Content uploaded by Alice Sullivan
Author content
All content in this area was uploaded by Alice Sullivan on Mar 24, 2014
Content may be subject to copyright.
Cultural Capital and Educational Attainment
Sullivan, A. 2001. ‘Cultural Capital and Educational Attainment’ Sociology. 35(4)
893-912.
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. I will conclude that cultural
capital is transmitted within the home and does have a significant
effect on performance in the GCSE (General Certificate of Secondary
Education) examinations. However, a large, direct effect of social
class on attainment remains when cultural capital has been controlled
for. Therefore, ‘cultural reproduction’ can provide only a partial
explanation of social class differences in educational attainment.
Education Attainment Inequality Cultural Capital
Social-class Gender
Word length: 6 800 words plus 5 tables, equivalent to 1 250 words.
2
1 Introduction
This paper will assess the merits of the cultural reproduction approach to the
examination of class and gender differentials in educational attainment.
I will address the following questions:
•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.
2 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. He
argues that 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 1977: 494)
3
Since, according to Bourdieu, the educational system presupposes the possession of
cultural capital, which only a minority of 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 ‘...minimise the risks by throwing a smoke-screen of vagueness
over the possibility of truth or error.’ (Bourdieu and Passeron 1977, 1990: 114)
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: 32)
In sum, Bourdieu’s view is that cultural capital is inculcated in the higher-class home,
and enables the higher-class student to gain higher educational credentials than the
lower class student. This enables higher-class individuals to maintain their class
positions, and legitimates the dominant positions that they typically go 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.
4
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. Indeed, he might himself
be accused of ‘…throwing a smoke-screen of vagueness over the possibility of truth
or error.’ 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. I will go on to discuss Bourdieu’s own attempt to apply empirical
evidence to his theory.
2.1 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: 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:
5
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.
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 one-to-one
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 écoles, 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 1977: 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 1981:145). In addition, the use of bivariate
analyses is crude. Clearly, a simple association between two variables is not
6
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. (Ibid.: 490-492).
However these figures are insufficient to 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.
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.
2.2 Operationalisation
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 by subsequent
researchers. 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 (1980) use this
proxy, as do Robinson and Garnier (1985) and Jonsson (1987).
7
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 (1982) and DiMaggio and Mohr (1985) 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, P.M. 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 capital, 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
cultural capital. 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 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, I
8
have used 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 (1997) and N.D. De Graaf et al. (2000)
follow P.M. De Graaf (1986, 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 (1997) and N.D. 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
9
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 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.
3 Methodology
I surveyed pupils in their final year of compulsory schooling (i.e. ‘year 11’ students,
about 16 years old) in England, in 1998. (I piloted the questionnaire in 1997). I chose
to survey year 11 pupils because this allowed me to follow up on the GCSE results
obtained subsequently by the pupils. The sample included four schools. Two of these
were co-educational, two single-sex. All were comprehensive. Cultural reproduction
theory is concerned with general processes, which are not contingent on any particular
school context. Therefore, while a representative sample of the year 11 population
10
might have been ideal, it should be borne in mind that I am not attempting to make
population estimates, but rather to examine processes that the theory of cultural
reproduction suggests should operate right across the educational system. As a student
collecting data independently, my sample size was inevitably restricted. Therefore, I
did not have the capacity to examine school type effects, and preferred to keep the
variable ‘school type’ constant as far as possible by restricting my sample to the
comprehensive sector.
I administered a questionnaire for self-completion by pupils. Pupils were not allowed
to confer while completing the questionnaire. Pupils and schools were assured of the
confidentiality of their responses.
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 83.5 per cent. The majority of the non-response was due to
absenteeism.
Taken as a whole, the sample provides a good spread in terms of social class. Schools
1 and 2 had a large proportion of service class families (44.8 per cent and 42 per cent
respectively), compared to schools 3 and 4 where 14.7 and 30.8 per cent of families
respectively were categorised as belonging to the service class1. The proportion of
families categorised as belonging to the skilled or unskilled manual classes was higher
in schools 3 and 4 (34.6 per cent and 28.6 per cent respectively) than in schools 1 and
2 (10.5 and 11 per cent respectively).
Parents’ social class and educational credentials were determined from pupils’
responses. The responses on parents’ occupations were re-coded using a six-category
11
version of the Goldthorpe class schema2, taking mother’s or father’s class, whichever
was the higher, as determined by a simplified version of Erikson’s (1984) dominance
schema3.
Mother’s or father’s qualifications were also selected according to which was higher.
The level of missing data on social class is 12 per cent (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). Therefore, I have
included these missing cases within my analyses as separate categories.
I have surveyed pupils on a broad range of possible components of cultural capital.
1. 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.
2. Cultural Knowledge
•Tested knowledge of famous cultural figures.
3. Language
•Active and passive vocabulary test scores.
For a detailed description of these variables, see Sullivan (2000).
12
The data on reading includes the types of books read as well as the amount read. Both
classic books and contemporary books of the sort that receive reviews in the quality
press were categorised as having cultural capital content. 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. 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.
