International Review for the

Sociology of Sport
http://irs.sagepub.com

The Paradox of Social Class and Sports Involvement: The Roles of

Cultural and Economic Capital
Thomas C. Wilson
International Review for the Sociology of Sport 2002; 37; 5
DOI: 10.1177/1012690202037001001
The online version of this article can be found at:
http://irs.sagepub.com/cgi/content/abstract/37/1/5

Published by:
http://www.sagepublications.com

On behalf of:
International Sociology of Sport Association

Additional services and information for International Review for the Sociology of Sport can be found
at:
Email Alerts: http://irs.sagepub.com/cgi/alerts
Subscriptions: http://irs.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.co.uk/journalsPermissions.nav
Citations http://irs.sagepub.com/cgi/content/refs/37/1/5

Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 5

INTERNATIONAL REVIEW FOR THE SOCIOLOGY OF SPORT 37/1(2002) 516

Copyright ISSA and SAGE Publications (London, Thousand Oaks, CA, New Delhi)
[10126902 (200203) 37:1;516; 021851]

THE PARADOX OF SOCIAL CLASS AND SPORTS

INVOLVEMENT
The Roles of Cultural and Economic Capital

Thomas C. Wilson
Florida Atlantic University, USA
Abstract Studies in the sociology of sport have found that the higher ones social class, the greater
is ones overall involvement in sports, but the less likely is ones involvement in what have come
to be called prole sports. Using data from the 1993 General Social Survey, this study tests two
explanations for this paradox, one stressing class-based differences in cultural capital and the other
emphasizing class-based differences in economic capital. Findings show that those who are richest
in cultural capital and those richest in economic capital are most likely to be involved in sports
generally, and that these tendencies are independent of one another. However, those richest in cultural
capital are least likely to be involved in prole sports, and economic capital has no bearing on prole
sports involvement. In all, cultural capital explains the paradox of social class and sports involvement
better than economic capital does. Inferences are drawn for the role of sports involvement in the
reproduction of social inequality, and for the cultural omnivore thesis.
Key words Americans sport consumption cultural capital social class

Sports Involvement and Social Class

Sociology of sport findings present a paradox. On the one hand, the higher ones
social class, the more likely one is to be involved in sports. But on the other hand,
the higher ones class, the less likely one is to be involved in certain sports
that have come as a result to be associated with the lower classes. Studies have
repeatedly shown that indicators of social class are positive predictors of sport
involvement in general and that members of the upper classes are more likely to
be both sports participants and sports spectators (Bourdieu, 1984; Coakley, 1998;
Curtis and Milton, 1976; Eitzen and Sage, 1991: 304; Erickson, 1996; Hughes
and Peterson, 1983; Leonard, 1998; Nixon and Frey, 1996; Scholsberg, 1987;
Yergin, 1986; Young and Willmott, 1973). However, social class is inversely
related to involvement in certain prole sports, so-called because they are
avoided by the upper classes and have therefore become associated with the
proletariat or working class (Curry and Jiobu, 1984; Eitzen and Sage, 1991;
Nixon and Frey, 1996). For example, Bourdieu (1978) found that the French
upper classes were more likely to play golf and tennis and to go skiing than the
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 6

INTERNATIONAL REVIEW FOR THE SOCIOLOGY OF SPORT 37(1)

