Académique Documents
Professionnel Documents
Culture Documents
VOLUME 23
NUMBER 2
2007
123137
123
Introduction
In this article we contribute to the literature on
cultural omnivorousness by introducing to the discussion a new but related dimension of leisure consumption. Since the concept of omnivorousness was first
defined by Peterson and Simkus (1992), research on it
has rested theoretically on the way cultural capital
involves an appreciation of a wide range of cultural
forms, including the fine arts, popular culture, and folk
culture. Accordingly, measures of omnivorousness have
been primarily based on the breadth of cultural tastes.
However, another important dimension of cultural
consumption relates to the frequency of participation
in relevant leisure activities. While there is a growing
literature on omnivorousness within the sociology of
culture, there has been no previous research that
attempts to theoretically locate the frequency of leisure
participation in relation to cultural omnivorousness.
The addition of the dimension of cultural participation
The Author 2006. Published by Oxford University Press. All rights reserved.
For permissions, please e-mail: journals.permissions@oxfordjournals.org
124
Revisiting Omnivorousness
The general goal of the research described here was to
problematize, revisit, and modify the contours of the
omnivore concept, through complementing it with a
measure based on both the range and the frequency of
cultural participation. Our reconsideration of omnivorousness stems from three sources. First, Bourdieus
(1984) emphasis on the differences in the way
individuals consume rather than on what they
consume led us to look more closely at the (voracious)
pattern of consumption in addition to the (omnivorous) contents of consumption. This emphasis stems
from a perspective that sees the expression of cultural
capital as embodied in consumer actions (Holt, 1997).
Holt asserts that research on the association between
cultural capital and social divisions should rely less
on cultural contents and more on consumption
practices. This is because the content of objectified
cultural capital has become a relatively weak exclusionary mechanism due to increasingly blurred cultural
hierarchies in contemporary societies.
Campbell (1987) had earlier linked the idea of the
insatiable consumer to the unique dynamic character
of modern-day consumption. High levels of individual consumption in contemporary developed societies
stem from the insatiability of consumers (the fact that
their wants appear never to be exhaustible). Hedonistic
consumers are unlimited in their appetites for intense
emotional experiences (Bell, 1979) and their focus of
desire is the experience the consumer anticipates of the
act of consumption (McCracken, 1990). If this desire
were provided not by leisure participation in itself
but by anticipation of a novel leisure activity, we
would expect modern consumerism to involve a high
turnover of goods and experiences, or, in our terms,
a voracious pattern of cultural participation. In a
different but related account of cultural behaviour,
Schulze (1992) offers a perspective in which cultural
repertoires are characterized by the aesthetics of
excitement (Spannungschema). For Schulze, it is selfrealization that drives the cultural preferences of the
affluent middle class, who engage in various highbrow,
125
126
Method
The Data
Our data comprised the first wave of Home OnLine,
a panel study undertaken by the Institute for Social
and Economic Research at the University of Essex,
England, of adult individuals in households in Britain.2
The first wave was conducted between October and
December (inclusive) of 1998, and was selected according to a qualified form of randomization, which
ensures inclusion of geographically clustered areas
with representation of different social strata similar
to that of the population. Selection of households was
The Variables
The first advantage of this data set for our purposes
was that it contained information on participation
in leisure activities both from questionnaire and
time use diary sources, making possible a choice
and a comparison of measures of voraciousness.
