Sport as culture: Social class,styles of cultural consumption and sports participation in Canada

Abstract

Cultural consumption writ large has had a prominent place in the sociological discipline since Pierre Bourdieu. While Bourdieu often considered sport in analyses of culture, there have since been relatively few studies that are focused on considering sport within the broader landscape of cultural consumption. This paper seeks to assess the place of sports participation within the cultural lifestyles of Canadians. To this end, this paper employs multiple latent class analyses of various cultural and sporting variables from a large-scale Canadian government survey. It also employs regression analyses of those latent class groups. The results show three primary groups of consumers, pointing strongly to confirming the omnivore thesis in Canada. The results of the core of the analysis show, however, strong delineations in which sports these different groups consume. Ultimately then, the cultural domain of sports may be an area where omnivores practise more distinctive consumption, eschewing the sports of other consumer groups.

Keywords

Bourdieu cultural consumption latent class analysis omnivore thesis social class sport participation

Introduction

This paper examines the place of direct sports participation within the cultural landscape of Canada. It engages three different theories of cultural consumption in their application to culture in Canada, with particular emphasis on sports participation. Along with Bourdieu’s (1978, 1984, 1986, Bourdieu and Wacquant, 1992) theories of habitus and homologies, this paper tests Peterson’s (1992, 1997; Peterson and Kern, 1996) omnivore–univore framework and individualisation theses of cultural consumption. With respect to these three theories, particularly considered all at once, there is little work done on sports. There exists much more work in other cultural domains (e.g. music). Even as Bourdieu frequently considered sport, few scholars have studied sport as a cultural form that can be consumed in the same way. Since Bourdieu, the tendency of scholars has been to treat culture and sport relatively separately, without combining them in a broader analysis of culture. These dynamics are especially so within the social context of Canada and North America more broadly. The North American context is especially important here because of the cultural consumption theories that have arisen from it (e.g. Peterson’s omnivore). There is also a breadth to the sporting landscape of North American that is of a different nature to other areas of the world. Despite the prominent role that sport plays in Canadian, and other contemporary societies, research is generally absent when it comes to the relationship between social consumption patterns of sport and other cultural activities. This paper is thus concerned with how participatory sport operates as another domain of culture generally. Specifically, it seeks to further understanding of how sport intersects with other forms of culture and fits into typologies of cultural consumption patterning. This work seeks to fill these gaps in the existing literature, provide innovative applications for sports studies that bring in more traditional sociological analyses of cultural consumption and provide an additional national context-specific case study for present and future cross-national comparison. Towards these ambitions, this work asks the following guiding research questions.

What typologies of cultural consumption exist in Canada and which theory (or theories) of cultural consumption are most relevant to the social patterning of these types?

Where does direct sports participation fit into these typologies of consumption and broader patterning of cultural consumption? What is their relationship?

Theories of cultural consumption

Bourdieu: Habitus and homologies

Bourdieu argued that the reification of cultural choices is a form of capital. This cultural capital is then used to symbolically and structurally reinforce social (class) position. It can also be exchanged in ways that accrue other forms of capital. The socialised ways of naturally appreciating and understanding consecrated forms of culture are what Bourdieu characterises as embodied cultural capital (Bourdieu, 1986). Objectified cultural capital is a manifestation of this embodied form. This happens through a process by which the privileged objectify their cultural knowledge. They convert their cultural capital and cultural knowledge into implicit and explicit tastes and consumption patterns of legitimate and exclusive forms of culture. The processes of social exclusion that are based upon cultural choices and styles of appreciation describe the concept from which Bourdieu’s seminal work draws its title, Distinction (Bourdieu, 1984). Distinction is both the produced outcome of and the manifestation of these judgements of cultural difference. The difference in cultural engagement and appreciation between various classes of society form this distinction and is in contrast, and in negative referent, to the cultural choices and aesthetic profiles of other classes (Bourdieu, 1984; Swartz, 1997; Veenstra, 2015). Bourdieu (1984: 49) describes the internal logics of the process of distinction this way:

Aesthetic intolerance can be terribly violent. Aversion to different life-styles is perhaps one of the strongest barriers between the classes; class endogamy is evidence of this. The most intolerable thing for those who regard themselves as the possessors of legitimate culture is the sacrilegious reuniting of tastes which taste dictates shall be separated.

Bourdieu suggests here that these internal logics form distinctions on the basis of cultural tastes and engagements. It arises out of the two distinct but inexorably linked parts of cultural capital. There were thus stark social distinctions based upon the type of culture that one engaged (objectified cultural capital) and the sophistication with which one could innately discuss and appreciate these highbrow art forms (embodied cultural capital).

Bourdieu and sport

The internal mechanisms by which cultural choices are structured are a series of class-based primary and secondary socialised constitutions. They are structurally patterned by one’s ‘habitus’. Bourdieu’s habitus is the class-based dispositions of the individual (Bourdieu, 1984). This habitus is what structures cultural taste and social action. Because cultural choices are informed and embedded by this habitus, homologies of taste across the cultural spectrum follow from these class-based constitutions. This includes the domain of sport.

With regards to sports, Bourdieu (1978, 1984) argued that this theoretical framework extended to this cultural domain through the mechanism of this class-based habitus and the pursuant homology of taste. There are numerous studies that support Bourdieu’s class-based theories as they pertain to sports participation. Indeed, many support the conclusion that participation in sports is stratified by social position (Bourdieu, 1978; Collins, 2003, 2014; Lamprecht and Stamm, 1996; Scheerder et al., 2002, 2005; Sugden and Tomlinson, 2000; Taks et al., 1995). For instance, Bourdieu (1978) asserts the importance of the upper class origins of sport that he saw manifestly present in both British and French sport during his time. Collins (2003, 2014) outlines the strong exclusionary effect that poverty and a lack of economic resources have for sports participation, highlighting the persistence of class-based difference in this sports engagement and leisure engagement more broadly. Of special importance to the national context of this paper, Gruneau (1975, 1999) explains how the social class histories of Canadian sports specifically entrench class-based sporting inequalities in the Canadian context.

In accordance with Bourdieu’s theories, some studies find that personal exercise acts as a social marker of cultural capital (Shilling, 1993; Stempel, 2005; Wilson, 2002). Wilson (2002) performed one of the most influential studies regarding cultural capital and sports engagement, positing a ‘paradox’ of social class and sports involvement. This paradox was that while those with higher levels of both economic and cultural capital were more likely to be involved in sports generally, those with high levels of cultural capital were unlikely to participate in what he termed ‘prole’ sports (Wilson, 2002). These are so-called because they are those sports that have become associated with the lower and working classes (Wilson, 2002: 5). Wilson argues that ‘class-based differences in economic capital enable upper class involvement in expensive sports, leaving “prole” sports largely relegated to the lower classes’ (Wilson, 2002: 6). This is consistent with Bourdieu’s assertions of cultural capital and how it manifests through objectified cultural consumption. Wilson also asserts that because those with higher capital possession were more likely to participate in most sports, this is also not inconsistent with Peterson’s omnivore thesis, and further adaptations thereof.