working class was, but less likely to be interested in boxing, rugby, bodybuilding, and football. American studies have produced similar results. Yergin (1986)
found that, while the upper classes are more likely to attend most sporting events,
they are less likely than the lower classes to attend wrestling and boxing matches.
Scholsberg (1987) came to the same conclusion for wrestling and boxing, and
also found that the upper classes are less likely to go bowling, lift weights, or be
among rodeo or roller derby spectators. In the same vein, Eitzen and Sage (1991)
identify bowling, wrestling, and contact sports in general as more attractive to the
working class than to the upper classes.
An explanation for this paradox can be drawn from Pierre Bourdieus concept of cultural capital. According to Bourdieu (1978, 1984; see also Collins,
1979; Holt, 1997, 1998), all cultural consumption including sports consumption
requires the appropriate preferences and tastes as well as skills and knowledge,
which he terms cultural capital. Cultural capital is gained from ones upbringing
and education. Critically, cultural capital varies by social class, and in fact serves
as a marker and a legitimater of social differences. Most sports are consistent with
the preferences of the upper classes, either because they exemplify virtues that the
upper classes hold dear or because like art, music, and pure academics, they are
pursued as ends in themselves rather than for instrumental purposes (Bourdieu,
1978; Lamont, 1992: 121). However, some sports are inconsistent with or even
antithetical to upper class preferences and are therefore rejected by the upper
classes as a negative assertion of their tastes (Bourdieu, 1978: 20, 1984: 56; see
also Holt, 1998; Lamont, 1992; Peterson, 1997). Grounds for this rejection are
varied, and include a sports emphasis on artifacts or skills that are devalued in
the upper class milieu, or a sports treatment of the body as an instrument toward
some end rather than as an object of cultivation for its own sake (Bourdieu,
1984). In contemporary America, the upper classes tend to avoid sports that stress
physical contact, toughness, asceticism, and hard manual labor, the so-called
prole sports (Eitzen and Sage, 1991; Nixon and Frey, 1996).
There is another explanation that stresses economic capital rather than
cultural capital. Sports involvement either as a participant or a spectator requires
both money and leisure time, and the upper classes have more of both (Bourdieu,
1978; Coakley, 1998; Eitzen, 1996; Nixon and Frey, 1996; for evidence of
erosion in class-based differences in leisure, see Rojek, 2000; Schor, 1991).
Prole sports are relatively inexpensive, however, and for this reason they are
particularly attractive to lower class persons (Mandell, 1984: 278; Nixon and
Frey, 1996: 206). In short, class-based differences in economic capital enable
upper class involvement in expensive sports, leaving prole sports largely
relegated to the lower classes.
Evaluating these two explanations would require assessing the separate
effects of cultural capital and economic capital on involvement in sports (Holt,
1997). But notwithstanding numerous studies linking class indicators with sports
consumption, an assessment of cultural and economic capitals respective independent effects has been reported in only a single study that has recently appeared
in this journal. Using data from the 1992 General Social Survey of Canada, White
and Wilson (1999) analyzed spectatorship at amateur and professional sporting
events. Following Bourdieu (1978, 1984), they measured cultural capital with
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 7

WILSON: SOCIAL CLASS AND SPORTS INVOLVEMENT

respondents education and economic capital with respondents household

income. Results of their analysis showed that both income and education were
directly and independently related to attendance at sporting events, from which
White and Wilson concluded that both economic and cultural capital promote
Canadians sports involvement.
In the following study, I extend this line of research by testing two hypotheses drawn from the implications of cultural and economic capital for sports
involvement reviewed above. The first hypothesis is that both cultural and economic capital independently promote sports involvement in general. White and
Wilsons (1999) study confirms this for sports attendance among their Canadian
sample. My analysis will focus on both attendance and sports participation, using
American data. The second hypothesis is that both cultural and economic capital
will retard involvement in prole sports. To my knowledge no prior study has
addressed this issue.

Data and Methods

The following analysis is based on data for a representative sample of Americans
contained in the 1993 NORC General Social Survey (Davis and Smith, 1998). In
the survey, respondents indicated if they had engaged in each of a list of leisuretime activities during the previous year. Two of the activities pertained to sports
involvement generally: attendance at any sports event and participation in any
sport. One pertained to a particular genre of prole sport: attendance at an auto,
stock car, or motorcycle race. As Aveni (1976) has noted, attendance at these
events may also imply participation, because at the amateur level many racing
spectators are also contestants. Prior studies report that involvement in sports of
this sort are inversely related to social class indicators (e.g. Leonard, 1998;
Martin and Berry, 1987; Scholsberg, 1987). Eitzen and Sage (1991) suggest that
their popularity among the lower classes may be cultural: such sports are not
school-related, they emphasize speed and violence, and the relevant artifacts and
skills (cars and driving) are familiar to lower class culture. Curry and Jiobu
(1984) provide some historical context, speculating that auto-racings appeal to
the lower classes is a logical extension of Appalachian whiskey running.
Involvement in these activities is shown in Table 1. Men are more involved
than women are. Roughly three in five men report attending a sports event, and
the same proportion has actively taken part in sport. Among women, only about
half have attended and half have participated. Men are also more likely to have
gone to an auto or cycle race, a bit less than a quarter of them having done so,
compared to just over one in ten women.
In the following analysis, I assess the impact of both economic and cultural
capital on the sports involvement indicators in Table 1. Following White and
Wilson (1999), I operationalize economic capital with respondents household
income, and cultural capital with respondents educational attainment. This
approach to measuring cultural capital is consistent with Bourdieus (1984: 23)
argument that education transmits class-based culture intergenerationally in the
form of dispositions directed both toward scholastic knowledge and also beyond
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 8