Having asked pupils to list the television programmes they watched regularly, I
categorised these programmes according to their cultural capital content. Factual TV
programmes on science, arts or humanities and politics were categorised as having
cultural capital content. Non-factual programmes that are sophisticated in terms of
vocabulary and cultural references were also categorised as having cultural capital
content. Pupils were given a point for each of the following categories that were
included in the programmes they said they watched regularly: science (e.g. Horizon);
arts (e.g. The Late Review); politics, current affairs and humanities (e.g. Newsnight);
literary adaptations (e.g. Pride and Prejudice); and sophisticated comedies (e.g.
Frasier).
The test of cultural knowledge consisted of asking pupils to categorise 25 famous
cultural figures according to whether these figures are associated with politics, music,
novels, art or science. 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.
13
The test of passive vocabulary was a conventional ‘sentence completion’ test (see for
instance Levy and Goldstein 1984). The test of active vocabulary demanded that
pupils provide several synonyms for each of five words given.
I have also surveyed pupils on their parents’ cultural activities. These activities
include reading (and number of books in the home), newspapers taken, type of music
and radio stations listened to, participation in ‘formal culture’, and the subjects
discussed by parents in the home. It would have been difficult to get information
directly from the parents, as many parents would no doubt have been reluctant to
participate. It could be argued that pupils’ responses regarding their parents are
unreliable, as shared activities may be over-reported by pupils. However, note that De
Graaf et al. (2000) find that respondents’ own cultural practices have no effect on
their reporting of their parents’ cultural practices.
4 Analysis
4.1 Parental Cultural Capital
The first step in assessing the theory of cultural reproduction is to examine the
distribution of cultural capital by social class and parental education. The parental
cultural capital variable has mean 4.78 and standard deviation 3.89. Its maximum is
16. Service-class parents have a mean cultural capital score of 7.2, while non-service-
class parents have a mean score of 3.6. Graduate parents have a mean score of 8.5,
while non-graduate parents have a mean score of 3.8. Both these differences are
significant at the 0.001 level.
14
4.2 Pupils’ Cultural Capital
Having established an association between parental social class and cultural capital,
we can move on to the question of whether cultural capital is transmitted within the
home. To what extent is parental cultural capital associated with pupils’ cultural
capital, controlling for background variables?
4.2.1 Activities
[Table 1 here]
I used linear regression to analyse the determinants of the activities component of
pupils’ cultural capital. Table 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. Having a graduate parent and having a higher service-class parent are
significantly positively associated with pupils’ cultural activities. Parents’ cultural
capital is introduced in Model 2. This shows that parents’ cultural capital (with an eta2
of 0.233) is by far the most important factor in accounting for the variation in pupils’
cultural activities. (The eta2 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
capital has been included. This shows that the effect of these background variables on
pupils’ cultural activities is mediated by parents’ cultural capital. The effect of school
attended is 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
15
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 ≤ 0.000). The strength of this relationship provides support for
Bourdieu’s view that cultural resources are strongly transmitted from parents to
children.
4.2.2 Language and Knowledge
Pupils’ tested vocabulary and cultural knowledge scores are modelled in tables 2 and
3. Parental cultural capital 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 capital 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.
[Table 2]
I modelled the pupils’ vocabulary score in stages, 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 category
of the variable). There is not space to show overall significance and eta2 values for
categorical variables in the tables. However, I will occasionally refer to these values
in the text, as this may help to clarify the patterns of effects.
In Model 2, parental cultural capital is added. This shows that the effects of parents’
social class and qualifications are partially mediated by parental cultural capital. The
16
overall effect of parents’ qualifications is reduced from an eta2 of 0.041 in model 1 to
0.024 in model 2. The overall effect of social class is reduced from an eta2 of 0.065 to
0.043. 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. Note that, in this and subsequent models, the overall cultural
participation score is removed from the model before inserting the component scores.
So, the italicised effects do not form part of the original version of Model 3. On
Crook’s (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.
[Table 3]
Using the same procedure for pupils’ cultural knowledge, 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
17
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 capital 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 eta2 =
0.073 to eta2 = 0.031. Model 3 shows that the effect of parental cultural capital is itself
partially mediated by the activities component of pupils’ cultural capital. However,
the direct effect of parents’ cultural capital 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.
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’ 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 intelligence or whether pupils’ reading and TV viewing habits
simply reflect their level of measured intelligence. 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
18
later performance in tests of ability and examinations controlling for the earlier ability
score.
4.2.3 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.
[Table 4]
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.
4.3 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.
[Table 5]
The effects of the background variables on pupils’ GCSE scores are shown by Model
1. Compared to unskilled manual backgrounds, all non-manual backgrounds are
associated with increased GCSE performance, with higher-service-class backgrounds
19
providing the strongest advantage. Parents’ qualifications in the degree and
intermediate categories (A level or vocational) were significantly associated with
GCSE scores. Model 2 shows that these effects are mediated to an extent by parental
cultural capital. For instance, the overall effect of parents’ class on pupils’ GCSE
attainment is reduced from an eta2 of 0.109 in Model 1, to an eta2 of 0.077 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 capital on pupils’ GCSE scores is partially mediated
by the activities component of pupils’ cultural capital.
