 0.00

Graduate parent 0.25 0.00 0.29 0.00

0.45 0.00 0.43 0.00 0.06 0.24 0.14 0.01 0.30 0.00

Language 0.31 0.00

0.39 0.00 0.57 0.00 0.10 0.03 0.16 0.00 0.58 0.00 0.25 0.00

Occupational aspirations 0.46 0.00 0.39 0.00 0.19 0.00 0.28 0.00 0.44 0.00 0.29 0.00 0.31 0.00

Parents’ cultural participation 0.43 0.00 0.49 0.00 0.22 0.00 0.22 0.00 0.45 0.00 0.54 0.00 0.33 0.00 0.30 0.00

Pupils’ cultural participation 0.41 0.00 0.44 0.00 0.24 0.00 0.22 0.00 0.41 0.00 0.41 0.00 0.36 0.00 0.37 0.00

Service class parent 0.41 0.00 0.34 0.00 0.13 0.01 0.15 0.00 0.37 0.00 0.51 0.00 0.29 0.00 0.31 0.00

Sex 0.00 0.95 −0.02 0.71 0.11 0.02 0.14 0.00 0.09 0.05 −0.03 0.56 0.04 0.36 0.05 0.32

Single or step family −0.11 0.02 −0.07 0.12 −0.14 0.00 −0.19 0.00 −0.24 0.00 0.02 0.75 −0.07 0.13 −0.02 0.67

Sibling number −0.14 0.00 −0.29 0.00 0.01 0.87 −0.06 0.17 −0.24 0.00 −0.19 0.00 −0.23 0.00 −0.16 0.00

0.27 0.00 0.24 0.00 0.12 0.01 0.13 0.00 0.29 0.00 0.27 0.00 0.14 0.00 0.20 0.00

Social capital

Social capital

Sibling number Single or step family

Sex

Parents’ cult. participation Pupils’ cult. participation Service class parent

σ p σ p σ p σ p σ p σ p σ p

Alevels

0.43 0.00 0.41 0.00 0.41 0.00 0.00 0.95 −0.14 0.00 −0.11 0.02 0.27 0.00

Cultural knowledge

0.49 0.00 0.44 0.00 0.34 0.00 −0.02 0.71 −0.29 0.00 −0.07 0.12 0.24 0.00

Educational attitudes

0.22 0.00 0.24 0.00 0.13 0.01 0.11 0.02 0.01 0.87 −0.14 0.00 0.12 0.01

FE 0.22 0.00 0.22 0.00 0.15 0.00 0.14 0.00 −0.06 0.17 −0.19 0.00 0.13 0.00

GCSEs 0.45 0.00 0.41 0.00 0.37 0.00 0.09 0.05 −0.24 0.00 −0.24 0.00 0.29 0.00

Graduate parent 0.54 0.00 0.41 0.00 0.51 0.00 −0.03 0.56 −0.19 0.00 0.02 0.75 0.27 0.00

Language 0.33 0.00 0.36 0.00 0.29 0.00 0.04 0.36 −0.23 0.00 −0.07 0.13 0.14 0.00

Occupational aspirations 0.30 0.00 0.37 0.00 0.31 0.00 0.05 0.32 −0.16 0.00 −0.02 0.67 0.20 0.00

Parents’ cultural participation 0.62 0.00 0.44 0.00 0.09 0.06 −0.20 0.00 −0.07 0.16 0.45 0.00

Pupils’ cultural participation 0.34 0.00 0.09 0.07 −0.08 0.09 −0.05 0.34 0.32 0.00

0.62 0.00

Service class parent 0.03 0.53 −0.18 0.00 0.00 0.97 0.20 0.00

0.44 0.00 0.34 0.00

Sex 0.04 0.42 −0.07 0.13 0.03 0.57

0.09 0.06 0.09 0.07 0.03 0.53

Sibling number 0.06 0.18 −0.11 0.02

−0.20 0.00 −0.08 0.09 −0.18 0.00 0.04 0.42

Single or step family −0.03 0.53

−0.07 0.16 −0.05 0.34 0.00 0.97 −0.07 0.13 0.06 0.18

0.45 0.00 0.32 0.00 0.20 0.00 0.03 0.57 −0.11 0.02 −0.03 0.53

Social capital

Appendix B Questionnaire

187

Appendix C Educational Attitudes Scale Table C.1 shows the questions on attitudes to education to which pupils responded either agree strongly, agree, neither agree nor disagree, disagree or disagree strongly. Those items that are marked with an asterisk are phrased negatively, and the scales for these items have been reversed.