INTERNATIONAL REVIEW FOR THE SOCIOLOGY OF SPORT 37(1)

Table 1 Frequency of Sports Involvement within Previous Year,

among Men and Women (%)

Attend an amateur or professional sports event

Participate in any sports activity such as softball,
basketball, swimming, golf, bowling, skiing or
tennis
Go to an auto, stock car, or motorcycle race

Men (N = 646)

Women (N = 812)

61.6

49.3

63.9
22.9

52.3
10.5

1993 GSS, N=1458. All differences between men and women are significant, p < .000.

the curriculum.1 Similary, Dimaggio and Useem (1978) have noted that, once
class-based preferences evolve, they are maintained intergenerationally in large
measure by educational reinforcement.
The analytic strategy in the analysis is Multiple Classification Analysis
(MCA), again following White and Wilson (1999).2 In the MCAs, I control four
demographic variables found in prior studies to be related to sports involvement:
race, age, region, and community size (Hughes and Peterson, 1983; Lamont,
1992: 121; Scholsberg, 1987; Yergin, 1986; White and Wilson, 1999).3 Separate
analyses will be presented for men and women because my preliminary analyses
of the GSS data showed significant differences by gender in the relationship of
education and income to sports involvement.

Results
Sports Attendance and Participation
Table 2 pertains to sports attendance and addresses the hypothesis that cultural
capital and economic capital each promote sports involvement. Model 1 presents
the bivariate relationships of income and education with sports attendance, and
shows for both men and women alike that more affluent people and better
educated people are more likely to attend sports events. In the mens and in the
womens analysis, the value of the eta coefficients are all of roughly the
same magnitude, indicating that the incomeattendance relationships and the
educationattendance relationships are all of approximately equal strength, and
that there is no difference in their magnitudes between men and women.
Model 2 presents the independent effects from a multivariate analysis for
economic and cultural capital, where the effect of each is adjusted for the other.
Again, among both men and women, education and income remain significantly
and directly related to sports attendance. This means that more affluent people are
more likely to attend sports events regardless of their education. A case in point:
women in the highest income category are more than twice as likely to have
attended than women in the lowest income category (68.9% vs 33.9%). And, for
men and women alike, better educated people are more likely to attend sporting
events, again regardless of their incomes. To illustrate, men who did not graduDownloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 9

WILSON: SOCIAL CLASS AND SPORTS INVOLVEMENT

Table 2 Frequency of Sports Attendance by Income and

Education (%)
Men (N = 646)

Women (N = 812)

(1)

(2)

(3)

(1)

(2)

(3)

Income
< $9999
$1019,999
$2029,999
$3039,999
$4049,999
$5059,999
$6074,999
$75,000 +

29.6
47.1
57.9
67.9
64.9
78.8
84.0
75.0

39.5
53.3
60.0
66.8
63.7
73.4
75.7
67.1

40.9
55.2
59.7
65.9
61.4
71.4
76.6
67.8

26.2
39.3
42.6
59.7
57.9
59.6
72.4
76.9

33.9
41.3
42.7
57.5
55.5
54.5
67.8
69.8

36.6
42.5
40.9
57.3
55.0
53.6
66.4
67.7

eta/beta

.323***

.209***

.194***

.326***

.239***

.215***

Education:
< high school graduate
high school graduate
some college
college graduate
postgraduate degree

30.6
58.7
71.7
74.2
82.4

38.1
58.7
69.7
71.5
77.7

43.4
59.1
66.4
70.2
76.0

25.8
42.7
59.3
61.1
74.0

35.0
45.1
56.1
54.7
65.6

39.2
45.0
54.5
52.9
64.7

eta/beta

.362***

.279***

.221***

.296***

.184***

.149***

1993 GSS, N=1458. Model 1: bivariate incomeattendance and educationattendance relationships.

Model 2: incomeattendance relationship net of education; educationattendance relationship net
of income. Model 3: incomeattendance relationship net of education, race, age, region and
community size; educationattendance relationship net of income, race, age, region and
community size. * p < .05; ** p < .01; *** p < .001.

ate high school are only about half as likely to attend compared to men with postgraduate degrees (38.1% vs 77.7%). Among both men and women, model 2 beta
coefficients are smaller than comparable eta coefficients in model 1, indicating
that the relationships of income and education with attendance are each somewhat attenuated when the other is controlled. Model 2 betas also show that among
men, the impact of education on attendance is somewhat greater than that of
education (beta = .279 vs .209), whereas if anything the opposite is the case
among women (beta = .184 vs .239).
Model 3 repeats the analysis, this time with additional controls for race, age,
region, and community size. Model 3 beta coefficients are consistently smaller
than their counterparts in model 2, indicating that the relationships of income and
education with sports attendance are further attenuated with the additional demographic controls. However, they consistently remain statistically significant and
robust. For example, for men and women in the highest income category or the
highest education category, at least two-thirds had attended a sporting event,
compared to fewer than half of those in the lowest income or education categories. Among men, the independent effects of income and education are of
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