Questionnaire measures of activity participation are
the most familiar means of data-collection on leisure
participation in the area of cultural consumption,
although time use measures are becoming increasingly applied (e.g. Garhammer, 1998; Robinson and
Godbey, 1999). Therefore, it was important to be
able to compare the new diary source with the more
standard measures. It has been shown that diary
estimates of time spent in different activities, where
people record their participation in those activities
with at least some degree of contemporaneity (i.e., in
their diaries), differ from estimates based upon
responses to retrospective questions. It can safely be
assumed that diary estimates are in fact the more
accurate, since they do not involve the same problems
127
128
Independent Variables
Results
Comparison of Measures
The first step was to compare the time use and
questionnaire measures of voraciousness. There has
been some criticism in the literature of time use diaries
as data collection instruments in the area of free time
and leisure activities (see the Round Table Discussion
on non-response bias in time use surveys in Leisure &
Society, 1998). One of the most common arguments
is that, while time use diaries may be able to accurately
record activities which are done frequently and regularly, they may be unreliable when it comes to less
frequently performed activities. The use of a weeklong
diary, however, seems to go a long way in meeting
these objections. In Table 1, the first questionnaire
measure of voraciousness is based on the number of
activities reported from the survey as being done at
least once a week. With a weekly diary one would
expect on the average to record all these activities. The
second questionnaire measure also contains those
activities that are reported on the survey question
as being done at least once a month. On average
one-quarter of these activities will appear in a weekly
diary. Assuming that the diary gives an accurate record
of activity participation we would therefore expect
the diary estimate to have a higher distribution and
mean than the first questionnaire measure, but a lower
Table 1 Distributions and means of the time use diary and questionnaire measures of voraciousness
Number of different
out-of-home leisure
activities
Questionnaire
measure 1 N (%)
Weekly diary
measure N (%)
Questionnaire
measure 2 N (%)
None
One
Two
Three
Four
Five
N = 100%
Mean
Correlation coefficient
(Spearmans rho) with
weekly diary measure
135 (19.7)
304 (44.2)
186 (27.1)
60 (8.7)
2 (0.3)
1 (0.1)
688y
1.26
103 (15.0)
232 (33.7)
227 (32.9)
102 (14.8)
24 (3.5)
688y
1.58
57 (8.3)
179 (26.0)
247 (35.8)
150 (21.8)
49 (7.1)
7 (1.0)
688y
1.96
0.363
0.350
The analyses are based on a sample of 1317 respondents. However, the total N in this and all subsequent analyses in the article are a result of
the sample weighting which corrects for both the original over-sampling of households with a personal computer and for patterns of non-response.
Source: Home OnLine, Britain, 1998 (first wave).
High management
Low management
Intermediate
Small employer
Low supervisor
Semi-routine
Routine
(Undefined)
Total
Mean (N)
1.79
1.83
1.77
1.56
1.49
1.39
1.35
(1.58)
1.64
(50)
(136)
(80)
(43)
(52)
(56)
(54)
(20)
(491)
Highest qualification
Mean (N)
Newspaper
read
Degree, nursing
A-levely, higher vocational
GCSEy, lower vocational
Failed GCSEy, youth training
1.89
1.63
1.59
1.47
Quality
Medium
Tabloid
(None)
Total
1.30 (167)
1.58 (696)
(141)
(122)
(241)
(26)
(None)
Total
Mean (N)
2.00 (81)
1.66 (161)
1.38 (176)
(1.55) (278)
1.58 (696)
129
130
131
Table 3 Multiple analysis of variance models showing effect of various variables on the number of different
out-of-home leisure activities per week
Independent variables
Covariate hours
worked per week
(employees only)
Covariate total
leisure time
P 0.000, beta0.20
1.78
1.86
1.62
1.28
P 0.001, beta 0.26
ns
ns
0.11
ns (beta 0.13)
1.81
1.77
1.57
1.30
P 0.002, beta0.20
ns
ns
0.08
P 0.033, beta0.14
1.83
1.74
1.55
1.33
ns
ns
P 0.000
0.12
P 0.000, beta0.20
1.82
1.64
1.55
1.48
P 0.008, beta0.23
ns
ns
0.07
NS (beta 0.12)
1.85
1.64
1.58
1.44
P 0.02, beta0.17
ns
ns
0.06
P 0.000, beta0.20
1.89
1.65
1.59
1.30
ns
ns
P 0.000
0.09
P 0.000, beta0.17
1.82
1.73
1.48
1.60
P 0.004, beta0.24
ns
ns
0.07
2.06
1.70
1.39
1.60
P 0.004, beta0.19
ns
ns
0.08
1.98
1.68
1.38
1.54
P 0.001, beta0.17
ns
P 0.000
0.09
Covariate net
monthly income
(employees only)
132
Cross-Time Comparisons
Two national-level British data-sets from the MTUS
archive: the 1975 BBC and the 1987 SCELI surveys11
(both of which included a week-long diary element)
permitted us to conduct a cross-time comparison to
see if our measure of voraciousness displayed the same
relationship with measures of human and economic
capital over time. As in any cross-time comparison
there were some questions of data compatibility to be
overcome. First, since the activity categories in the
harmonized MTUS data-sets are slightly different
from those used in the Home OnLine survey, the
voraciousness variable (number of different outof-home leisure activities participated in during the
week) was calculated slightly differently, on the basis
of a more detailed list of such activities. Secondly,
one of the problems of the MTUS archive is known
to be the relative lack of variables reflecting aspects
of human capital and social status, which confined us
to level of education, and a 3-category income variable
(for 1987 only). However, with these limitations in
mind, it was nevertheless possible to construct effectively the same dependent variable, and to test
its relationship with these aspects of human and
economic capital in multiple analyses of variance in
the same way as for the Home OnLine data.