The cultural omnivore

In what was seen as a North American update to Bourdieu’s theorisations of the relationship between social class and cultural consumption, American sociologist Richard Peterson and colleagues developed the ‘omnivore thesis’ (Peterson, 1992, 1997; Peterson and Kern, 1996; Peterson and Simkus, 1992). The omnivore in this framework is one who consumes a broad range of culture, spanning the traditional highbrow/lowbrow divide of Bourdieu’s conceptualisations. Omnivorous consumption, however, is not the paradigm of the majority of cultural consumers for Peterson. Rather, it is high status individuals that are more omnivorous (Peterson, 1992; Peterson and Kern, 1996). Therefore, there is still a social class divide in the patterning of cultural consumption. Those with lower social status and capital resources were observed to be ‘univores’, possessing a taste pattern that only consumed one (lowbrow) form of, in this case, music (Peterson, 1992). Therefore, while the consumption profiles of the higher and lower classes in the USA as observed by Peterson and in France as observed by Bourdieu were different, the patterning of cultural consumption by social class is consistent with both theories.

Many studies since Peterson have confirmed these findings. They find that omnivorous cultural consumption maps strongly onto socio-economic difference (Bryson, 1996; Chan and Goldthorpe, 2007; Peterson and Kern, 1996; Peterson and Simkus, 1992; Sintas and Alvarez, 2002, 2004; Sullivan and Katz-Gerro, 2007; Tampubolon, 2008). However, using musical taste, Bryson (1996: 897) suggests that the openness to a wide range of culture that omnivorism represents may ultimately also suggest that ‘tolerance itself may separate high-status culture from other group cultures’.

Other studies more explicitly highlight the compatibility of the omnivore with Bourdieu. Indeed, many argue that the cultural omnivore can also be a manifestation Bourdieu’s habitus and homology theories (Bennett et al., 2009; Coulangeon and Lemel, 2009; Leguina, 2015; Lizardo and Skiles, 2012; Tampubolon, 2008). Lizardo and Skiles (2012) argue that omnivorousness is ultimately a manifestation of Bourdieu’s aesthetic disposition. This is similar to the argumentation of Friedman (2011), who contends that Bourdieu’s aesthetic disposition can be extended to lowbrow or popular cultural forms. Friedman (2011: 351) argues that Bourdieu’s arguments for the possession of embodied cultural capital among the socially privileged provides a way by which they can consume more popular forms of culture in a more exclusive manner. Pursuant to this line of reasoning, this disinterested aesthetic disposition is transposable to all existing cultural domains and also to new cultural domains (Lizardo and Skiles, 2012). Therefore, it is this disinterested aesthetic that is the common thread between Bourdieu and Peterson. Rather than homologies of taste per se, it is thus more of a homology of application of the disinterested aesthetic that distinguishes omnivorousness and ties it to the theories of Bourdieu.

Omnivorousness, then, may not eliminate distinctive judgements of consumption. Rather, omnivorism may merely reconfigure these judgements to draw boundaries fully within each cultural domain. Therefore, it is more about how one consumes the cultural form, regardless of its general social status (Bennett et al., 2009; Friedman, 2011; Jarness, 2015; Savage et al., 2015). These kinds of dynamics continue to perpetuate Bourdieusian conceptions of ‘distinction’ within otherwise omnivorous cultural profiles. Because the data of this analysis is firstly quantitative and, secondly, not as differentiated at the level of taste within each activity, much of this more qualitative analysis of styles of consumption is beyond the scope of this paper. Bourdieu’s disinterested aesthetic has allowed the literature on cultural consumption a significant pathway into a Bourdieusian understanding of how his theories may be reconciled with, or at least nuance, the omnivore thesis and the existence of emerging forms of culture. Other scholars suggest that other more cultural mechanisms, such as ‘authenticity’ and ‘exoticism’ (Johnston and Baumann, 2007), or ‘openness’ (Ollivier, 2008), serve to perpetuate forms of distinction in more omnivorous styles of consumption.

Few studies have assessed any kind of omnivore concept with respect to sport as a form of cultural consumption and the sporting omnivore remains under-researched (Widdop et al., 2014). Some have focused on omnivorism in the domain of sport (Gemar, 2019; Widdop and Cutts, 2013; Widdop et al., 2014; Wilson, 2002), but these studies did not fully account for a varied range of cultural engagement or assess the place of sport within wider cultural lifestyles. This paper is thus concerned with more comprehensively understanding how sports participation is connected to other cultural domains.

Individualisation theses of cultural consumption

Yet another set of theories of cultural consumption argue that both Bourdieu and Peterson’s class-based theories are either not as powerful as many sociologists asserted, or have waning relevance in most modern societies. In terms of cultural consumption, these theories all posit a consumption paradigm that is individually driven, rather than patterned by more structural forces. Theories of cultural consumption that downplay the effect of social class in the lifestyle decisions of persons (e.g. Beck, 1992; Giddens, 1991) have been termed ‘individualisation’ theories (Chan and Goldthorpe, 2010; Chan and Goldthorpe, 2007; Gerhards et al., 2013). These theories proceed from the assertion that the cohesion of social class structures is, and has been, in decline, primarily as a function of increased societal wealth and mobility in modern societies (Atkinson, 2010; Bauman, 2002; Beck, 1992; Chan and Goldthorpe, 2010; Giddens, 1991). This decline is argued to be a product of modernisation (Gerhards et al., 2013). Therefore, the influence of social class structures in structuring the cultural consumption of individuals within the society is argued to be not nearly as pronounced as other theories suggest. Rather, Chan and Goldthorpe (2010: 6) describes that:

Rising standards of living, greater geographical and social mobility and exogamy and a growing awareness of alternative social bases of identity – for example, gender, ethnicity or sexuality – all help to free individuals from class constraints and status preoccupations and allow them to develop their own lifestyles as a matter of personal choice.

These theories as they apply to cultural consumption can be seen ‘if not as a more or less direct contradiction of the homology argument, then at all events as an attempt to restrict the validity of that argument to the past’ (Chan and Goldthorpe, 2007: 2). With respect to our testing of these three theories, if we find that the composition of the cultural consumption profiles found in this paper are not structured by the socio-economic variables, then this would suggest that theories of individualisation may hold sway in the patterning (or not) of cultural behaviour in Canada. Individualisation theories are thus ultimately based upon the waning influence of class in the structuring of cultural choices. However, Giddens (1991) does suggest that certain elements of identity can themselves be utilised as a form of cultural capital in producing distinctive consumption that mirrors and reinforces social class divisions.

Canadian context

Research on cultural consumption in Canada finds somewhat mixed results. Veenstra (2010) confirms social class differences in inter-domain consumption. Veenstra (2010) suggests that the omnivore theory might also be relevant to the Canadian context. This conclusion comes because almost all cultural activities in Canada were found to be associated with higher class persons. This does suggest omnivorousness, but it is not conclusive. Indeed, in an intra-domain investigation of musical tastes in English-speaking urban Canada, Veenstra (2015) found little evidence of omnivorism and did not confirm this theory. Through qualitative interviews, Ollivier (2008) finds distinctively cultural ways that omnivores create distinction, echoing the arguments that omnivorism may simply reconfigure the more structural bases of consumption found by Bourdieu.