10

Page 10

INTERNATIONAL REVIEW FOR THE SOCIOLOGY OF SPORT 37(1)

Table 3 Frequency of Sports Participation by Income and

Education (%)
Men (N = 646)

Women (N = 812)

(1)

(2)

(3)

(1)

(2)

(3)

Income:
< $9999
$1019,999
$2029,999
$3039,999
$4049,999
$5059,999
$6074,999
$75,000 +

33.8
47.1
59.6
70.8
71.9
78.8
76.0
83.8

45.0
54.8
62.3
69.9
70.9
72.0
66.2
73.7

7.5
58.2
61.1
67.4
66.8
69.7
68.4
75.6

24.2
42.3
52.5
62.2
65.8
61.7
72.4
78.2

34.4
45.1
52.2
59.0
62.2
55.5
66.6
69.7

37.5
46.5
50.4
58.1
61.6
54.5
64.8
67.3

eta/beta

.325***

.188***

.162***

.344***

.227***

.193***

Education:
< high school graduate
high school graduate
some college
college graduate
postgraduate degree

31.3
55.8
77.6
79.4
87.9

37.6
56.0
76.1
77.0
83.4

46.9
56.8
70.7
73.8
80.7

22.6
46.3
65.7
67.3
74.0

32.3
47.9
62.4
61.3
66.5

38.1
48.0
60.0
59.2
64.4

eta/beta

.416***

.345***

.246***

.348***

.236***

.179***

1993 GSS, N=1458. Model 1: bivariate incomeattendance and educationattendance relationships.

Model 2: incomeattendance relationship net of education; educationattendance relationship net
of income. Model 3: incomeattendance relationship net of education, race, age, region and
community size; educationattendance relationship net of income, race, age, region and
community size. * p < .05; ** p < .01; ** p < .001.

roughly equal magnitude (betas = .194 and .221), while among women the effect
of income is somewhat stronger than that of education (betas = .215 vs .149).
In all, Table 2 confirms the first hypothesis, showing the cultural capital indicated by education, and economic capital indicated by income, each independently promotes sports attendance. These results for the American GSS sample
are similar to White and Wilsons (1999) results for their Canadian sample. In
Table 3, additional results confirming the hypothesis are presented for sports
participation. Model 1 bivariate relationships show that, among both among men
and women alike, more affluent people and better educated people are more
likely to be sports participants. Among men, the education relationship is somewhat stronger than the income relationship (eta = .416 vs .325), whereas the
relationships are of equal strength among women (eta = .348 vs .344).
Table 3s model 2 shows that income and education each influence sports
participation independent of one another. Among both men and women, better
educated people are more likely to participate regardless of income, and more
affluent people are more likely to participate regardless of education. Comparing
model 2 beta coefficients with comparable eta coefficients in model 1, the respecDownloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 11