C.1

Reliability

The reliability for the whole 16-item scale is standardised item α = 0.80. The reliability for the 7-item sub-scale for attitudes to education as an intrinsic good is standardised α = 0.73. The reliability for the 5-item sub-scale for attitudes to education as an instrumental good is less high, at standardised item α = 0.43. Nevertheless, I think that use of this sub-scale is justifies in that it meets my theoretical purpose. Also, one must bear in mind that reliability analysis is based on an assumption that the responses for each item are normally distributed. This assumption is strongly violated in the case of this educational attitude scale. Factor analysis shows (see table C.2) one factor accounting for 27% of the variance, with further factors being much less important. (The second factor accounts for 9% of the variance, the third for 8%). I have limited the number of factors extracted to three. The first component can be seen as reflecting students’ general attitude towards education. The highest loading items are ‘I don’t like studying’ and ‘I enjoy studying’. The highest loading items in the second component are ‘The more qualifications you get, the better the job 204

APPENDIX C. EDUCATIONAL ATTITUDES SCALE

STUDY3 STUDY4 STUDY8 STUDY9 STUDY14 STUDY15 STUDY16

STUDY2 STUDY5 STUDY6 STUDY7 STUDY1

STUDY1 STUDY10 STUDY11 STUDY12

205

Intrinsic Studying is worthwhile for its own sake I enjoy studying Studying increases your confidence Studying increases your ability to think clearly *I don’t like studying Studying improves your ability to be creative *Studying is only ever worthwhile if it leads to a job Instrumental The more qualifications you get, the better the job you are likely to get. *Qualifications are useless for getting jobs Studying can help you to gain skills which will be useful at work *These days, it doesn’t matter how many qualifications you get, you still won’t get a job These days, if you are unqualified, you can’t get a job Neither *School is a waste of time *I muck about in lessons *Studying is irrelevant to real life I take school seriously Table C.1: Attitudes to education scale

you are likely to get’, ‘Qualifications are useless for getting jobs’, and ‘These days, it doesn’t matter how many qualifications you get, you still won’t get a job’. This may suggest that component 2 reflects the ‘instrumental’ attitudes to some degree. However, the overall picture for components 2 and 3 is rather messy, and may simply reflect differences in the distributions of responses for each item.

APPENDIX C. EDUCATIONAL ATTITUDES SCALE

STUDY1 STUDY2 STUDY3 STUDY4 STUDY5 STUDY6 STUDY7 STUDY8 STUDY9 STUDY10 STUDY11 STUDY12 STUDY13 STUDY14 STUDY15 STUDY16

1 0.592 0.420 0.403 0.711 0.300 0.499 0.305 0.631 0.637 0.557 0.377 0.698 0.175 0.725 0.515 0.336

Component 2 −0.009 0.474 0.285 −0.428 0.413 0.037 0.394 0.153 0.171 −0.456 0.258 −0.236 0.169 −0.373 0.271 −0.145

206

3 0.305 −0.252 0.103 −0.166 0.120 0.043 0.478 −0.251 −0.153 0.024 0.220 0.122 −0.652 −0.134 −0.151 0.479

Table C.2: Factor analysis of attitudes to education

Appendix D Coding of occupational expectations and aspirations 1. Business Owner

15. Architect

2. Manager

16. Associate Professional

3. Professional Other 4. Engineer

17. Teacher (Secondary or freelance)

5. Doctor

18. Teacher (Primary)

6. Dentist

19. Nurse

7. Vet

20. Technician

8. Barrister

21. Skilled manual

9. Solicitor

22. Mechanic

10. Lawyer

23. Electrician

11. Computer Programmer/Analyst

24. Engineer (non-professional)

12. Academic/Researcher

25. Other

13. Scientist

26. Unskilled manual

14. Accountant

27. Factory Work 207

APPENDIX D. CODING

208

28. Cleaner

53. Something in computing

29. Other

54. Computer operator

30. Armed Forces

55. Other

31. Army

56. Working with animals

32. RAF pilot

57. Veterinary Nurse

33. Other

58. Working with children

34. Routine non-manual other 35. Clerical/Office work 36. Secretary 37. Protective Services 38. Police officer

59. Nanny 60. Nursery nurse 61. Childminder 62. Other 63. Carer

39. Fire fighter 64. Old people’s home 40. Other 65. Something in law 41. Travel and Tourism 42. Holiday Rep 43. Air Hostess 44. Other 45. Hair and Beauty 46. Own salon