To summarize the findings, we do see from the
earlier data some of the same effects as in the Home
OnLine data regarding the direction of the overall
relationship between voraciousness and measures of
human and economic capital.12 And in an analysis
of covariance equivalent to that shown in Table 3
(i.e. with education level, sex, and family status as
the independent variables, and time in paid work and
time in leisure as covariates in two separate models)
the only difference between the results from the Home
OnLine data and those from the earlier surveys
was the clear inverted J-shape of the relationship in
the earlier surveys between educational level
Discussion
133
134
Notes
1. Feldman and Hornik (1981) coined the term
timestyle to suggest that the choice of daily
activities constitutes the locus of the meaning of
time to individuals.
2. The original data collection was funded by British
Telecommunications plc.
3. For further details of the MTUS see Gershuny
et al., 2000.
4. We also derived a randomly selected sample of one
adult individual per household, and all analyses
were performed for both these samples. The results
for the individual sample were identical to those
for the sample of individual and couple households; given the greater numbers in the latter, we
present here only those results.
5. The response categories were: most days; at
least once a week; at least once a month;
6.
7.
8.
9.
10.
11.
12.
13.
References
Bell, D. (1979). The Cultural Contradictions of
Capitalism. London: Heinemann.
Bihagen, E. and Katz-Gerro, T. (2000). Culture
Consumption in Sweden: The Stability of Gender
Differences. Poetics, 27, 327349.
Bittman, M. (2002). Social Participation and Family
Welfare: The Money and Time Costs in Leisure in
Australia. Social Policy and Administration, 36,
408425.
Bittman, M. and Wajcman, J. (2000). The Rush Hour:
The Character of Leisure Time and Gender Equity.
Social Forces, 79(1), 165189.
135
136
Authors Addresses
Oriel Sullivan (to whom correspondence should be
addressed), Department of Behavioural Sciences,
Ben-Gurion University, Beer Sheva 84105, Israel.
Email: sullivan@bgu.ac.il
137
Appendix 1 Net pay, hours of work and leisure time by measures of human, economic, and cultural capital
Usual work
hours/week (N)
Total leisure
time/week (N)
1,314 (81)
1,241 (60)
842 (104)
842 (57)
302yy
0.000
40.9 (114)
39.9 (93)
36.4 (175)
34.9 (91)
472y
0.014
33.2 (139)
33.7 (119)
35.0 (239)
41.2 (190)
688
0.000
1,327 (115)
817 (54)
1,189 (50)
666 (72)
291
0.000
41.7 (181)
32.4 (79)
43.1 (93)
30.9 (102)
454
0.000
31.5 (184)
33.7 (79)
33.0 (95)
34.2 (109)
467
0.312
1,591 (41)
841 (65)
900 (83)
1,079 (113)
302
0.000
44.4 (61)
36.1 (103)
36.7 (114)
37.4 (193)
472
0.005
38.7 (79)
38.7 (158)
36.6 (174)
33.6 (275)
687
0.002
Educational level:
Degree, nursing qualification
A-level, higher vocational
GCSE, lower vocational
None
N
Statistical significance from one-way
analysis of variance
Social status of job:
Management
Intermediate
Self-employed, low supervisors
Semi-skilled, unskilled routine
N
Statistical significance from one-way
analysis of variance
Type of newspaper read:
Quality
Medium
Tabloid
None
N
Statistical significance from one-way
analysis of variance