WILSON: SOCIAL CLASS AND SPORTS INVOLVEMENT

11

tive influence of income and of education is each attenuated somewhat when the
other is controlled, but without exception they remain significant and strong. For
example, women in the lowest income category are again less than half as likely
to be sports participants compared to highest income women (34.4% vs 69.7%),
and women not graduating high school are less than half as likely to participate
than those with postgraduate degrees (32.3% vs 66.5%). For men, income results
are only a little less striking, showing a 73.7 percent participation rate among the
most affluent compared to 45.0 percent for those of lowest income. Mens education results are somewhat stronger than those observed for women, however, showing that while 83.4 percent of graduate-degree holders participate in sports, only
slightly more than a third, 37.6 percent, of men not graduating high school do.
Model 3 in Table 3 adds the demographic controls, and the influences of
income and education are further attenuated (as indicated by lower betas coefficients in model 3 compared to their counterparts in model 2). But for men and
women alike the influences remain significant and dramatic. Among men, for
example, 75.6 percent of the most affluent and 80.7 percent of the best educated
participate in sports, compared to fewer than half of the least affluent or the least
educated. Among women, roughly two-thirds of both the most affluent and of the
best educated participated in sports, compared to little more of a third of those
with lowest incomes or least education. The influence of income and of education on sports participation are roughly equal for women (betas =.193 and .179),
whereas for men the influence of education is somewhat stronger than for income
(betas = .246 vs .162).
Involvement in Prole Sports
Table 4 addresses the hypothesis that cultural and economic capital each retard
involvement in prole sports, and focuses on attendance at auto and cycle races.
Model 1 shows that economic capital as reflected by household income has little
to do with auto and cycle racing attendance. Among both genders there is some
tendency for people with low and modest incomes (under $10,000 for men, and
under $30,000 for women) and those with high incomes (over $75,000 for both
men and women) to attend racing events less often than those with income falling
between these extremes. But the incomeattendance relationship is never significant in model 1, nor in the multivariate analyses in models 2 and 3.4
Racing attendance is strongly related to cultural capital, however, at least
among men. Model 1 shows that less educated men are far more likely to go to
auto and cycle races compared to better educated men. The contrast is sharpest
between high school graduates, fully a third of whom went to racing events,
and men with graduate degrees, only 6.7 percent of whom did so. There is an
anomaly: men not graduating high school attend racing less often than men with
high school but not college diplomas and those with some college, though more
often than those with college and graduate degrees. Model 2 repeats the analysis
controlling for income and model 3 adds demographic controls, and results
remain virtually unchanged.
Cultural capital has a somewhat weaker influence on womens racing attendance. Model 1 shows that the educationattendance relationship is not signifiDownloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

12

Page 12

INTERNATIONAL REVIEW FOR THE SOCIOLOGY OF SPORT 37(1)

Table 4 Frequency of Attendance at Auto or Motorcycle Race by

Income and Education (%)
Men (N = 646)

Women (N = 812)

(1)

(2)

(3)

(1)

(2)

(3)

Income:
< $9999
$1019,999
$2029,999
$3039,999
$4049,999
$5059,999
$6074,999
$75,000 +

16.9
26.5
25.4
26.4
28.1
21.2
22.0
13.8

16.1
24.4
23.3
24.8
28.0
20.9
25.5
20.2

16.4
24.8
22.1
23.9
27.2
20.8
25.9
23.1

5.4
9.8
9.8
15.1
15.8
12.8
12.1
7.7

4.9
9.4
10.0
15.1
15.3
13.7
12.6
8.9

5.7
10.1
9.2
14.4
15.2
13.4
11.9
8.9

eta/beta

.112

.078

.068

.113

.116

.102

Education:
< high school graduate
high school graduate
some college
college graduate
postgraduate degree

21.6
33.1
27.0
15.5
6.7

22.6
32.4
26.8
15.7
6.5

25.8
31.9
25.7
14.9
5.7

9.0
9.4
14.8
10.6
4.1

11.7
9.4
13.7
9.6
3.2

14.5
9.4
12.6
8.8
2.0

eta/beta

.210***

.204***

.208***

.100

.094

.111*

1993 GSS, N = 1458. Model 1: bivariate incomeattendance and educationattendance

relationships. Model 2: incomeattendance relationship net of education; educationattendance
relationship net of income. Model 3: incomeattendance relationship net of education, race, age,
region and community size; educationattendance relationship net of income, race, age, region and
community size. * p < .05; ** p < .01; *** p < .001.

cant, and that while attendance is lowest among the best educated (4.1% for holders of graduate degrees) it is highest not among the least educated (9% among
those without high school diplomas) but among women with some college education (14.8%). Model 2 controls for income, and these results change little.
However, with demographic controls added in model 3, a significant influence of
education does emerge. Women holding postgraduate degrees remain least likely to attend racing (2.0%) and this time least educated women, those without high
school diplomas, are the most likely to do so (14.5%). However, attendance is
nearly as high among women having some college education (12.6%) and differs
little between high school graduates and college graduates (9.4% vs 8.8 %).

Discussion
This study has addressed a paradox in the sociology of sport: the upper classes
are more involved in sports overall, but they are less involved in certain prole
sports. Theoretically this paradox can be explained by class-based differences in
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 13

