Sport omnivorism: Social stratification of sports practice in Chile

Abstract

Using Chan and Goldthorpe’s theoretical framework of cultural consumption, this study analyses the patterns of social stratification in sports practice in Chile. Based on the data of the National Survey of Physical Activity and Sports Habits in the Population Aged 18 and Over 2015 in Chile (n = 5511), a latent class analysis was used to identify and characterize types of sports practice in this country. After this, logistic and multinomial regressions were performed to relate the types of sports practice found and its social determinants. The results support current literature in sports participation, finding evidence in favour of both the homology hypothesis and omnivorous–univorous hypothesis. However, in the same way as previous studies, it is shown that these arguments are reconfigured in the field of sport.

Keywords

age Chile gender latent class analysis omnivorousness social class social stratification sociology of culture sociology of sport sports practice

Introduction

According to the National Survey of Physical Activity and Sports Habits of the Chilean Population Aged 18 and Over, Chileans who practise sports or physical activity (PA) only reach 31.50% of the population. Moreover, when considering only people who exercise at least three sessions per week of at least 30 minutes each, that is, non-sedentary people, this rate reaches 20% in this country (Ministerio del Deporte de Chile, 2015). From a global perspective, these rates are very low. For example, in the case of the European Union, the rate of PA participation is about 61% (Van Tuyckom and Scheerder, 2010) and increases more when considering Scandinavian countries (Llopis-Goig, 2016).

This great distinction between Chile and Europe shows that the social stratification of sport is an issue that needs to be addressed. The international promise of “sport for all” has been demystified by a wave of studies addressing social inequities in sports and physical activities. In this respect, it is possible to distinguish two theoretical currents.

The first current focuses on explaining the differences among participants and non-participants (Kahma, 2010). Two approaches are considered to study sports disparities from this current: the epidemiological; and the sociological. On the one hand, the epidemiological approach focuses on the rate of sports inactivity as a function of various social determinants. The purpose of this approach is to understand physical inactivity as a source of social disadvantages that affect health (Willey et al., 2010). On the other hand, in the sociological approach, the objective is to understand the sources of social inequality that are involved in the field of sports. However, despite having different research objectives, both approaches agree on three points. First, background variables such as social class, education, and incomes are strong determinants of sports practice (Andersen and Bakken, 2019; Breuer et al., 2011; Kahma, 2010; Kamphuis et al., 2008; Lusmägi et al., 2016; Willey et al., 2010). Second, there is a significant gender gap in sports practice, even in the countries with the highest rates of PA (Cooky et al., 2016; Van Tuyckom and Scheerder, 2010; Van Tuyckom et al., 2012). Third, age has become a crucial factor in mitigating these differences, with younger age groups more likely to engage in sports and physical activities than the older ones (Lusmägi et al., 2016; Van Tuyckom and Scheerder, 2010). Both approaches also consider ethnicity to explain inequalities (Marshall et al., 2007).

The second theoretical current focuses on understanding the heterogeneity of the types of sports consumption and practice. This approach mainly uses studies of culture and lifestyles as a conceptual framework and investigates the differences in the kinds of sports practised according to social stratification variables. Sports consumption and practice studies trace back to Bourdieu (1989, 1993), who is considered by many researchers as one of its main influencers. However, the development of this field is quite recent in national contexts, and although there are a growing number of publications (Gemar, 2018a, 2018c; Lefèvre and Ohl, 2012; Lefèvre et al., 2020; Stempel, 2005; Widdop and Cutts, 2013; Widdop et al., 2016), there is little evidence for Latin American countries. In this connection, the following research aims to analyse the patterns of social stratification in sports practice in Chile. Specifically, it seeks to identify and characterize the types of sports practice present in the Chilean adult population and study its connection to the social stratification variables most reported in the literature, by answering the following questions:

How do sports and PA practice relate to social stratification variables in Chile?

Which cultural theory is best framed in the Chilean context in the field of sport?

Sports practice and theories of cultural consumption

According to Chan and Goldthorpe (2005, 2006, 2007) there are three arguments, which lead to an understanding of cultural consumption: (a) the homology argument; (b) the omnivore–univore argument; and (c) the individualization argument. These arguments have generated extensive literature about which theories best fit different areas of culture, including the practice and consumption of sports. Below are some findings in this field.