Specifically regarding sports in Canada, both Erickson (1996) and Veenstra (2007) downplay class differences in terms of sports knowledge. Indeed, they suggest that both Toronto (Erickson, 1996) and British Columbia (Veenstra, 2007) are mostly classless in this aspect of sports consumption. In terms of attendance at sporting contests, however, White and Wilson (1999) find strongly classed dynamics, with those of higher economic and cultural capital more likely to attend sporting events. Finally, an omnivorous pattern of consumption has been found in the following of professional sports in Canada (Gemar, 2019). However, this professional sports omnivore was not associated with higher socio-economic status (Gemar, 2019). Rather, those who had distinctive sports that they were most likely to follow (or not) showed the greatest levels of social differentiation (Gemar, 2019). While Veenstra (2010) included a small handful of sports-related variables in his otherwise comprehensive analysis of cultural consumption in Canada, a comprehensive analysis of sports participation from a cultural consumption perspective is still absent in this context (and many others). This paper seeks to address this gap in the literature.

Data and methods

The Time Use Survey of Canada’s General Social Survey (GSS) is conducted every five years. However, the 2005 data set is the most recent survey year to ask questions of both cultural and sporting participation to the same subset of the respondents (n = 9747). Therefore, this is the most up to date data from Canada that can be used to analyse the place of sports within wider consumption profiles. One limitation to this data is its age, being now more than a decade old. However, because of the structure of the GSS Time Use Surveys, this 2005 survey is the most up to date survey that includes enough information to answer the core research questions of this analysis. The 2005 GSS cycle is also the largest sample that they ever collected for the Time Use section, and therefore it is still a good and useful measure of participation and provides valuable insights for the relationship between sports participation and broader cultural engagement. Indeed, other recent studies have utilised data from 2005 to make a scholarly contribution to the sociology of culture (Gemar, 2019; Cutts and Widdop, 2017). Table 1 shows the make-up of the GSS sample. The survey data regrettably does not include information on race. Table 2 shows the relative frequencies of the various cultural activity variables in this analysis. Table 3 shows the relative frequencies of the various sporting activity variables of this survey. Both tables also report the Cramer’s V measures of association between the operationalised measures of cultural and economic capital of this analysis. The results of these association measures are discussed in the results section of this paper.

Table 1.

Relative frequencies of socio-economic and demographic variables.

Variable	GSS
Personal education
Some secondary/elementary/none	19.4%
High school diploma	16.1%
Some uni/comm college	13.2%
Diploma/comm/tech college	27.4%
Bachelor’s degree	17.0%
Grad/professional School	6.9%
Household income (C$/year)
<20,000	13.6%
20–39,999	23.3%
40–59,999	21.8%
60–79,999	15.2%
80–99,999	10.0%
>100,000	16.1%
Age group
25–44	39.5%
45–64	39.0%
65+	21.5%
Sex
Female	57.3%
Male	42.7%
Region
Quebec	19.5%
Atlantic Region	20.0%
Ontario	28.5%
Prairie/Western Provinces	31.9%
Population centre
Larger/urban population centres	76.7%
Smaller/rural population centres	23.3%

GSS: General Social Survey.

Table 2.

Relative frequencies and economic and cultural capital Cramer’s V scores for cultural variables.

	Frequency	Education	Household income
Read magazines	69.5%	.200***	.184***
Read newspapers	66.6%	.066***	.130***
Watch videos	55.2%	.204***	.205***
Read books	41.3%	.195***	.036
Go to historic sites	34.5%	.244***	.204***
Go to zoo/aquarium^a	31.4%	.232***	.190***
Go to movies	25.1%	.229***	.188***
Cultural/art festival	22.4%	.254***	.147***
Theatrical performance	13.9%	.121***	.083**
Pop music performance	9.2%	.076**	.027
Classical/symphony perf.	4.8%	.120***	.036
Public art gallery	4.2%	.161***	.074***
Other museums	2.5%	.098***	.075***

p < .05; **p < .01; ***p < .001.

Also includes botanical garden, planetarium or observatory.

Table 3.

Relative frequencies and economic and cultural capital Cramer’s V scores for sports participation variables.

	Frequency	Education	Household income
	Frequency	(cultural capital)	(economic capital)
Golf	5.8%	.083***	.144***
Ice Hockey	3.1%	.070***	.134***
Swimming	2.4%	.072***	.033
Cycling	1.7%	.058***	.047*
Downhill skiing	1.6%	.114***	.132***
Curling	1.5%	.022	.055**
Tennis	1.3%	.089***	.058***
Baseball	1.3%	.030	.055**
Soccer	1.1%	.061***	.062***
Volleyball	1.1%	.068***	.049**
X-country skiing	0.9%	.070***	.053**
Softball	0.9%	.032	.047*
Basketball	0.8%	.042*	.035
Ten-pin bowling	0.8%	.050**	.042*
Badminton	0.6%	.050**	.033
Five-pin bowling	0.6%	.036	.027
Squash	0.4%	.074***	.059***
Equestrian	0.3%	.033	.036
Football	0.3%	.024	.027
Sailing/yachting	0.2%	.037*	.065***
Gymnastics	0.2%	.024	.019
Martial Arts	0.2%	.025	.017
Rugby	0.1%	.030	.030
Weightlifting	0.2%	.034	.042*
Boxing	0.1%	.035	.036
Field hockey	0.1%	.039*	.043*
Archery	0.1%	.034	.030
Rowing	0.1%	.029	.024
Track and field	0.1%	.028	.015
Triathlon	0.1%	.058***	.045*
Water skiing	0.1%	.029	.049**
Figure skating	0.1%	.032	.024
Lacrosse	0.1%	.031	.031

p < .05; **p < .01; ***p < .001.

Regarding sports participation, the GSS very particularly outlines the parameters of its question. The survey asks, ‘Did you regularly participate in any sports during the past 12 months?’. In the survey the term ‘regularly’ is defined by stipulating that ‘regularly means at least once a week during the season or for a certain period of the year’ (Statistics Canada, 2012). It also stipulates what is to be included in the range of sports of this question. The survey instructs respondents to exclude ‘non-competitive aerobics, aquafit, bicycling for recreation/transportation only, body building/body sculpting, car racing, dancing, fishing, fitness classes, hiking, jogging, lifting weight (non-competitive), motorcycling, snowmobiling and non-competitive walking’ (Statistics Canada, 2012). Therefore, some personal fitness activities, such as jogging, are not included in the data and thus cannot be included in this analysis. The survey thus excludes some unstructured forms of physical activity. This is especially disappointing, given many of the social class connotations of those types of personal fitness activities (Bennett et al., 2009; Stempel, 2005). However, the list of over 30 sports that is able to be included in this paper does serve as a substantial, although not entirely comprehensive, introduction to the relationship between sporting activity and broader cultural and leisure lifestyles.