WILSON: SOCIAL CLASS AND SPORTS INVOLVEMENT

13

cultural capital and by class-based differences in economic capital. To evaluate

these explanations, I have tested two hypotheses drawn from them: first, cultural
capital and economic capital each promote sports involvement generally; and
second, both cultural and economic capital retard involvement in prole sports.
Operationalizing cultural capital with educational attainment and economic
capital with household income, the analysis presented here has found unambiguous support for the first hypothesis. Among both men and women alike, economic capital and cultural capital promote attendance at sporting events and
participation in sports, each does so independent of the other, and each does so
independent of selected demographic variables as well.
Findings for the second hypothesis were mixed. There is no evidence that
economic capital exerts any influence on prole sport involvement, at least as
reflected by attending auto and cycle racing. However, there is some evidence for
women and far stronger evidence for men that those with greater cultural capital
are less involved in racing sports compared to those whose cultural capital is
limited.
Considered together, these results provide little support for class-based
differences in economic capital as an explanation for the paradox of social class
and sports involvement. Those rich in economic capital are more involved in
sports generally, presumably because they can better afford their cost, both in
terms of money and leisure time. But those with limited economic capital show
no particular affinity for prole sports, regardless of those sports relative affordability.
Bourdieus concept of cultural capital provides a far better explanation.
Those rich in cultural capital are more prone to sports involvement generally but
less likely to be involved in prole sport, and this implies that sports consumption is to a large degree motivated by preferences, tastes, skills, and knowledge
that vary by social class. Apparently, most sports well fit the tastes and preferences of the upper classes, but some sports like auto and cycle racing do not.
Sports of this latter sort are therefore avoided by the upper classes. As a result of
upper class avoidance, and perhaps also because these sports better correspond to
lower class tastes (a possibility that has not been feasible to test in this study),
auto and cycle racing along with other so-called prole sports attract participants
and spectators drawn largely from the lower classes. Critically, all of this occurs
independent of class-based differences in economic capital, so the influence of
cultural capital on sports consumption need have nothing to do with the ability to
pay.
I draw two inferences relevant to broader stratification issues from this
studys findings. The first pertains to the reproduction of social inequality.
Bourdieu has argued that class differences in taste are means of reproducing
status-based social networks that in turn provide access to material and symbolic
goods (Bourdieu, 1984; see also Collins, 1979; Douglas and Isherwood, 1979).
Others have argued similarly, noting that taste functions as a means of ritual
identification in the construction of social relations. Those with high-cultural
tastes prefer interacting with each other, but neither they nor those they exclude
necessarily intend or even recognize the social reproductive implications
(DiMaggio, 1987; DiMaggio and Ostrower, 1990; Holt, 1997; Lamont, 1992).
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

14

29/1/02

10:15 am

Page 14

INTERNATIONAL REVIEW FOR THE SOCIOLOGY OF SPORT 37(1)

Empirical evidence for these claims is limited but generally supportive and shows
that similarity in taste does in fact influence ones choice of friends, associates,
and spouse (DiMaggio and Mohr, 1985; Lamont et al., 1996). It has been
suggested that tastes in sports may also function in this way (Booth and Loy,
1999; Eitzen and Sage, 1991). The findings presented in this study showing that
sports tastes are linked not only to economic capital but particularly to classbased differences in cultural capital strongly imply that, along with other class
differences in taste, sports tastes do in fact function to accommodate and reinforce the existing structure of social inequality.
The second inference pertains to the emergence of the cultural omnivore. A
number of studies suggest that high-brow snobbery, centered around upper class
cultural pursuits and involving the repudiation of common tastes, has been
replaced as a status marker by more cosmopolitan and eclectic tastes characteristic of what has been called the cultural omnivore (DiMaggio, 1987; Erickson,
1996; Lamont, 1992; Levine, 1988; Peterson, 1997). Much of this research has
focused on tastes in music. For example, Peterson and Simkus (1992) report that
among the elite only a minority consider classical music to be their favorite
musical genre, and more favor country music than favor opera. In the same vein
Peterson and Kern (1996) found that fans of classical music and opera are more
likely than others to also enjoy middle-brow and low-brow music genres. At
least one study suggests that omnivorism has not entirely replaced snobbery,
however. Bethany Bryson (1996) reports in her study based on the same 1993
GSS data set I have used in this study that the greater ones education (and
implicitly, ones cultural capital) the broader ones musical tastes, but the broader
ones tastes, the more likely one is to dislike those musical genres that are most
favored by the least educated. She concludes that, while high-status cultural
tolerance seems to be the current rule, it is not indiscriminate and instead continues to reject markedly low-status genres. My findings for sports tastes parallel
Brysons findings for music. That those richest in cultural capital are generally
more involved in sports is consistent with the cultural omnivore thesis (though
strictly speaking my findings show only that the elites participation is more
frequent, not necessarily more varied). However, my findings also show that
those richest in cultural capital apparently dislike prole sports like auto and
cycle racing, suggesting that there are strict limits to any cultural omnivorism in
sports tastes among the elite. It thus appears that, to some extent at least, cultural
capital continues to involve the classification of consumption items into the more
and the less valued, and to promote the elites disdain for the latter.