The homology argument and sports

The homology argument, elaborated in Pierre Bourdieu’s extensive work, is considered to be one of the main starting points for studies on sports practice and consumption. In particular, “Sport and social class” (Bourdieu, 1978), Distinction (Bourdieu, 1989), and “How can one be a sports fan?” (Bourdieu, 1993) are the most influential works in the studies that investigate sports and its sources of social stratification (Sánchez García and Moscoso Sánchez, 2015). This argument states that there exist differences between classes and class fractions about the dispositions, appraisals, and benefits that participants can get from the practice of different sports (Bourdieu, 1989). In this line, Kahma (2010) notes that from the Bourdieusian framework, the dominant class prefers individual and unique sports such as golf, horseback riding or tennis. Meanwhile, the middle class adopts a similar culture in their concerns for health but in a more rational manner. Thus, this class chooses team sports such as basketball, handball or football. Conversely, the working class favours muscular bodies, which translates into the choices of sports such as bodybuilding, wrestling, and other combat sports (Bourdieu, 1978, 1989; Kahma, 2010). In this way, according to Ohl (2000), the homology hypothesis allows us to clearly understand the place of social classes in the sporting spectrum.

Researchers have found evidence in favour of this argument. They suggest that a non-static reading of the relevant concepts involved in the homology framework, such as class, field, and habitus, allow understanding the relationship between the practice of sports and social positions in a clear and visible way (Ohl, 2000; Sánchez García and Moscoso Sánchez, 2015). For example, Stempel (2005) studied the relation between sports practice and cultural capital and economic capital, for 15 sports in the adult population of the United States. Using logistic regressions, the author specifically detailed the associations between these sports and the class fractions, inferring the social distances present in these practices. His findings show that tennis, golf, and running were reported as the most practised sports by high cultural and economic capital classes. Likewise, the author indicates that, according to the Bourdieusian framework, not only vertical class divisions are found, but also within the dominant classes. Specifically, it is shown that individuals of the cultural dominant class invest more time and energy in ascetic–feminine sports while those of the economically dominant class invest more in competitive–masculine sports. Moreover, years later Stempel (2019) observes that the Bourdieusian scheme is still present in US society. On this occasion the author observes in greater detail the distinctions within the dominant classes, finding that the sports practices between these fractions remain differentiated.

The omnivorous–univorous argument and sports

Notwithstanding the above, it has been noted that the context of Bourdieu’s analysis of sports practice is far from what is happening in today’s societies (Stempel, 2005). In this way, Ohl (2000) acknowledges that the explanatory capacity of social positions, as understood from this theory, may not be enough to account for the sports practice in post-modern society. For instance, Lefèvre and Ohl (2012) note that, although finding support for distinctive sports practices, there is more evidence for an argument of omnivority and massification of sports in the case of France. In this line, the omnivorous–univorous argument has been used commonly in studies of culture and lifestyles (Chan and Goldthorpe, 2005, 2006, 2007) and, more recently, in the sports field (Gemar, 2018a, 2018c; Lefèvre and Ohl, 2012; Widdop and Cutts, 2013; Widdop et al., 2016).

The omnivorous–univorous argument is introduced by Peterson and Kern (1996) who analyse the passage from the homology argument to a new form of distinction: the cultural omnivores. This argument suggests that due to a multiplicity of factors in the cultural field, the tastes of high-status groups become oriented towards a wider repertoire of activities. Thus, the dominant classes would not only consume and practise distinguished activities, but also those traditionally preferred by low-status groups. These low-status groups, however, continue to partake in only one or a few low-status activities, thus earning the moniker of univores. In this way, diverse studies, mainly in Europe and the United States, have used this argument to investigate different cultural fields such as music, theatre and, newspaper reading (Chan and Goldthorpe, 2005, 2006, 2007). And, although there are strong defenders of this argument, the literature has shown that the presence or absence of cultural omnivores depends on the contexts researched, in which these factors can manifest themselves in different ways (Lefèvre and Ohl, 2012; Lefèvre et al., 2020; Stempel, 2005; Widdop and Cutts, 2013).