66. Something in the arts 67. Designer 68. Artist 69. Writer 70. Journalist

47. Waiter

71. Actor

48. Shop work

72. Musician

49. Stacking shelves

73. Photographer

50. Chef/Catering

74. Other arts or media job

51. Own restaurant

75. Something in sport

52. Profession/Head/Top Chef

76. Sports person

APPENDIX D. CODING

209

77. coach/trainer/instructor

84. Model

78. Work in leisure centre or gym

85. Student

79. PE teacher

86. Unemployed

80. Other

87. Stay at home

81. Fantasy other

88. Too vague to categorise

82. Film star

89. Other

83. Pop star

90. Don’t know

Bibliography Arnot, M. (1994). In Stone, L., editor, The Education Feminism Reader. Routledge and Kegan Paul, New York, London. Becker, G. S. (1964). Human Capital. University of Chicago Press, Chicago. Bernstein, B. (1973). Class, Codes and Control, volume 1. Paladin, London. Boudon, R. (1974). Education Opportunity and Social Inequality. Wiley, New York. Boudon, R. (1994). The Art of Self Persuasion. Polity Press, Cambridge. Bourdieu, P. (1967). Systems of education and systems of thought. International Social Science Journal, 19(3). Bourdieu, P. (1974). The school as a conservative force: Scholastic and cultural inequalities. In Eggleston, J., editor, Contemporary Research in the Sociology of Education. Methuen, London. Bourdieu, P. (1977a). Cultural reproduction and social reproduction. In Karabel, J. and Halsey, A. H., editors, Power and Ideology in Education. OUP, Oxford. Bourdieu, P. (1977b). Outline of a Theory of Practice. CUP, Cambridge. Bourdieu, P. (1989). Language and Symbolic Power. Polity Press, Cambridge. Bourdieu, P. (1990). In Other Words. Polity Press, Cambridge. Bourdieu, P. and Boltanski, L. (1981). The education system and the economy: Titles and jobs. In Lemert, C., editor, Rupture and Renewal Since 1968. CUP, New York. Bourdieu, P. and Passeron, J.-C. (1990). Reproduction in Education, Society and Culture. Sage, London, second edition. 210

BIBLIOGRAPHY

211

Bourdieu, P. and Saint-Martin, M. (1974). Scholastic excellence and the values of the educational system. In Eggleston, J., editor, Contemporary Research in the Sociology of Education. Methuen, London. Bowles, S. and Gintis, H. (1976). Schooling in Capitalist America. Basic Books, New York. Breen, R. and Goldthorpe, J. (1997). Explaining educational differentials: Towards a formal rational action theory. Rationality and Society, 9(3):275–305. Brennan, G. (1990). What might rationality fail to do? In Cook, K. S. and Levi, M., editors, The Limits of Rationality. University of Chicago Press, Chicago. Carmines, E. and Zeller, R. (1979). Reliability and Validity Assessment. Sage, Beverley Hills. Cheung, S.-Y. (1997). Meritocracy Revisited: A Disaggregated Approach to the Study of Educational and Occupational Attainment in Britain. PhD thesis, Oxford University. Coleman, J. (1961). The Adolescent Society. The Free Press, Glencoe, IL. Coleman, J. (1990). Equality and achievement in education. Westview, Boulder, Colorado. Collins, R. (1971). Functional and conflict theories of educational stratification. American Sociological Review, 36:1002 – 1019. Cooper, B. (1976). Bernstein’s codes: a classroom study. volume 6 of Occasional Paper. University of Sussex Education Area. Crook, C. J. (1997). Cultural Practices and Socioeconomic Attainment: The Australian Experience. Greenwood Press, Westport, Connecticut. De Graaf, N. D., De Graaf, P., and Kraaykamp, G. (2000). Parental cultural capital and educational attainment in the Netherlands: A refinement of the cultural capital perspective. Sociology of Education, 73:92–111. De Graaf, P. (1986). The impact of financial and cultural resources on educational attainment in the Netherlands. Sociology of Education, 5(9). De Graaf, P. (1988). Parents’ financial and cultural resources, grades and transition to secondary school in the federal republic of germany. European Sociological Review, 4(3):209–231.