Notes
1.

As Lamont and Lareau (1988) have noted, Bourdieus concept of cultural capital has been
operationalized in various ways including educational attainment but also high culture knowledge and participation (see e.g. DiMaggio, 1982; DiMaggio and Mohr, 1985; Dimaggio and
Useem, 1978). Holt (1997) contends that Bourdieus concept has a twofold meaning: a fieldspecific form, exemplified by specific tastes in art, food, music, and the like; and an abstracted
virtual form consisting of generic transposable dispositions, tastes, knowledge and the like,
accumulated primarily through social class background. It is this latter sense of cultural capital,
for which educational attainment is an appropriate operationalization, that is used in this paper.
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

29/1/02

10:15 am

Page 15

WILSON: SOCIAL CLASS AND SPORTS INVOLVEMENT

2.

3.

4.

15

In preliminary analyses, I also tried OLS regression and logistic regression, and results did not
differ substantively from those reported here. In the end, I chose MCA because in the GSS data
the relationship between income and sports involvement (particularly in the racing genre as
shown in Table 4) departs somewhat from linearity, making both OLS regression and logistic
regression inappropriate. Additionally OLS regression is, strictly speaking, inappropriate for this
studys dichotomous dependent variables.
Race is a dichotomy coded for nonwhite. Age is at respondents last birthday. Region is a series
of four dummies, respectively coded for east, midwest, south, and west, based on US Census
divisions. Community size is based on the GSS variable XNORCSIZ recoded 1 = open country;
2 = unincorporated area < 2500; 3 = town or village of 25009999; 4 = small city of
10,00049,999, 5= SMSA where central city is 50,000250,000; 6= SMSA where central city is
over 250,000.
This is not an artifact of the income variables coding. I repeated the Table 4 analyses trying
alternative income codings with as many as 21 categories and as few as three. Income was never
significantly related to mens attendance at auto and cycle races. In a single case for women,
with income coded 1 = < $30,000; 2 = $30,00075,000; 3 = $75,000+, there was a significant
positive bivariate relationship but none when demographic controls were introduced.

References
Aveni, A.F. (1976) Alternative Stratification Systems: The Case of Interpersonal Respect among
Leisure Participants, Sociological Quarterly 17: 5364.
Booth, D. and Loy, J. (1999) Sport, Status, and Style, Sport History Review 30: 126.
Bourdieu, P. (1978) Sport and Social Class, Social Science Information 17: 81940.
Bourdieu, P. (1984) Distinction: A Social Critique of the Judgement of Tast, trans. Richard Nice.
Cambridge, MA: Harvard University Press.
Bryson, B. (1996) Anything but Heavy Metal: Symbolic Exclusion and Musical Dislikes,
American Sociological Review 61: 88499.
Coakley, J.J. (1998) Sport in Society (6th edn). Boston, MA: Irwin McGraw-Hill.
Collins, R. (1979) The Credential Society: An Historical Sociology of Education and Stratification.
New York: Academic Press.
Curry, T.J. and Jiobu, R.M. (1984) Sports: A Social Perspective. Englewood Cliffs, NJ: Prentice-Hall.
Curtis, J.E. and Milton, B.G. (1976) Social Status and The Active Society: National Data on
Correlates of Leisure-Time Physical and Sport Activities, in R.S. Gruneau and J.G. Albinson
(eds) Canadian Sport: Sociological Perspectives, pp. 30329. Don Mills, Ontario: AddisonWesley.
Davis, J.A. and Smith, T.W. (1998) General Social Surveys, 19721998 [machine-readable data file].
Principal investigator, James A Davis; director and co-principal investigator, Tom W. Smith; coprincipal investigator, Peter V. Marsden. Chicago, IL: National Opinion Research Center.
DiMaggio, P. (1982) Cultural Capital and School Success: The Impact of Status Culture Participation
on the Grades of U.S. High School Students, American Sociological Review 47: 189201.
DiMaggio, P. (1987) Classification in Art, American Sociological Review 52: 44055.
DiMaggio, P. and Mohr, J. (1985) Cultural Capital, Educational Attainment and Marital Selection,
American Journal of Sociology 90: 123161.
DiMaggio, P. and Ostrower, F. (1990) Participation in the Arts by Black and White Americans,
Social Forces 68: 75378.
DiMaggio, P. and Useem, M. (1978) Social Class and Arts Consumption: The Origins and Consequences of Class Differences in Exposure to the Arts in America, Theory and Society 5:
14161.
Douglas, M. and Isherwood, B. (1979) The World of Goods. New York: Basic Books.
Eitzen, S.D. (1996) Classism in Sport: The Powerless Bear the Burden, Journal of Sport and Social
Issues 20: 95105.
Eitzen, S.D. and Sage, G.H. (1991) Sociology of North American Sport (5th edn). Madison, WI:
Brown & Benchmark.
Erickson, B.H. (1996) Culture, Class and Connections, American Journal of Sociology 102: 21751.
Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009