The specific questions for cultural participation inquire about a time frame of the last 12 months. For the various cultural categories, the questions ask, ‘During the past 12 months, (as a leisure activity) did you…?’ (Statistics Canada, 2012). If the respondent answered in the negative, then it was marked down that they did not participate in that activity at all. If, however, the respondent answered in the affirmative, then the interviewer would follow up by asking ‘how often’. For each of the various activities, the two most frequent response choices for engagement were chosen. These response frequencies varied by the reasonably expected frequency of engagement, as determined by Statistics Canada for this survey year. For instance, the most frequent response option for reading a newspaper was ‘daily’ and for reading magazines was ‘at least once a week’. The top two frequency options are coded as regular participation for this study. This is because in the study of lifestyles, scholars should be focused on what people do more consistently with their cultural engagement. Therefore, this method is able to distinguish between those that go to a film or theatre production once or twice a year for a family member’s birthday and those who frequently engage in these cultural activities. This coding therefore also captures taste in a way that the former method of coding could not. However, ultimately a detailed outline of taste is not available in the cultural variables of this paper. This is because the questions do not differentiate between, for example, genres of film or popular music, some of which may be more distinguishably traditional highbrow forms.

This paper uses latent class analysis (LCA) to form clusters of cultural behaviour. From these clusters, patterns of consumption can be identified. This can aid in the distinctive identification of, for example, the cultural omnivore. However, because of the necessarily class-based conceptualisations of Bourdieu and Peterson, the modal latent class values of each respondent will be tested using binomial logistic regression. These are performed against a series of socio-economic and demographic variables. Regression methods are used here because of their ability to isolate and control for these variables and generate predictive measures. From this analysis, we can start to see which theories of cultural consumption are most supported by this analysis. In this paper, economic capital is operationalised by household income. Cultural capital is operationalised by personal education. To account for those who are likely not of an age to have obtained their highest educational qualification, the category of 15–24-year-olds has been excluded from this analysis. Table 3 shows the relative frequencies for the independent variables that are included in the regression analysis of this paper. We now move on to the results of the LCA.

Results

Capital association statistics for each cultural and sporting variable

Tables 2 and 3 followed the example of Veenstra (2010) in using Cramer’s V association statistics. These are used to assess the bivariate relationship between education, income and each of the cultural and sporting variables. The results help to illuminate the general social position of each variable. This will aid in the interpretation of the latent class groups, particularly surrounding any signs of omnivorism. For each of the 13 cultural variables, education is significantly associated with participation. Similarly, 10 of the 13 variables, with the exceptions of reading books and going to pop and classical music performances, are significantly associated with income levels. Eleven of the 13 activities are more strongly associated with education than income. Only reading the newspaper and watching videos (barely) were more strongly associated with income.

The sporting variables show a lower percentage of significant associations. This could be a function of these sports being legitimately less tied to economic and cultural capital, or it could be mostly a function of their relatively low frequency levels within the survey. Of the 34 sports, 21 show a statistically significant association with education, income or both. Sixteen show association with educational attainment. Eighteen show association with household income. Ten sports show a stronger association with education over income, while the other eleven sports show a stronger association with income over education. These measures show that the influence of cultural and economic capital is more similar for sports. Like the cultural variables, none of these sports showed an association that reflects a relationship where less education or income corresponds to a higher likelihood of participation.

Latent class analysis of cultural activity variables

The LCA of the data included in this section takes the 13 cultural activities from the GSS and seeks to find a latent class model that best fits the data. LCA is able to identify typological patterning in the data by which these latent classes are created. Table 4 shows the results of the latent class modelling for the data when one to six latent class solutions were specified as possibilities for model solutions. As the table shows, the first solution, the one latent class model, shows non-significant chi-squared p-values. This chi-squared p-value is used to determine which models show good fit with the data. However, the two, three, four, five and six latent class models all show statistically significant p-values. Therefore, all four of these models show good fit with the data. Because of this, these models have to be differentiated somehow so that one can be chosen for more comprehensive analysis. We must therefore compare these models against each other. For these more comparative measures we use an alternative measures of fit based upon the model’s log-likelihood statistics. This is in contrast to the p-value for each model, which is derived from the model’s chi-squared statistics. Of these five significant latent class solutions, the three-class solution best fits the data. This is because the Bayes Information Criterion (BIC) and Consistent Akaike Information Criterion (CAIC) reach their optimal point for this model solution. The BIC, specifically, is argued to be the most statistically rigorous measure (Asparouhov and Muthen, 2006). We therefore select this three-class model solution for the analysis of this paper.

Table 4.

Latent class analysis of cultural activities and behaviours.

		BIC(LL)	CAIC(LL)	N par	L²	df	p-value
Model1	1-Cass	43,009.0217	43,022.0217	13	4022.4919	3261	7.5e-19
Model2	2-Class	42,041.3474	42,068.3474	27	2941.5049	3247	1.00
Model3	3-Class	41,907.1258	41,948.1258	41	2693.9705	3233	1.00
Model4	4-Class	41,917.6680	41,972.6680	55	2591.2000	3219	1.00
Model5	5-Class	41,936.8902	42,005.8902	69	2497.1094	3205	1.00
Model6	6-Class	41,974.9927	42,057.9927	83	2421.8992	3191	1.00

BIC: Bayes Information Criterion; CAIC: Consistent Akaike Information Criterion.

The bolded latent class model is the one chosen for the analysis of this paper.

As seen in Table 5, this model divides the respondents from the GSS into three clusters, which constitute approximately 43%, 39% and 18% of the sample, respectively. The first cluster is distinguished by the high probability of people in this group to engage in what might be termed more ‘pop culture’ activities. The activities included in this cluster that have probabilities of inclusion that are higher than their overall relative probabilities are watching movies, watching videos, going to historic sites, reading magazines, reading newspapers, attending cultural/art festivals, going to zoos/aquariums/botanical gardens/planetariums and attending pop music performances. We label this group, then, the ‘pop culture’ cluster.

Table 5.

Latent class analysis profile.

Latent class	One	Two	Three	Overall relative frequency
Cluster size	0.4319	0.3869	0.1812
Indicators
Magazine	0.8388	0.6916	0.9031	0.695
Newspaper	0.7227	0.6476	0.8643	0.666
Videos	0.7797	0.4634	0.5672	0.552
Books	0.4926	0.3631	0.7643	0.413
Historic Sites	0.6164	0.2776	0.7718	0.345
Zoo/aquarium/etc.	0.5988	0.1994	0.5946	0.314
Movies	0.5481	0.1754	0.4953	0.251
Cultural/art festival	0.5021	0.1585	0.6189	0.224
Theatrical performances	0.0888	0.0674	0.4080	0.139
Pop music performances	0.1010	0.0381	0.1903	0.092
Symph/classical music perf	0.0071	0.0226	0.1993	0.048
Public art galleries	0.0309	0.0090	0.3605	0.042
Other museums	0.0168	0.0029	0.2049	0.025

Bolded values represent those who have probabilities higher than their overall relative frequencies for the sample.