The research agenda for the empirical study of cultural omnivores is in an early phase in the field of sport. The progress has focused on Europe and North America, mainly in France (Lefèvre and Ohl, 2012; Lefèvre et al., 2020), England (Widdop and Cutts, 2013; Widdop et al., 2016), Spain (Lefèvre et al., 2020), the USA (Stempel, 2019), and Canada (Gemar, 2018a, 2018c) and there is little research outside these geographical areas. Among these investigations, it is necessary to mention the case of England, which represents a significant theoretical and methodological example in the understanding of the cultural omnivore in the field of sport. In England, Widdop and Cutts (2013) found five different types of sports practice using the latent class analysis (LCA) methodology. The authors showed the existence of an omnivorous class associated with the upper strata, but also how this coexists with other types of classical sports participation such as highbrow or lowbrow. In this sense, this study is one of the first to argue that cultural omnivorism in the field of sport, far from existing in a dichotomous logic of the omnivore as opposed to the univore, can coexist with other sports lifestyles in the same national context.

Furthermore, cultural omnivorism has been studied not only regarding sports practice but also from the perspective of the consumption of sport as a spectator. In this case, the results show a different pattern of characterization. As an example, Gemar (2018c) uses LCA to investigate omnivorism for professional sport in Canada. The results of this study are significant because the author shows that no type of professional follower is entirely highbrow or lowbrow, without finding evidence for the univorous. Moreover, he determines that the most omnivorous group is, at the same time, the least socially distinguished, finding proof against Petersen’s argument and its later Bourdieusian reformulations (Lizardo and Skiles, 2012). In this way, other studies have found that the differences in sports consumption as a spectator are explained better by factors such as gender and age than by determinants of social position, as Kahma (2010) concluded in the national context of Finland. Therefore, as far as the relationship between sport and social positions is concerned, it is shown that the study of sports practice could lead to different conclusions from those obtained in the spectator’s field.

The individualization argument and sports

Lastly, the individualization argument is characterized by leaving the hypothesis of homology behind, proposing that the sources of social differentiation around cultural consumption would no longer be given by variables of origin, but by others such as gender, age, ethnicity or sexuality (Chan and Goldthorpe, 2005). The pioneer of this argument is Ulrich Beck, who with a critical look at social stratification studies declares the decadence of the concept of social classes (Becker and Hadjar, 2013). He argues that the expansion of education and economic growth have produced social changes that have brought down classical inequities (Becker and Hadjar, 2013: 212).

From this perspective, cultural consumption can be reinterpreted following Bauman (2000) as a choice for the construction of one’s own identity in a consumer society. Thus, the emphasis of this argument is on the ability of individuals to decide which sports to play or follow by their means and not because of some form of social structure. However, despite the interest of this argument, most studies on sport practice in national contexts show that the social stratification effect remains strong.

From cultural omnivorism to sports omnivorism

Understanding sport as a cultural practice requires adopting a theory of cultural consumption. As stated above, both the argument of homology and the omnivorous–univorous argument have shown empirical support in different national contexts for consumption and sports participation. However, these same findings have reconfigured the expected hypotheses of these conceptual frameworks in the field of sport in at least three senses.

First, in the coexistence of arguments of both the homology argument and the omnivorous–univorous argument in the types of sports practice that can be found in a certain context. There are a growing number of studies that interpret their findings in light of both hypotheses in at least some respect (Gemar, 2018a, 2018b, 2018c; Lefèvre and Ohl, 2012; Stempel, 2005; Widdop and Cutts, 2013; Widdop et al., 2016). This phenomenon has already tried to be explained for another cultural field by Lizardo and Skiles (2012), who postulate that cultural omnivorism will be acquired through socialization at an early age in the family setting and then develop in school and in the labour market, becoming a form of distinction in Bourdieu’s (1989) terms. However, it is not yet clear whether cultural omnivores in the field of sport would respond to this logic.