BIBLIOGRAPHY

212

DiMaggio, P. (1982). Cultural capital and school success. American Sociological Review, 47(2):189–201. DiMaggio, P. and Mohr, J. (1985). Cultural capital, education attainment and martial selection. American Journal of Sociology, 90(6):1231–1261. DiMaggio, P. and Mohr, J. (1995). The intergenerational transmission of cultural capital. Research in Social Stratification and Mobility, 14. Egerton, M. (1997). Occupational inheritance: The role of cultural capital and gender. Work, Employment and Society, 11(2):263–282. Elliott, J. and Richards, M. (1991). Children and divorce: Educational performance and behaviour before and after parental separation. International Journal of Law and the Family, 5:258–76. Elster, J. (1987). Ulysses and the Sirens. Cambridge University Press, Cambridge. Elster, J. (1990). When rationality fails. In Cook, K. S. and Levi, M., editors, The Limits of Rationality. University of Chicago Press, Chicago. Erikson, R. (1984). Social class of men, women and families. Sociology, 18(4):500–514. Erikson, R. and Jonsson, J. (1996). Can Education be Equalized? Westview Press, Boulder. Farkas, G., Grobe, R. P., Sheehan, D., and Shuan, Y. (1990). Cultural resources and school success: Gender, ethnicity and poverty groups within an urban school district. American Sociological Review, 55:127–42. Floud, J., Halsey, A., and Martin, F. (1956). Social class and educational opportunity. Heinemann, London. Gambetta, D. (1987). Were they Pushed or did they Jump? University Press, Cambridge.

Cambridge

Gewirtz, S., Ball, S., and Bowe, R. (1995). Markets, Choice and Equity in Education. Open University Press, Buckingham. Gipps, C. and Murphy, P. (1994). A fair test? : assessment, achievement and equity. Assessing assessment. Open University Press, Buckingham.

BIBLIOGRAPHY

213

Giroux, H. (1982). Power and resistance in the new sociology of education: Beyond theories of social and cultural reproduction. Curriculum Perspectives, 2(3). Graetz, B. (1988). The reproduction of privilege in australian education. British Journal of Sociology, 39(3):358–76. Halsey, A. H., Heath, A., and Ridge, J. (1980). Origins and Destinations. OUP, Oxford. Heath, A. and Britten, H. (1984). Women’s jobs do make a difference. Sociology, 18(4):475–490. Hellevik, O. (1997). Class inequality and egalitarian reform. Acta Sociologica, 40(4):377–98. Hurrel, P. (1995). Do teachers discriminate? Sociology, 29(1):59–72. Jencks, R. (1972). Inequality: A reassessment of the effect of family and schooling in America. Basic Books, New York. Jenkins, R. (1992). Pierre Bourdieu. Routledge, London. Jonsson, J. and Mills, C. (1993a). Social class and educational attainment in historical perspective: A swedish english comparison. part 1. British Journal of Sociology, 44(2):213–47. Jonsson, J. and Mills, C. (1993b). Social class and educational attainment in historical perspective: A swedish english comparison. part 2. British Journal of Sociology, 44(3):403–28. Jonsson, J. O. (1987). Class origin, cultural origin and educational attainment. European Sociological Review, 3(3):229–242. Kahneman, D., Slovic, P., and Tversky, A. (1982). Judgement Under Uncertainty. Cambridge University Press, Cambridge. Kalmijn, M. and Kraaykamp, G. (1996). Race, cultural capital and schooling: An analysis of trends in the United States. Sociology of Education, 69:22– 34. Katsillis, J. and Rubinson, R. (1990). Cultural capital, student achievement and educational reproduction: The case of Greece. American Sociological Review, 55(2):270–279.

BIBLIOGRAPHY

214

Kiernan, K. (1992). The impact of family disruption in childhood on transitions made in young adult life. Population Studies, 46:213–34. Labov, W. (1972). The logic of nonstandard english. In Giglio, P. P., editor, Language and Social Context. Penguin. Lawton, D. (1968). Social Class, Language and Education. Routledge and Kegan Paul. Layard, R. and Psacharopoulos, G. (1974). The screening hypothesis and the returns to education. Journal of Political Economy, 82:985–98. Levy, P. and Goldstein, H. (1984). Tests in Education. Academic Press, London. Lindsay, K. (1926). Social Progress and Educational Waste. Routledge, London. Manski, C. F. (1993). Adolescent econometricians. In Clotfelter, C. T. and Rothschild, M., editors, Studies of Supply and Demand in Higher Education. University of Chicago Press, Chicago. Marshall, G., Swift, A., and Roberts, A. (1997). Against the Odds? Social Class and Social Justice in Industrial Societies. Clarendon Press, Oxford. Micklewright, J. (1989). Choice at sixteen. Economica, 56:25–39. Mills, C. W. (1959). The Sociological Imagination. Oxford University Press, New York. Mohr, J. and DiMaggio, P. (1995). The intergenerational transmission of cultural capital. Research in Social Stratification and Mobility, 14. Murphy, J. (1981). Class inequality in education: Two justifications, one evaluation but no hard evidence. British Journal of Sociology, 32(2):182– 201. Murphy, J. (1990). A most respectable prejudice: Inequality in educational research and policy. British Journal of Sociology, 41(1):29–54. Nash, R. (1990). Bourdieu on education and social and cultural reproduction. British Journal of Sociology of Education, 11(4). NCDS (1993). National Child Development Survey. ESRC Data Archive, University of Essex.