Cluster two is defined by the fact that none of the cultural activity variables have probabilities for inclusion in this group that meet or exceed their overall relative frequencies in the sample. This cluster is thus determined to be an ‘inactive’ cluster. This group is distinguished by its relative inactivity and non-engagement across the cultural spectrum. It is the second largest group.

In cluster three we see the ‘omnivore’. As the table shows, this group has probabilities of inclusion that are higher than their overall relative frequencies for each of the cultural activities, with no exceptions. The fact that there is a distinct omnivorous group is important to note. We can now see if the make-up of this cluster is distinct from the other clusters, how it is so and if it is indeed associated with any specific patterning of sports participation.

The make-up of the cultural clusters

What, then, is the demographic and socio-economic make-up of these various latent classes? To analyse this make-up, binary logistic regressions are performed on the modal latent class values associated with each respondent. The results of these regressions can be found in Table 7. Table 6 reports the latent class probabilities for each of the socio-economic and demographic variables. Analysing these latent classes for these variables can further the analysis of which prevailing theories of sports consumption are most indicative of sports consumption in Canada. One of the primary findings of the first part of this paper is that an omnivorous consumption of culture exists in Canada. But what are the characteristics of this, and indeed the other latent classes from our LCA?

Table 6.

Latent class probabilities of socio-economic and demographic variables.

Variable	Pop culture	Inactive	Omnivore
Personal education
Some secondary/elementary/none	0.0566	0.1461	0.0353
High school diploma	0.1098	0.1640	0.0724
Some uni/comm college	0.1352	0.1425	0.1236
Diploma/comm/tech college	0.3029	0.3012	0.2599
Bachelor’s degree	0.2733	0.1754	0.3141
Grad/professional School	0.1213	0.0653	0.1917
Household Income (CAD/year)
<20,000	0.0480	0.0909	0.0539
20–39,999	0.1449	0.1707	0.1412
40–59,999	0.1755	0.1766	0.1518
60–79,999	0.1613	0.1243	0.1458
80–99,999	0.1194	0.0823	0.1027
>100,000	0.2050	0.1388	0.2364
Age group
25–44	0.5087	0.3654	0.3202
45–64	0.3869	0.4051	0.4581
65+	0.1044	0.2296	0.2217
Sex
Female	0.5835	0.6186	0.6361
Male	0.4165	0.3814	0.3639
Region
Quebec	0.2133	0.2515	0.1936
Atlantic Region	0.1842	0.1875	0.1586
Ontario	0.2939	0.2912	0.3162
Prairie/Western Provinces	0.3087	0.2698	0.3317
Population centre
Larger/urban population centres	0.8277	0.7535	0.8396
Smaller/rural population centres	0.1452	0.2130	0.1297

Conditional probabilities

Table 6 presents the conditional probabilities of memberships for each cluster according to the various demographic and socio-economic variables used in the analysis. The pop culture group contains individuals that are generally well educated and have elevated levels of household income. Individual members of this group also appear to come predominantly from the youngest two age categories, with the youngest age category being most prevalent in this cluster. This group also mostly lives in large urban population centres. Members of this group are also mostly female and live in Ontario and the West, although these probabilities are not overwhelming given the make-up of the sample.

The inactive cluster shares few characteristics with the pop culture cluster. Probabilities for individuals at the top and bottom of the education spectrum are reversed, with higher probabilities in the lower two educational categories and lower probabilities for the highest two categories. Similar dynamics can be seen in the income categories, although they are not as pronounced. This group is more balanced in terms of age. The conditional probabilities for the oldest age category are highest for this group. This group is more balanced for the type of population centre of individual members. It is also more probabilistically female than the cluster one, and more likely to include those from Quebec versus the Prairie and Western provinces.

The omnivore group extends the trends of the pop culture cluster for education and income. This group has the highest probabilities for individual respondents to be in the highest categories for education and income. It has the highest probability levels for individual members to be from the middle age category, and probabilities for the youngest and oldest categories just below the inactive cluster. This group also has the highest probabilities for females to be included in this group, the highest probabilities for those from Ontario and the West and the highest probability for including those from large urban centres. While this exercise is useful, it is not as statistically robust an analysis of the make-up of these clusters. To achieve a more rigorous statistical testing, this analysis performs binomial logistic regressions.

Binary logistic regression

As the reader can see in Table 7, and mirroring the probabilities of Table 6, there are many variables that present as significant parameters for predicting latent class membership. For the first cluster, the pop culture group, the significant parameters are educational attainment, age and population centre. Those in the highest educational category are almost twice as likely to be in this group than those in the lowest educational category. Those in large/urban population centres are almost one and a half times more likely to be in this group. Age, however, is the most substantial predictor for this group. Those in the youngest age group are approximately three times more likely to be in this cluster than those in the oldest age category.

Table 7.

Binary logistic regression odds ratios of socio-economic and demographic variables.

Variable	Pop culture	Inactive	Omnivore
Personal education
Some secondary/elementary/none	.531**	6.135***	.110***
High school diploma	.912	4.673***	.138***
Some uni/comm college	.737*	3.10***	.374***
Diploma/comm/tech college	.729	2.849***	.430***
Bachelor’s degree	.747	1.543*	.793
Grad/professional School	1.00	1.00	1.00
Household income (CAD/year)
<20,000	.704	1.754**	.709
20–39,999	.889	1.314	.798
40–59,999	.963	1.339*	.645*
60–79,999	1.161	.960	.835
80–99,999	1.261	.912	.765
>100,000	1.00	1.00	1.00
Age group
25–44	2.99***	.703*	.238***
45–64	1.82***	.800	.529***
65 +	1.00	1.00	1.00
Sex
Female	1.12	1.046	1.398**
Male	1.00	1.00	1.00
Region
Quebec	1.245	1.692***	.593**
Atlantic Region	.986	1.328*	.574**
Ontario	1.188	1.430**	.764
Prairie/Western Provinces	1.00	1.00	1.00
Population centre
Larger/urban population centres	1.40**	.594***	1.492*
Smaller/rural population centres	1.00	1.00	1.00
Negelkerke R square	.081	.134	.137

p < .05; **p < .01; ***p < .001.

While the first cluster had three significant predictive parameters, the second group, the inactive cluster, has five. These five are education, income, age, region and population centre. Those in the oldest age category are slightly more likely to be in this group, while those in small/rural population centres and the general east of the country are more likely to be in this group. Importantly for the class-based theories that this paper examines, the two operationalised measures of economic and cultural capital are the most predictive of membership in this class. Those in the lowest income category are almost twice as likely to be in this group than those with the highest levels of income. The most predictive, however, is educational attainment. Those with the highest levels of education are significantly less likely than each lower income category to be in this inactive group. Indeed, the lowest income category is more than six times more likely to be included in this cluster.