On the contrary, precisely the second reconfiguration of cultural theories in the sports field corresponds to the possibility of the presence of a multiplicity of omnivorous groups that do not necessarily stand out because they are socially distinguished. Thus, although some groups of omnivores may effectively obey Bourdieusian logic, there may be others who do not. For example, it can be expected that there are omnivorous categories associated with the lower and lower middle-classes (Gemar, 2018c; Widdop and Cutts, 2013) and others determined by sex or age (Widdop and Cutts, 2013). Thereby, the omnivore cannot always be equated with the notion of social distinction (Lefèvre and Ohl, 2012; Warde et al., 2007). Likewise, this multiplicity of omnivorous profiles gives scope to the idea of the omnivorous as cosmopolitan individuals open and tolerant to new forms of cultural practices and consumption as the one explained by Chan (2019). However, it is still necessary to analyse for each case whether, in the case of sports, cosmopolitanism would obey a form of social distinction (Gemar, 2018b; Ollivier, 2008), or a transversal change of attitudes towards new experiences (Chan, 2019). Considering this, the fact that on the sports field it is plausible to find more than one omnivorous group indicates that each of them must be analysed on the basis of its unique context.

On the last point, the third reconfiguration is the possibility of the appearance of new forms of social stratification in the lower and middle classes, different from those expected in the above-mentioned frameworks. Specifically, researchers’ results indicate that the middle and the less advantaged social strata are no longer anchored in the practice of a single sport or PA, as traditional hypotheses would suggest, but on the contrary omnivorous patterns can be expected in them (Gemar, 2018c; Widdop and Cutts, 2013). In this sense, the absence of the concept of cultural univore is a plausible postulate.

All of the above shows that cultural omnivorism, whether understood from its classical version (Peterson and Kern, 1996) or from its Bourdieusian reinterpretation (Lizardo and Skiles, 2012) or as an expression of cosmopolitan postmaterialism (Chan, 2019), finds in the field of sport a more dynamic version of the distribution of these practices in the social space, which in this study I will call sports omnivorism. This is an integrated view of the hypotheses of the theories of the sociology of culture applied to the studies of the sociology of sport, which seeks to understand in a contextual manner the lifestyles of sport and the PA of a given population.

Cultural consumption in Chile: theoretical and methodological challenges in the field of sport

In general terms, the Chilean cultural panorama has been described as strongly stratified (Gayo et al., 2009, 2013; Guëll Villanueva et al., 2011), giving support to the argument of homology. For example, Gayo et al. (2009) show how cultural consumption is associated with the most favoured groups. Likewise, the culturally inactive people are associated with older age groups or lower educational levels.

However, studies in Chile have rarely focused on cultural subfields such as sport. In Chile, research has incorporated sports as a generic variable in the analysis of cultural consumption. In this respect, one of the findings to be considered lies in a cultural polarization of genders in the Chilean middle class. According to Gayo et al. (2013), this group experiences a dichotomy between masculinity and sport on the one hand and femininity and intellectuality on the other. Yet, the information used in this and other preceding studies refers only to the reading of sports magazines, sports programmes and PA as general activities, but does not explicitly detail what type of consumption or specific practices they denote. Consequently, there is a theoretical gap in the analysis of heterogeneity behind the practices of PA and sports in Chile.

This gap can also be understood from methodological aspects. Studies of culture and lifestyles in Chile have been based primarily on the technique of multiple correspondence analysis (Gayo et al., 2009, 2013). This method has found in countless studies of culture an attempt at replication in national contexts (Bourdieu, 1989; Gayo et al., 2009; Herrera-Usagre, 2011). The advantage of this method is that it allows visualizing the relations between variables of cultural consumption and thus to test the hypothesis of the homology (Herrera-Usagre, 2011). Although research of this type has proven to be a necessary step, this method has two main limitations. First, it does not allow delving adequately into a single cultural field (Herrera-Usagre, 2011) and, second, it does not permit adequate testing of the omnivore–univorous hypothesis. Taking this into account, this study seeks to address these theoretical and methodological challenges.

Methodology

Data

This research uses data from the National Survey of Physical Activity Habits and Sports in the Population Aged 18 and Over (Ministerio del Deporte de Chile, 2015). This is a face-to-face survey that aims to identify the sport and physical exercise habits of the Chilean adult population in different dimensions of daily life. Regarding the studied population, the universe corresponds to “men and women aged 18 and over, of all socioeconomic levels, who belong to households located in urban and rural communes of the 15 regions of continental Chile” (Ministerio del Deporte de Chile, 2015: 8). The sample design was stratified by clusters and multistage, with a random selection at all stages of the process. First, using information from the National Statistics Institute, it was stratified geographically by administrative regions and communes. Then, the cases were allocated proportionally to the size of each commune. Finally, the households were selected randomly from the selected cases. The final sample was 5511 cases. The results were weighted to take into consideration the size of each region, and also the sex, the age and the socioeconomic variables (Ministerio del Deporte de Chile, 2015: 8–12).