BIBLIOGRAPHY

215

Nisbet, J. (1953). Family environment and intelligence. Eugenics Review, XLV:31–42. Office of Population Censuses and Surveys (1991). Standard Occupational Classification, volume 1-3. HMSO, London. Payne, J. (1987). Unemployment, apprenticeships and training: Does it pay to stay on at school? British Journal of Sociology of Education, 8(4):425–445. Power, S., Whitty, G., Edwards, T., and Wigfall, V. (1998). Schools, families and academically able students: contrasting modes of involvement in secondary education. British Journal of Sociology of Education, 19(2):157– 76. Robinson, R. and Garnier, M. (1985). Class reproduction among men and women in France. American Journal of Sociology, 91(2):250–280. Savage, M. and Egerton, M. (1997). Social mobility, individual ability and the inheritance of class inequality. Sociology, 31(4):645–72. Shavit, Y. and Blossfeld, H.-P. (1993). Persistent Inequality : changing educational attainment in thirteen countries. Social inequality series. Westview, Boulder, Colo. Smith, D. and Tomlinson, S. (1989). The School Effect: A Study of MultiRacial Comprehensives. Policy Studies Institute, London. Sokal, A. and Bricmont, J. (1998). Intellectual Impostures. Profile Books, London. Sullivan, A. and Cheung, S.-Y. (2000). The british case. In Ballarino, G., editor, Innovation, Flexibility, Training and Education: Link with the Family Situation. E.U. Commission. Walford, G. (1994). Choice and Equity in Education. Cassell, London. Waquant, L. (1989). Towards a reflexive sociology: A workshop with Pierre Bourdieu. Sociological Theory, 7. Weiner, G., Arnot, M., and David, M. (1997). Is the future female? In Halsey, A. H., Lauder, H., Brown, P., and Wells, A. S., editors, Education: Culture, Economy and Society. Oxford University Press, Oxford. Willis, P. (1977). Learning to Labour. Saxon House, Farnborough.

BIBLIOGRAPHY

216

Willis, P. (1983). Cultural production and theories of reproduction. In Barton, L. and Walker, S., editors, Race, Class and Education. Croom Helm, London.

Questionnaire – Private and Confidential Section 1 Education 1.

Are you: Male Female

2.

How many subjects are you studying for GCSE?

1

1

2

………….

2

What subjects are you studying for GCSE? …………………………………………………………… …………………………………………………………… …………………………………………………………… …………………………………………………………… …………………………………………………………… …………………………………………………………… ……………………………………………………………

3-20

…………………………………………………………… …………………………………………………………… …………………………………………………………… 3.

What grades do you think you are most likely to get in your GCSE’s? Tick one box only Mostly As

1

Mostly As and Bs

2

Mostly Bs and Cs

3

Mostly Cs and Ds

4

Mostly Ds and Es Mostly Es and Fs Mostly Fs and Gs

21

5

6

7

Mostly Gs and ungraded 8

1

4.

5.

6.

What do you expect to do after your GCSE’s? Tick one box only. Study for a BTEC

1



If so, which subject are you likely to do? ……………………………………………

22

Study for a GNVQ

2



If so, which subject are you likely to do? ……………………………………………

23

Retake GCSEs

3

Study for A levels

4



If so, which subjects are you likely to do? …………………………………………… ……………………………………………

Get a job

5

Join a training scheme

6

Other (please specify)

7



…………………………………………… If so, what sort of job? ……………………………………………



……………………………………………

Do you intend to go to university? Yes

1

No

2

24

How much do you agree or disagree with each of these statements about studying?

Agree strongly (a)

School is a waste of time

(b)

The more qualifications you get, the better the job you are likely to get

(c)

Studying is worthwhile for its own sake

(d)

I enjoy studying

(e)

Qualifications are useless for getting jobs

(f)

Studying can help you to gain skills which will be useful at work

Agree

Neither agree nor disagree

Disagree

Disagree strongly 25

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

26

27

28

29

30

2

Agree strongly

Agree

Neither agree nor disagree

Disagree strongly

Disagree

These days, it doesn’t matter how many qualifications you get, you still won’t get a job

1

2

3

4

5

(h)

Studying increases your confidence

1

2

3

4

5

(i)

Studying increases your ability to think clearly

1

2

3

4

5

(j)

I muck about in lessons

1

2

3

4

5

(k)

Studying is irrelevant to real life

1

2

3

4

5

(l)

I take school seriously

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

(g)

(m)

These days, if you are unqualified, you can’t get a job I don’t like studying

(n)

31

32

33

35

36

37

38

(o)

Studying improves your ability to be creative

1

2

3

4

5

(p)

Studying is only ever worthwhile if it leads to a job

1

2

3

4

5

7.