There are significant parameters in each of the variables for predicting membership in the final, omnivorous class. Like the first cluster, we see those in large/urban centres more likely to be members of this group. In the reverse of the second cluster, those in the Prairie and Western provinces are most likely to be members of the omnivorous group. The results for age are approximately the inverse of cluster one, with those in the oldest age group almost two and three times more likely to be in this group than those in the middle and youngest age categories, respectively. There is some suggestion that those in the highest income category are more likely to be in this group, although these results do not appear statistically compelling. The strongest predictive parameter of any of these regressions is the educational attainment for membership in the omnivore cluster. Those with the highest levels of educational attainment are more than nine times more likely to be members of this group.

The place of participatory sport within the broader cultural landscape

Now that we have observed where groupings exist in cultural consumption, we now answer our broader research question of where direct sports participation fits within people’s cultural lifestyles. To start to answer this question we will include each of the sports with our LCA model for the cultural activity variables that we performed earlier in this paper. After that, we will look at the regression results for these variables.The conditional probabilities for each of the sports are reported in Table 8. The odds ratios from the regression analysis for these variables can be seen in Table 9.

Table 8.

Latent class analysis cluster probabilities for each of the sports participation variables.

	Pop culture	Inactive	Omnivore	Overall
Ice hockey	0.0563 ^a	0.0345	0.0317	0.031
Basketball	0.0124	0.0080	0.0125	0.008
Soccer	0.0186	0.0069	0.0166	0.011
Volleyball	0.0246	0.0161	0.0165	0.011
Baseball	0.0182	0.0105	0.0118	0.013
Downhill skiing	0.0327	0.0209	0.0494	0.016
Tennis	0.0246	0.0126	0.0443	0.013
Swimming	0.0400	0.0311	0.0422	0.024
Cycling	0.0297	0.0179	0.0375	0.017
Golf	0.0986	0.0774	0.1073	0.058
Equestrian	0.0047	0.0023	0.0059	0.003
Water skiing	0.0007	0.0015	0.0003	0.001
Squash	0.0089	0.0030	0.0111	0.004
Field hockey	0.0037	0.0020	0.0005	0.001
Tennis	0.0246	0.0126	0.0443	0.013
Badminton	0.0127	0.0056	0.0135	0.006
X-country skiing	0.0166	0.0097	0.0240	0.009
Triathlon	0.0033	0.0016	0.0007	0.001
Five-pin bowling	0.0055	0.0078	0.0056	0.006
Rowing	0.0015	0.0021	0.0020	0.001
Track and field	0.0004	0.0006	0.0011	0.001
Figure skating	0.0016	0.0009	0.0028	0.001
Lacrosse	0.0010	0.0001	0.0002	0.001
Martial arts	0.0027	0.0031	0.0004	0.002
Boxing	0.0031	0.0010	0.0006	0.001
Archery	0.0015	0.0008	0.0016	0.001
Ten-pin bowling	0.0072	0.0117	0.0051	0.008
Curling	0.0233	0.0183	0.0250	0.015
Weightlifting	0.0034	0.0030	0.0006	0.002
Softball	0.0174	0.0092	0.0097	0.009
Rugby	0.0015	0.0006	0.0019	0.001
Gymnastics	0.0038	0.0022	0.0049	0.002
Football	0.0053	0.0034	0.0038	0.003
Sailing/yachting	0.0030	0.0023	0.0098	0.002

Bolded values represent those who have probabilities higher than their overall relative frequencies for the sample.

Italics represent the highest probability of that sport in any of the three latent classes.

Table 9.

Binary logistic odds ratios for each of the sports participation variables.

Latent class	Pop culture	Inactive	Omnivore
Variable
Golf	1.262	.648**	1.336
Gymnastics	.996	.581	2.089
Baseball	2.500**	.484*	.525
Basketball	1.424	.618	1.151
Equestrian	1.916	.465	1.010
Football	1.596	.620	.925
Ice hockey	2.006***	.618*	.548*
Rugby	1.195	.568	1.853
Soccer	2.188*	.282**	1.395
Softball	1.958*	.547	.748
Volleyball	1.699*	.636	.776
Weightlifting	1.495	1.243	.000
Cross-country skiing	1.530	.433*	1.579
Curling	1.308	.603	1.374
Ten-in bowling	.475	2.418*	.664
Badminton	1.849	.342*	1.502
Archery	1.195	.517	1.853
Boxing	5.990	.310	.000
Triathlon	2.993	.620	.000
Downhill skiing	1.202	.521**	1.845**
Martial arts	1.195	1.554	.000
Sailing/yachting	.746	.281	4.806**
Tennis	1.168	.308***	2.932***
Swimming	1.311	.675	1.173
Squash	2.253	.147*	1.972
Water skiing	.674	3.109	.000
Field hockey	3.594	.517	.000
Lacrosse	1.932E+9	.000	.000
Figure skating	1.794	.000	3.713
Track and field	.000	1.553	5.562
Rowing	.597	1.554	1.111
Cycling	1.559*	.369***	1.801*
Five-pin bowling	1.087	1.166	.583

p < .05; **p < .01; ***p < .001.

As the reader can see, there are many sports included in each cluster. For the pop culture cluster, only four sports do not have probabilities of inclusion that are higher than their overall relative frequencies for the data. These four sports are track and field, water skiing, five-pin bowling and ten-pin bowling. Eleven sports have their highest probabilities in this cluster. Eight of these 11 are team sports. The binary logistic regressions show fewer sports that have statistically significant predictive odds ratios for inclusion in this cluster. Baseball, ice hockey, soccer, softball, volleyball and cycling are all significantly predicted to be in this cluster. Because these sports are all sports with relatively high frequencies in the sample, it is likely that more of the sports from the LCA model would be statistically significant with increased numbers.

The inactive group has the highest number of sports that are excluded from the cluster and the lowest number of sports that are included with their highest probabilities of the three clusters. In total, 12 sports do not have probabilities of inclusion in this cluster that are higher than their overall relative frequencies. This cluster has four sports that have their highest probabilities in this cluster. These four are water skiing, rowing, five-pin and ten-pin bowling. This inactive group has the highest number (11) of statistically significant predictive odds ratios for these sports. Ten of these sports are significantly predicted to not be included in this cluster. The one sport that is significantly more likely to be in this cluster is ten-pin bowling.

The omnivore group has nine sports not included in this cluster. These nine are mostly ones that are highly included in either one or both of the other two cultural clusters. The omnivore cluster also has by far the highest number of sports that have their highest probability of inclusion. Eighteen of the sports have their highest probability in this cluster. In contrast to the pop culture cluster, most of these 18 are individual sports. Indeed, the four sports that are significantly predicted to be included in this cluster are individual sports. Cycling, downhill skiing, tennis and sailing/yachting are all more likely to be included in this cluster, while ice hockey is significantly less likely to be in this cluster.

Discussion and conclusion

What did we learn regarding prevailing theories of cultural consumption in Canada? Firstly, it appears that we can dismiss the individualisation thesis as representative of cultural consumption in Canada. This is because all of the cultural variables in this analysis are associated with higher levels of education or income, or both. Therefore, the results do not support the assertion that social class no longer structures consumption. However, this quantitative investigation is not able to capture all of the nuanced ways in which elements of identity might be bound up in underlining the class-based structuring of these patterns of consumption. Also, because it is still possible for the broader patterning of this consumption to be organised by more individualistic measures, the subsequent analysis of this paper is critical for providing further judgement on this theory.