Methods

The principal methodology adopted in this research is LCA. According to Collins and Lanza (2010), LCA looks for subtypes of individuals who exhibit similar patterns of individual characteristics. Thus, this technique is suitable to find different types of sports practice. In general terms, this methodology allows us to relate categorical indicators with a latent categorical variable, as shown in Figure 1. In this case, the latent variable corresponds to the type of sports practice, which explains the sporting patterns, the time per session and the weekly frequency of practice. LCA is a relatively new technique in sports practice studies (Gemar, 2018a, 2018b, 2018c; Widdop and Cutts, 2013; Widdop et al., 2016), but it has been used in other cultural fields (Chan and Goldthorpe, 2005, 2006, 2007).

Figure 1.

Latent class analysis path diagram for types of sports practice.

Regarding estimation, LCA provides two main parameters: latent class probabilities; and conditional probabilities. The first one refers to the likelihood p of an individual i to belong in a particular level j of the latent class C and the second one refers to the likelihood that an individual i, in a level j of a latent class C belongs to a particular level of an observed variable X (Linzer and Lewis, 2011). Both parameters are estimated using the maximum likelihood method according to the procedure given in Linzer and Lewis (2011).

As for the determination of the number of classes, the LCA technique does not provide a fixed solution. On the contrary, it is the researcher who must determine it based on different statistical criteria of parsimony and goodness-of-fit, observing if these fit or not with the theory that sustains his/her study (Widdop and Cutts, 2013). According to Collins and Lanza (2010), the goodness-of-fit statistics frequently used in the selection of the most parsimonious models that fit the observed data are the Akaike information criterion (AIC) and the Bayesian information criterion (BIC). In both criteria, the solution in which the statistical value is minimum must be chosen. Additionally, Cramer’s V-scores are reported before estimation of LCA. This statistic gives a first approximation to the association patterns of the variables (Gemar, 2018a).

After identifying the latent classes, logistic and multinomial regressions are performed to predict the membership of the types of sports practice found. This procedure is known as a three-step approach (Vermunt, 2010). Based on this analysis it is possible to compare the different groups found and predict which theory of cultural consumption best fits in Chile.

Variables

Two types of variable are considered for the LCA: those referring to the sports and physical activities practised; and those related to the frequencies of these practices. On the one hand, the practice of PA or sport’s variables are constructed from the question “What PA and/or sport do you practice and how many times a week?”. 16 sports categories are considered.¹ A large number of indicators have been taken into account because, as Widdop and Cutts (2013) pointed out, the complexity of the sports field must be captured from a wide range of sports that represent the symbolic barriers of this cultural domain. On the other hand, the frequency variables are the weekly frequency and the duration of the session. In this sense, these variables allow us to consider the cultural commitment and the degree of voraciousness (in terms of frequency) that individuals have with sports practices (Gemar, 2018a; Lefèvre et al., 2020) but also to investigate the differences between active (those who practise in at least one sport or PA, regardless of its duration) and non-sedentary groups (those who are physically active three times a week and for at least half an hour).

The variables to be used for regression analysis correspond to the most relevant determinants of sports practice that have been indicated in the literature. Particularly, as indicators of social position the social class of origin² and the educational level of both the interviewee and the head of household (institutionalized cultural capital) are considered. However, a limitation is that the income variable is not available. Likewise, the variables of gender, age and ethnic identification are regarded. The analysis is also controlled by the variables of the situation of the couple (de facto) and permanent health problems (deafness, dumbness, physical difficulty, and mental problems). The frequencies of the social stratification variables are presented in Table 1 and the frequencies of the variables to be used for the LCA are given in Table 2.

Table 1.

Descriptive statistics of the social stratification variables.