34

39

40

In general, how do you rate your academic abilities as compared to other pupils at your school? Excellent Above average Average Below average

1

41

2

3

4

Poor 5

Employment 1.

What sort of job would you most like to be doing in 10 years time?

42

……………………………………………………………….. 2.

What sort of job do you think you will actually be doing in 10 years time?

43

……………………………………………………………….. 3

3.

In choosing a job, what things about it do you think are important? Tick as many boxes as you like.

(a)

The job should involve working with my hands

(b)

It should make others respect me

(c)

It should be an outdoor job

(d)

It should involve using my head and need thought and concentration

(e)

It should allow me to excel

(f)

It should not get in the way of my family responsibilities

(g)

It should give me control over others

(h)

It should involve variety

(i)

It should be well paid

(j)

It should allow me to be creative

(k)

It should have convenient hours and conditions

(l)

It should give me the chance of being with other people

(m)

It should be a clean job

(n)

It should offer chances of promotion

(o)

The job should let me be my own boss

(p)

It should be a secure job

(q)

It should not have too much responsibility

(r)

It should give me the opportunity of helping others

44 1 45 1 46 1 47 1 48 1 49 1 50 1 51 1 52 1 53 1 54 1 55 1 56 1 57 1 58 1 59 1 60 1 61 1

Now look at the list again. Which of the items is: Most important 2nd most important 3rd most important

……… ……… ………

(just give the letter of the item)

62 63 64

4

Activities 1.

How often do you do each of these spare time activities (not organised by the school). Often

(a)

Watching television

(b)

Playing outdoor games and sports

(c)

Playing indoor sports

(d)

Going to parties at friends’ homes

(e)

Going to art galleries or museums

(f)

Voluntary work to help others

(g)

Going to the cinema

(h)

Going to see plays

(i)

Going to classical concerts

(j)

Going to pop concerts

(k)

Playing an instrument

(l)

Listening to classical music

(m)

2.

Listening to pop music

Sometimes

Hardly ever

Never 65

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

66

67

68

69

70

71

72

73

74

75

76

77

Which TV programmes do you watch regularly? Name as many as you can …………………………………………………………………………………….. …………………………………………………………………………………….. ……………………………………………………………………………………..

78-100

…………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. ……………………………………………………………………………………..

5

3.

What films have you seen at the cinema recently? (If you have not been to the cinema recently, move on to the next question). …………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. ……………………………………………………………………………………..

4.

About how often do you read books that are not connected to your schoolwork? Never / Hardly ever 1 per month 1 per fortnight 1 per week 2 per week

1

101-120

121

2

3

4

5

3 or more per week 6

5.

State the titles and/or authors of any books you have read recently that are not connected with schoolwork. ……………………………………………………………………………………..

122-130

…………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. 6.

7.

Are you a member of a public library? Yes

1

No

2

Yes

1

No

2

Do you have any favourite authors?

If so, who are they? …………………………………………………………………………………….. …………………………………………………………………………………….. …………………………………………………………………………………….. 8.

131

Do you sometimes read a national or local newspaper? If so which one(s)? …………………………………………………………………………………….. ……………………………………………………………………………………..

132

133-150

151-180

……………………………………………………………………………………..

6

Family 1.

Who do you live with? Adults:

mother and father

1

mother and stepfather

2

father and stepmother

3

just mother

4

just father other (please specify) ……………………………………………

181

5

6

Brothers and sisters (including step/half brothers and sisters) who are living with you, or who have left home in the last 3 years: 182

Number of brothers ………………. Number of sisters ……………….

Ages of brothers ……………………. Ages of sisters …………………….

183 184-190 191-200

In the following questions, I will be asking you about your mother and father. For some of you, your “natural” parents may not be relevant – answer in terms of who you see as your mother and father. 2.

3.

Is your mother employed?

Is your father employed?

Yes

1

No

2

Yes

1

No

2

201

202

7

4.

(a) (b) (c)

(d)

Please tell me about your parents’ jobs. If they are not working at the moment, tell me about their most recent jobs. Father What is the name of the job? ………………… ………………… What kind of work do they do? ………………… ………………… What sort of organisation do they work for? ………………… ………………… Do they have their own business?