The findings from the cultural analysis support the argument that Peterson’s omnivore thesis most accurately describes cultural engagement in Canada. This is because the omnivorous cultural consumption profile includes those who possess higher levels of cultural and economic capital. This is true, even as the age gap in consumption between the pop culture group and other latent classes also suggests a relative generational divide, something that other recent studies have observed in other national contexts (e.g. Bennett et al., 2009; Lizardo and Skiles, 2015; Purhonen et al., 2010; Roose, 2015; Savage et al., 2015). However, while age is the primary predictor of membership in the pop culture cluster, the most important variables for predicting membership in the inactive and omnivore clusters are capital possession measures. This is especially true of cultural capital, which is the most important predictor for these clusters. These results seem to refute the individualisation thesis, confirm the omnivore theory and suggest a primacy of cultural capital in patterning cultural omnivorism (and inactivity).

What did we learn towards the primary aim of understanding the landscape of sports participation in Canada and the broader cultural engagement of those who participate in sport? Firstly, based upon the inclusion of the sporting variables in the LCA, the patterning of the numerous sporting variables within the cultural clusters is not distinctively delineated. Indeed, the majority of sports had relatively high probabilities of being included in each latent class. However, there were sports that were only found in one of the clusters. Interestingly, the cluster with the most sports that were exclusive to it was the inactive cluster. The two variations of bowling (both five and ten pin), along with water skiing, were exclusive to the inactive cluster. The inclusion of water skiing is an interesting result. Other studies have treated water skiing as an upper class activity or found this to be the case (Widdop and Cutts, 2013). Bowling, however, has previously been identified as a working class sport (Eitzen and Sage, 1991). Because of the population centre results of the inactive cluster, the inclusion of water sports may provide a hint at a small number of this cluster who are higher capitally possessed rural persons who live permanently around bodies of water, without easy access to other cultural activities.

The significant sporting results for the pop culture class primarily include team sports. In Bourdieu’s framework, the association of team sports in the pop culture cluster makes sense. It can be understood through Bourdieu’s (1978) assertions that the mass spectacle of sport is appealing to these larger social classes, especially those who also consume other popular forms of culture. The spectacle of mass sporting performance, generally a product of team sports, thus influences consumers of other forms of popular culture in an application of Bourdieu’s homology thesis. In the case of Canada, the example of this dynamic par excellence is ice hockey. Ice hockey is Canada’s most popular sport. By law, it is its national winter sport. It was also one of the most strongly associated sports with the pop culture cluster. Importantly, it was significantly not associated with the other two cultural consumption clusters.

In furtherance of this example, the omnivores in this analysis eschew ice hockey. Ice hockey is the only sport from the regression analysis that is significantly much less likely to be participated in by omnivores. Omnivores are thus perhaps symbolically defining their omnivorism against specific activities in which they will conspicuously not engage (Bryson, 1996). They may also eschew ice hockey because of its violence (Bourdieu, 1978), viewing it more as a ‘prole’ sport for this reason (Wilson, 2002). This second assertion, however, is complicated by ice hockey’s status as representing normative Canadian whiteness and masculinity (Krebs, 2012). Because the omnivore group is also the most likely to be female, there may be a conscious or unconscious aversion to this example of hyper-masculinity in Canada. Alternatively, it may be a reflection of a ‘feminisation’ of highbrow culture (Purhonen et al., 2010), or of Bourdieu’s assertion that cultural capital, the primary predictor of omnnivorousness here, is to the feminine what economic capital is to the masculine (Bourdieu, 1978). Veenstra (2010) likewise finds results that mirror this gendered dynamic of economic and cultural capital in Canada. Critically, however, these considerations of masculinity in ice hockey (and other sports) also come from the top down. Therefore, outside of factors affecting taste, top down manifestations of these dynamics also affect participation. This importantly includes access to participation in sport, which has been and continues to generally be lower for women than for men.

In conclusion, overall cultural engagement in Canada supports the omnivore thesis. However, the patterning of sports participation fits well with Bourdieu’s homology thesis. Pursuant to the many scholars that have suggested a conceptual bridge between these two theories, the analysis in this paper suggests the same. Even as the broader patterning of cultural engagement in Canada conforms with the theory of the omnivore, sporting participation appears to be a specific domain of engagement by which the omnivore, as a high status group, generally does not engage in the same activities as the other two cultural profiles. Further research might find other cultural domains where otherwise omnivorous persons utilise that domain for the formation of distinction.

Footnotes

Acknowledgements

Special thanks to Statistics Canada for providing SPSS spreadsheets of the GSS results. Thanks also to two anonymous reviewers of this journal for their invaluable input on this paper.

Disclaimers

This paper is the sole work of the author and should not be construed to represent Statistics Canada or the Government of Canada.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

References

Atkinson

(2010) Class, Individualization and Late Modernity: In Search of the Reflexive Worker. Basingstoke: Palgrave MacMillan.

Asparouhov

Muthen

(2006) Robust chi square difference testing with mean and variance adjusted test statistics (Mplus Web Notes, No. 10). UCLA.

Bauman

(2002) Society under Siege. Cambridge, UK: Polity.

Beck

(1992) Risk Society: Towards a New Modernity. London: SAGE.

Bennett

Savage

Bortolaia

, et al. (2009) Culture, Class, Distinction. London: Routledge.

Bourdieu

(1978) Sport and social class. Social Science Information 17(6): 819–840.

Bourdieu

(1984) Distinction: A Social Critique of the Judgement of Taste. Cambridge, MA: Harvard University Press.

Bourdieu

(1986) The forms of capital. In: Richardson

(ed.) Handbook of Theory and Research for the Sociology of Education. New York: Greenwood, pp. 241–258.

Bourdieu

Wacquant

LÌJD

(1992) An Invitation to Reflexive Sociology. Chicago, IL: University of Chicago Press.

10.

Bryson

(1996) “Anything but heavy metal”: Symbolic exclusion and musical dislikes. American Sociological Review 61(5): 884–899.

11.

Chan

Goldthorpe

(2007) Social stratification and cultural consumption: Music in England. European Sociological Review 23(1): 1–19.

12.

Chan

Goldthorpe

(2010) Social status and consumption. In: Chan

(ed.) Social Status and Cultural Consumption. Cambridge: Cambridge University Press, pp. 1-27.

13.

Collins

(2003) Social exclusion from sport and leisure. In: Houlihan

(ed.) Sport and Society: A Student Introduction. London: SAGE, pp. 67–88.

14.

Collins

(2014) Sport and Social Exclusion. New York: Routledge.

15.

Coulangeon

Lemel

(2009) The homology thesis: Distinction revisited. In: Karen

Chris

(eds) Quantifying Theory: Pierre Bourdieu. Dordrecht: Springer, pp. 47–60.

16.