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. 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 capital on GCSE attainment are indeed
mediated in this way. In Model 3, parents’ cultural capital and the activities
component of pupils’ cultural capital both have highly significant effects. Once
pupils’ vocabulary and cultural knowledge scores are included, in Model 4, parents’
cultural capital 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 capital or pupils’ cultural activities. That the effects
of these variables are entirely mediated by cultural knowledge and language ability is
20
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 (barring the missing category) is rendered
insignificant by the inclusion of knowledge and language scores in the model.
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 eta2 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.
5 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.
21
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 capital. 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 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 TV viewing habits each 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 (1997) and N.D. 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
22
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’ educational 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 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 teachers are prejudiced against working-
class pupils because of their lack of cultural capital. (Note that Hurrell (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 élite (cultural or otherwise) who are prejudiced against non-élite pupils.
23
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.
So, 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 provide a complete account of these
differentials. In line with Bourdieu’s theory, cultural capital is associated with social
class, and is transmitted from parents to children. Again, in line with Bourdieu’s
theory the possession of cultural capital does have a significant effect on GCSE
24
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’.
ACKNOWLEDGEMENTS Many thanks to Anthony Heath and Geoffrey Walford.
This work was funded by an ESRC studentship.
25
1 The term ‘service class’ may be misleading, as it may suggest the service sector of the economy. In fact, in the context
of the Goldthorpe class schema, the term ‘service class’ denotes positions of ownership and control.
2 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)
VIIa+VIIb=6 (unskilled manual)
3 1 dominates 2, 2 dominates 4, 4 dominates 3, 3 dominates 5, 5 dominates 6.
4This is a normal level of non-response in national surveys where people are asked about their parents’ occupations,
such as the British Election Survey.
Bibliography
BOURDIEU, P. 1974. The School as a Conservative Force. In Eggleston, J. (ed.) Contemporary
Research in the Sociology of Education. London: Methuen.
BOURDIEU, P. 1977. Cultural Reproduction and Social Reproduction. In KARABEL, J. and
HALSEY, A.H. (eds.) Power and Ideology in Education. Oxford: Oxford University Press.
BOURDIEU, P. and PASSERON, J-C. 1977, 1990. Reproduction in Education, Society and
Culture. London: Sage.
BOURDIEU, P. and BOLTANSKI, L. 1981. The Education System and the Economy. In LEMERT,
C.C. (ed.) Rupture and Renewal Since 1968. New York: Columbia University Press.
CROOK, C.J. 1997. Cultural Practices and Socioeconomic Attainment: The Australian Experience.
Westport, Conneticut: Greenwood Press.
DALE, R. and PIRES, E. 1984. Linking People to Jobs. in BROADFOOT, P. (ed.) Selection,
Certification and Control. London: Falmer.
DE GRAAF, P.M. 1986. ‘The Impact of Financial and Cultural Resources on Educational
Attainment in the Netherlands’. Sociology of Education 59, 4: 237-46.
DE GRAAF, P.M. 1988. ‘Parents’ financial and cultural resources, grades, and transition to
secondary school in the Federal Republic of Germany.’ European Sociological Review. 4, 3: 209-
221.
DE GRAAF, N.D., DE GRAAF, P.M. 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.
DIMAGGIO, P. 1982. ‘Cultural Capital and School Success’. American Sociological Review 47:
189-201.
DIMAGGIO, P. and MOHR, J. 1985. ‘Cultural Capital, Educational Attainment, and Marital
Selection’. American Journal of Sociology 90, 6: 1231-61.
EGERTON, M. 1997. ‘Occupational Inheritance: The Role of Cultural Capital and Gender’. Work,
Employment and Society 11, 2: 263-282.
ERIKSON, R. 1984. ‘Social Class of Men, Women, and Families.’ Sociology 18, 4: 500-514.
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: 127-142.
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. Oxford: Clarendon
Press.
HURRELL, P. 1995. ‘Do Teachers Discriminate?’. Sociology 29, 1: 59-72.
JONSSON, J.O. 1987. ‘Class Origin, Cultural Origin and Educational Attainment’. European
Sociological Review 3, 3: 229-42.
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’. American Sociological Review 55, 2: 270-279.
LEVY, P. and GOLDSTEIN, H. 1984. Tests in Education. London: Academic Press.
MACKINNON, D., STATHAM, J. and HALES, M. 1995. Education in the UK: facts and figures.
Buckingham: Open University Press.
ROBINSON, R. and GARNIER, M. 1985. ‘Class Reproduction among Men and Women in
France’. American Journal of Sociology 91, 2: 250-58.
SAVAGE, M. and EGERTON, M. 1997. ‘Social Mobility, Individual Ability and the Inheritance of
Class Inequality’. Sociology 31, 4: 645-72.
SULLIVAN, A. 2000. Cultural Capital, Rational Choice, and Educational Inequalities. D.Phil.
Thesis, Oxford University.
WAQUANT, L.J.D. 1989. ‘Towards a Reflexive Sociology: A Workshop with Pierre Bourdieu’.
Sociological Theory 7, 3: 26-63.
3
4