Yes No

(e)

Do they work full or part time?

Full time Part time

5.

6.

Mother ………………… ………………… ………………… ………………… ………………… ………………… Yes

1

205

206 2

Full time

1

207 1

Part time

2

204

1

No

2

203

208 2

If you live with only one parent, or with a stepparent, are you in regular contact with the other parent (i.e. at least once a fortnight)? Yes

1

No

2

209

What is the highest qualification that each of your parents has? Father No qualifications CSE O level A level / Scottish higher Vocational qualification (please specify)

Mother 210

1

1

2

2

3

3

4

4

5

5

211

212

213

……………… Degree Post graduate qualification (masters, Ph.D. or D. Phil) 7.

……………….

6

6

7

7

Please state if either of your parents is studying for any of the above qualifications ……………………………………………………………………. …………………………………………………………………….

214 215

8

8.

Which of the following have you heard your parents discuss? Often

(a)

Sport

(b)

Art

(c)

Politics

(d)

Books

(e)

TV

(f)

Films

(g)

Work

(h)

Friends

(i)

Science

(j)

Current affairs

(k)

Music

9.

Sometimes

Never 216

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

217

218

219

220

221

222

223

224

225

226

Do your parents get a newspaper regularly (at least once a week)? Yes

1

No

2

10. Which newspaper(s) do they get? ………………………………………………… ………………………………………………… …………………………………………………

227

228-250

11. Do your parents listen to the radio?

If so, which station(s) do they listen to? ………………………………………………… ………………………………………………… …………………………………………………

Yes

1

No

2

251

252-270

9

12. Approximately how many books are in your home? None, or very few (0-10)

271 1

A few books (11-25)

2

One bookcase full (26-100)

3

Two bookcases full (101-200)

4

Three bookcases full (201-300)

5

Four bookcases full (301-400)

6

About a room full (401-600)

7

More than a room full (601 or more)

8

13. How many rooms are there in your home? Count all rooms including bathrooms ………..

272

14. Does your family own a car?

273

Yes

1

No

2

If your family owns more than one car, how many cars does it own? ……….. 15. Does your family have a computer?

16. Did your parents go on holiday last year?

274

Yes

1

No

2

Yes

275

276

1

No

2

If so, where did they go? ……………………………………..

277

17. For each of these activities, do your parents do it often, sometimes, rarely or never? Often (a)

Outdoor sports

(b)

Indoor sports

(c)

Watching television

(d)

Listening to music

(e)

Parties at friends’ homes

(f)

Going to art galleries or museums

Sometimes

Rarely

Never 278

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

279

280

281

282

283

10

Often (g)

Voluntary work to help others

(h)

Going to the cinema

(i)

Reading novels

(j)

Reading non fiction

(k)

Going to see plays

(l)

Going to concerts

(m)

Playing a musical instrument

(n)

Political meetings

(o)

Evening or daytime classes

Sometimes

Rarely

Never 284

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

285

286

287

288

289

290

291

292

18. If your parents listen to music, what kind(s) of music do they listen to? ………………………………………………….. …………………………………………………..

293-310

19. Do your parents have any friends who have the following jobs? (a)

Secretary

(b)

Nurse

(c)

Doctor

(d)

Manual worker

(e)

Unemployed

(f)

Academic

(g)

Teacher

(h)

Social worker

(i)

Cleaner

(j)

Student

(k)

Manager

(l)

Police

(m)

Foreman or supervisor

(n)

Clerical worker

(o)

Business owner

311 1 312 1 313 1 314 1 315 1 316 1 317 1 318 1 319 1 320 1 321 1 322 1 323 1 324 1 325 1

11

(p)

Technician

(q)

Accountant

(r)

Lawyer

(s)

Shop Worker

(t)

Housewife

326 1 327 1 328 1 329 1 330 1

Section 2 Famous people Each of the following names is a person you may have heard of. For each person listed, which do you associate him or her with most out of the following categories: politics, music, novels, art or science? If you do not know, do not guess, just tick “don’t know”. Politics Example:

Tony Blair

(a)

Albert Einstein

(b)

Jane Austen

(c)

Bill Clinton

(d)

Graham Greene

(e)

Wolfgang Mozart

(f)

Claude Monet

(g)

Karl Marx

(h)

Andy Warhol

(i)

Virginia Woolf

(j)

Marie Curie

(k)

Rachmaninov

(l)

Martin Amis

(m)

Galileo

(n)

Pablo Picasso

(o)

Rembrandt

Music

Novels

Art

Science

Don’t know

 331 1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

332

333

334

335

336

337

338

339

340

341

342

343

344

345

12

Politics (p)

Miles Davis

(q)

Charles Dickens

(r)

Louis Pasteur

(s)

Mahatma Gandhi

(t)

George Gershwin

(u)

John F. Kennedy

(v)

Stephen Hawking

(w)

Vincent van Gogh

(x)

Gordon Brown

(y)

Johannes Brahms

Music

Novels

Art

Science

Don’t know 346

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

347

348

349

350

351

352

353

354

355

Comprehension Underline the word that correctly completes the sentence. If you do not know which word is correct, just move on to the next question. Example: You (fill, milk, boil, match, paint) water to make tea. 1.