Variable	Categories	%	n
Social class	Working class	22.7	1223
	Middle class	55.4	2990
	Service class	21.9	1183
	Primary education or less	30.50	1681
Interviewee education	High school incomplete or equivalent	13.05	719
	Full high school or equivalent	24.70	1361
	Incomplete university	13.79	760
	Complete university	17.96	990
Gender	Male	46.16	2544
	Female	53.84	2967
Age group	18–29	21.27	1172
	30–39	17.15	945
	40–49	15.93	878
	50–59	18.08	996
	60–69	14.08	776
	>70	13.48	743
Head of household education	Primary education or less	24.08	1327
	High school incomplete or equivalent	12.50	689
	Full high school or equivalent	24.95	1375
	Incomplete university	10.47	577
	Complete university	27.99	1542
Ethnic identification	Other	8.71	480
	Chilean	91.29	5031
Couple status	Without partner	51.00	2812
	With partner	49.00	2699
Health problems	Yes	15.48	853
	No	84.52	4658

Table 2.

Frequency of the variables for the latent class analysis and Cramer’s V-scores.

	Frequency (%)	Social class	Education (interviewee)	Gender	Age group	Education (household)
Frequency variables of sports practice
Week frequency		0.064***	0.098***	0.186***	0.167***	0.120***
Not practising	68.2
Less than twice a week	11.9
Three or more times a week	19.8
Session duration		0.056***	0.098***	0.201***	0.165***	0.115***
Not practising	68.2
Less than one hour	8.50
More than one hour	23.2
Participation in sports
Fitness	10.5	0.044***	0.069***	0.106***	0.072***	0.083***
Football	9.50	0.010	0.064***	0.326*	0.210***	0.049***
Running	7.50	0.103***	0.130***	0.096***	0.149***	0.077***
Cycling	5.50	0.046***	0.099***	0.044**	0.095***	0.104***
Combat sports	5.50	0.052***	0.055***	0.048***	0.073***	0.074***
Outdoor sports	2.20	0.027	0.056***	0.020	0.014	0.054***
Swimming and other water sports	1.20	0.068	0.066***	0.020	0.031	0.066***
Yoga	1.10	0.044***	0.065***	0.045***	0.037	0.077***
Tennis	1.10	0.044***	0.066***	0.067***	0.034	0.062***
Extreme sports	0.70	0.021	0.050	0.055*	0.058***	0.066***
Recreational sports	0.60	0.003	0.028	0.044*	0.032	0.030
Weightlifting	0.60	0.018	0.008	0.056***	0.063***	0.052***
Team sports	0.60	0.022	0.061***	0.047*	0.079***	0.056***
Athletics	0.30	0.014	0.034	0.023*	0.030	0.043**
Gymnastics	0.30	0.014**	0.022	0.006	0.024	0.034
Club sports	0.10	0.017	0.022	0.025*	0.040	0.032

Note: ^***p < 0.01; ^**p < 0.05; ^*p < 0.1.

Results

First approach to LCA

Table 2 presents Cramer’s V-scores for each variable to be used in the LCA and shows whether or not there is a statistical association between the sports practice variables and the social stratification variables. This test has been used in other sports studies (Gemar, 2018a; Veenstra, 2007) and, according to Gemar (2018a) allows for a general notion of the social position of each variable, helping to interpret the analysis of latent classes better. In this case, it is shown that for frequency variables there is an association for all social stratification variables. However, associations do not exist in all cases for sports activity variables. Particularly, it is shown that there are sports that can be associated with certain social classes or educational levels, and others that are dominated by one sex or age group.

Appropriate number of latent classes

The first step in performing LCA is to determine the appropriate number of classes (Widdop and Cutts, 2013). Table 3 shows the goodness-of-fit statistics for different estimated latent class solutions.

Table 3.

Goodness-of-fit statistics for the latent class analysis.

Number of classes	Akaike information criterion (log-likelihood)	Bayesian information criterion (log-likelihood)	G2	Degrees of freedom
1	33,300.33	33,432.62	14301.82	5491
2	21,351.97	21,623.16	2311.459	5470
3	20,731.67	21,141.77	1649.165	5449
4	20,382.79	20,931.79	1258.278	5428
5	20,251.7	20,939.61	1085.194	5407
6	20,151.2	20,978.02	942.6979	5386

Note: Bold numbers represent the model to be chosen according to each criterion.