03_IRS 37/1 articles

16

29/1/02

10:15 am

Page 16

INTERNATIONAL REVIEW FOR THE SOCIOLOGY OF SPORT 37(1)

Holt, D.B. (1997) Distinction in America? Recovering Bourdieus Theory of Taste from its Critics,
Poetics 25: 93120.
Holt, D.B. (1998) Does Cultural Capital Structure American Consumption?, Journal of Consumer
Research 25: 125.
Hughes, M. and Peterson, R.A. (1983) Isolating Cultural Choice Patterns in the US Population,
American Behavioral Scientist 26: 45978.
Lamont, M. (1992) Money, Morals, and Manners: The Culture of the French and American UpperMiddle Class. Chicago, IL: University of Chicago Press.
Lamont, M. and Lareau, A. (1988) Cultural Capital: Allusions, Gaps and Glissandos in Recent
Theoretical Development, Sociological Theory 6: 15368.
Lamont, M., Schmalzbauer, J., Waller, M. and Weber, D. (1996) Cultural and Moral Boundaries in
the United States: Structural Position, Geographic Location, and Lifestyle Explanations,
Poetics 24: 3156.
Leonard, W.M., II (1998) A Sociological Perspective of Sports (5th edn). Boston, MA: Allyn &
Bacon.
Levine, L.W. (1988) Highbrow/Lowbrow: The Emergence of a Cultural Hierarchy in America.
Cambridge, MA: Harvard University Press.
Mandell, R.D. (1984) Sport: A Cultural History. New York: Columbia University Press.
Martin, T.W. and Berry, K.J. (1987) Competitive Sport in Post-Industrial Society: The Case of the
Motocross Racer, in Andrew Yiannakis, T.D. McIntire, M.J. Melnick and D.P. Hart (eds) Sport
Sociology: Contemporary Themes (3rd edn), pp. 26975. Dubuque, IA: Kendall/Hunt.
Nixon, H.L., II, and Frey, J.H. (1996) A Sociology of Sport. Belmont, CA: Wadsworth.
Peterson, R.A. (1997) The Rise and Fall of Highbrow Snobbery as a Status Marker, Poetics 25:
7592.
Peterson, R.A. and Kern, R.M. (1996) Changing Highbrow Taste: From Snob to Omnivore,
American Sociological Review 61: 9007.
Peterson, R.A and Simkus, A. (1992) How Musical Tastes Mark Occupational Status Groups, in M.
Lamont and M. Fournier (eds) Cultivating Differences: Symbolic Boundaries and the Making of
Inequality, pp. 15286. Chicago, IL: University of Chicago Press.
Rojek, Chris (2000) Leisure and the Rich Today: Veblens Thesis after a Century, Leisure Studies
19: 115.
Scholsberg, J. (1987) Who Watches Television Sports?, American Demographics 9: 459, 59.
Schor, J.B. (1991) The Overworked American: The Unexpected Decline of Leisure. New York: Basic
Books.
White, P. and Wilson, B. (1999) Distinctions in the Stands: An Investigation of Bourdieus
Habitus, Socioeconomic Status and Sport Spectatorship in Canada, International Review for
the Sociology of Sport 34: 24564.
Yergin, M.L. (1986) Who Goes to the Game?, American Demographics 8: 423.
Young, M. and Willmott, P. (1973) The Symmetrical Family: A Study of Work and Leisure in the
London Region. London: Routledge & Kegan Paul.

Thomas C. Wilson is Professor of Sociology, Florida Atlantic University in

Boca Raton, Florida. His current research focuses on class-based differences in
culture consumption as well as class- and race-based determinants of premarital
fertility patterns. His most recent work has been published in Social Forces and
Sociological Perspectives.
Address: Florida Atlantic University, Boca Raton, FL 33431, USA.
Email: [email protected]

Downloaded from http://irs.sagepub.com at GLASGOW CALEDONIAN UNIVERSITY on October 5, 2009