About a week after (ordering, receiving, emptying, managing, upsetting) this letter, I had a telegram from my brother.

2.

You can arrange a (money, summer, journey, job, weather) through a travel agent.

3.

She had been dieting for a month, but her weight had not (shown, increased, shrunk, decreased, grown).

4.

Despite the (convenience, expense, poverty, economy, pleasure) of living in this area, I have managed to save some money.

5.

The building that was going to be demolished had been (derisive, animated, vicious, derelict, derivative) for some months.

6.

The (process, proceeds, page, gain, progress) from the sale of the book were given to a worthwhile charity.

356

357

358

359

360

361

13

7.

It is not sufficient to know the means of preventing and curing disease; it is equally (bad, necessary, useful, sufficient, good) to provide these means and even to compel their use.

8.

The town council wanted to improve the (amounts, limits, amenities, places, inmates) of their district.

9.

As we both come from the same town, my wife and I have a great many (typical, mutual, friable, arable, viable) friends.

10. The discovery made by the explorers in the Arabian desert far (overtook, tormented, exceeded, exposed, passed) their wildest dreams. 11. When the speaker asked if there were any questions he was (involved, immured, inundated, implied, instructed) with queries. 12. Having had her expectations so much raised, it was very (realistic, discrediting, uplifting, disconcerting, discriminating) to have them suddenly shattered. 13. It took three men to (refrain, prohibit, restrain, catch, restrict) the youth from rushing into the burning shed to rescue his pet dog. 14. The boss firmly denied any accusation of discrimination; he claimed that the policies of his firm were not dictated by (shareholders, prejudice, incrimination, accusations, profit). 15. Although I won the debate, my opponent failed to acknowledge that I had (remaindered,

362

363

364

365

366

367

368

369

370

refuted, disputed, conquered, denied) his argument.

14

Synonyms Synonyms are words that mean the same or approximately the same as each other. For example happy and cheerful are synonyms. For each of the following questions think of as many synonyms as you can for each word (stop at five for each word). Here are two examples: Disgusting ………………………………………… repellent, nasty ………………………………………… vile ………………………………………… unpleasant ………………………………………… obscene …………………………………………

1.

Big

enormous ………………………………………… gigantic ………………………………………… huge ………………………………………… massive ………………………………………… obese …………………………………………

Small

…………………………………………

371

………………………………………… ………………………………………… ………………………………………… ………………………………………… 2.

Stupid

………………………………………… ………………………………………… ………………………………………… …………………………………………

372

………………………………………… 3.

Angry

………………………………………… ………………………………………… ………………………………………… ………………………………………… …………………………………………

373

15

5.

Sad

…………………………………………

374

………………………………………… ………………………………………… ………………………………………… ………………………………………… 6.

Odd

………………………………………… ………………………………………… ………………………………………… …………………………………………

375

…………………………………………

You have reached the end of this questionnaire. Please use any remaining time to go over your answers.

16

D.Phil Note, the PDF file has the wrong date. Correct citation: Sullivan, A. 2000. Cultural Capital, Rational Choice, and Educational Inequalities. D.Phil. Thesis, Oxford University. Publications from the doctorate: Sullivan, A. 2001. ‘Cultural Capital and Educational Attainment’ Sociology. 35(4) 893-912. Sullivan, A. 2002. ‘Bourdieu and Education: How Useful is Bourdieu’s Theory for Researchers?’ Netherlands Journal of Social Sciences. 38(2) 144-166 Sullivan, A. (2006) ‘Students as Rational Decision-Makers: the question of beliefs and attitudes’ London Review of Education 4(3) 271-290. Sullivan, A. 2007. ‘Cultural Capital, Cultural Knowledge and Ability’ Sociological Research Online 12(6).

Smile Life

When life gives you a hundred reasons to cry, show life that you have a thousand reasons to smile

Get in touch

© Copyright 2015 - 2024 PDFFOX.COM - All rights reserved.