Cutts

Widdop

(2017) Reimagining omnivorousness in the context of place. Journal of Consumer Culture 17(3): 480-503

17.

Eitzen

Sage

(1991) Sociology of North American Sport. Madison, WI: Brown & Benchmark.

18.

Erickson

(1996) Culture, class, and connections. American Journal of Sociology 102: 217–251.

19.

Friedman

(2011) The cultural currency of a good sense of humour: British comedy and new forms of distinction. The British Journal of Sociology 62(2): 347–370.

20.

Gemar

(2019) Which sports do you like? Testing intra-domain omnivorousness in Canadian following of professional sport. International Review for the Sociology of Sport 54: 813–836. DOI: 10.1177/1012690217749243.

21.

Gerhards

Mutz

Hans

(2013) Social class and cultural consumption: The impact of modernisation in a comparative European perspective. Comparative Sociology 12(2): 160–183.

22.

Giddens

(1991) Modernity and Self-Identity: Self and Society in the Late Modern Age. Cambridge: Polity Press.

23.

Gruneau

(1975) Sport, social differentiation and social inequality. In: Ball

Loy

(eds) Sport and Social Order: Contributions to the Sociology of Sport. Reading, MA: Addison-Wesley Publishing Company, pp. 117–184.

24.

Gruneau

(1999) Class, Sport and Social Development. Champaign, IL: Human Kinetics.

25.

Jarness

(2015) Modes of consumption: From ‘what’ to ‘how’ in cultural stratification research. Poetics 53: 65–79.

26.

Johnston

Baumann

(2007) Democracy versus distinction: A study of omnivorousness in gourmet food writing. American Journal of Sociology 113(1): 165–204.

27.

Krebs

Andreas

(2012) Hockey and the reproduction of colonialism in Canada. In: Joseph

Darnell

Nakamura

(eds) Race and Sport in Canada: Intersecting Inequalities. Toronto, ON, Canada: Canadian Scholars’ Press, pp. 81–106.

28.

Lamprecht

Stamm

(1996) Age and gender patterns of sport involvement among the Swiss labor force. Sociology of Sport Journal 13(3): 274–287.

29.

Leguina

(2015) Musical distinctions in England – Understanding cultural homology and omnivourism through a methods comparison. Bulletin of Sociological Methodology / Bulletin de Méthodologie Sociologique 126(1): 28–45.

30.

Lizardo

Skiles

(2012) Reconceptualizing and theorizing “omnivorousness”: Genetic and relational mechanisms. Sociological Theory 30(4): 263–282.

31.

Lizardo

Skiles

(2015) Musical taste and patterns of symbolic exclusion in the United States 1993-2012: Generational dynamics of differentiation and continuity. Poetics 53: 9–21.

32.

Ollivier

(2008) Modes of openness to cultural diversity: Humanist, populist, practical, and indifferent. Poetics 36: 120–147.

33.

Peterson

(1992) Understanding audience segmentation: From elite and mass to omnivore and univore. Poetics 21(4): 243–258.

34.

Peterson

(1997) The rise and fall of highbrow snobbery as a status marker. Poetics 25(2–3): 75–92.

35.

Peterson

Kern

(1996) Changing highbrow taste: From snob to omnivore. American Sociological Review 61(5): 900–907.

36.

Peterson

Simkus

(1992) How musical tastes mark occupational status groups. In: Lamont

Fournier

(eds) Cultivating Differences: Symbolic Boundaries and the Making of Inequality. Chicago, IL: University of Chicago Press, pp. 152–186.

37.

Purhonen

Gronow

Rahkonen

(2010) Nordic democracy of taste? Cultural omnivorousness in musical and literary taste preferences in Finland. Poetics 38(3): 266–298.

38.

Roose

(2015) Signs of ‘emerging’ cultural capital? Analysing symbolic struggles using class specific analysis. Sociology 49(3): 556–573.

39.

Savage

Cunningham

Devine

, et al. (2015) Social Class in the 21st Century. London: Pelican, an imprint of Pelican Books.

40.

Scheerder

Vanreusel

Taks

, et al. (2002) Social sports stratification in Flanders 1969-1999: Intergenerational reproduction of social inequalities? International Review for the Sociology of Sport 37(2): 219–245.

41.

Scheerder

Vanreusel

Renson

(2005) Social stratification patterns in adolescents’ active sports participation behaviour: A time trend analysis 1969-1999. European Physical Education Review 11(1): 5–27.

42.

Shilling

(1993) The Body and Social Theory. London: SAGE.

43.

Sintas

Alvarez

(2002) Omnivores show up again: The segmentation of cultural consumers in Spanish social space. European Sociological Review 18(3): 353–368.

44.

Sintas

Alvarez

(2004) Omnivore versus univore consumption and its symbolic properties: Evidence from Spaniard’s performing arts attendance. Poetics 32(6): 471–491.

45.

Statistics Canada (2012) Cycle 19 Time Use Main Survey – Questionnaire package. Available at: http://www23.statcan.gc.ca/imdb-bmdi/instrument/4503_Q1_V4-eng.pdf

46.

Stempel

(2005) Adult participation sports as cultural capital. International Review for the Sociology of Sport 40(4): 411–432.

47.

Sugden

Tomlinson

(2000) Theorizing sport, social class and status. In: Coakley

Dunning

(eds) Handbook of Sports Studies. London: SAGE, pp. 309–321.

48.

Sullivan

Katz-Gerro (2007) The omnivore thesis revisited: voracious cultural consumers. European Sociological Review 23(2): 123-137.

49.

Swartz

(1997) Culture and Power: The Sociology of Pierre Bourdieu. Chicago, IL: University of Chicago Press.

50.

Taks

Renson

Vanreusel

(1995) Social stratification in sport: A matter of money or taste? European Journal of Sport Management 2(1): 4–14.

51.

Tampubolon

(2008) Revisiting omnivores in America circa 1990s: The exclusiveness of omnivores? Poetics 36(2–3): 243–264.

52.

Veenstra

(2007) Who the heck is Don Bradman? Sport culture and social class in British Columbia. Canadian Review of Sociology 44(3): 319–344.

53.

Veenstra

(2010) Culture and class in Canada. The Canadian Journal of Sociology 35(1): 83–111.

54.

Veenstra

(2015) Class position and musical tastes: A sing-off between the cultural omnivorism and Bourdieusian homology frameworks. Canadian Review of Sociology 52(2): 134–159.

55.

White

Wilson

(1999) Distinctions in the stands: An investigation of Bourdieu’s ‘habitus’, socioeconomic status and sport spectatorship in Canada. International Review for the Sociology of Sport 34(3): 245-264.

56.

Widdop

Cutts

(2013) Social stratification and sports participation in England. Leisure Sciences 35(2): 107–128.

57.

Widdop

Cutts

Jarvie

(2014) Omnivorousness in sport: The importance of social capital and networks. International Review for the Sociology of Sport 51(5): 596–616.

58.

Wilson

(2002) The paradox of social class and sports involvement: The roles of cultural and economic capital. International Review for the Sociology of Sport 37(1): 5–16.