As shown, both goodness-of-fit statistics provide different solutions. Specifically, the AIC suggests a six-class model and the BIC suggests a four-class model. Faced with this dilemma, it was decided to analyse the solution that best fits the theory according to the characteristics of the classes, opting for the solution of four latent classes. Likewise, the literature concerning the information criteria for the selection of latent classes is not entirely conclusive, and the recommendations depend on the structures of the analysed data (Dziak et al., 2012).

In particular, according to Dziak et al. (2012) the BIC should be favoured when looking for larger and distinguishable latent classes. In contrast, the AIC should be used to favour a greater number of classes at the cost of losing parsimony in the analysis. Thus, the BIC has been favoured, since the solution of six classes does not provide more information than that of four classes and the loss of parsimony is lower.

Types of PA and sports practice

The next step is the description of the latent classes found. Table 4 shows the conditional probabilities of practising some sport or PA and Table 5 shows the conditional probabilities of the frequency variables. Both tables summarize the types of practice found and their respective sizes in the Chilean adult population.

Table 4.

Conditional probabilities of participating in a sport or physical activity.

	Class 1	Class 2	Class 3	Class 4
	Non-Participants	Highbrow Omnivore	Fitness Omnivore	Univore
Size	68.23%	12.59%	10.45%	8.73%
Fitness	0.00	0.00	1.00	0.02
Weightlifting	0.00	0.03	0.03	0.00
Cycling	0.00	0.32	0.05	0.03
Combat sports	0.00	0.04	0.02	0.00
Athletics	0.00	0.02	0.00	0.00
Football	0.00	0.05	0.03	1.00
Running	0.00	0.30	0.08	0.07
Club sports	0.00	0.01	0.00	0.00
Extreme sports	0.00	0.03	0.00	0.00
Recreational sports	0.00	0.04	0.00	0.02
Outdoor sports	0.00	0.17	0.04	0.01
Team sports	0.00	0.07	0.00	0.04
Yoga	0.00	0.05	0.02	0.00
Gymnastics	0.00	0.00	0.00	0.00
Tennis	0.00	0.01	0.00	0.03
Swimming and other water sports	0.00	0.04	0.02	0.02

Note: probabilities higher than the national average are in boldface type.

Table 5.

Conditional probabilities of frequency variables.

		Class 1	Class 2	Class 3	Class 4
		Non-Participants	Highbrow Omnivore	Fitness Omnivore	Univore
Size		68.23%	12.59%	10.45%	8.73%
Week frequency	Not practising	1.00	0.00	0.00	0.00
	Less than twice a week	0.00	0.39	0.33	0.58
	Three or more times a week	0.00	0.60	0.66	0.41
Session duration	Not practising	1.00	0.00	0.00	0.00
	Less than one hour	0.00	0.33	0.28	0.11
	More than one hour	0.00	0.66	0.72	0.88

Note: probabilities higher than the national average are in boldface type.

The first class corresponds to the Non-Participants and is characterized by not participating in any sport or PA. This is the largest group, with a size of 68.23%. This number is important when compared to European countries where this group is significantly smaller (Llopis-Goig, 2016). For example, in England, this group has a size of 41% (Widdop and Cutts, 2013), and in Finland, it is approximately 30% (Kahma, 2010).

The second class has a size of 12.59% and, using the taxonomy of Widdop et al. (2016), was called Highbrow Omnivore. This group is considered highbrow because it is the class that concentrates the highest proportion of sports and activities usually associated with the upper-middle and upper classes. Concretely, this is the only group that has an average practice higher than the national average in social club sports (rugby, golf, horseback riding, and rowing). Likewise, the Highbrow Omnivores practise, more predominantly than the national average, other sports such as cycling, running and outdoor sports, which makes them an omivorous group in two senses. First, in the number of sports they practise, which far exceeds those practised by the other classes. Second, that even if within the sports that this class practises there are some associated with high culture, there are also others in a massification process such as running and even other sports historically associated with the lower classes such as cycling. In this sense, this class evidences openness to different forms of sport practice.

The third class corresponds to the Fitness Omnivore, and its size is 10.45%. This group is characterized by the practise of physical activities and fitness sports such as physical conditioning exercises, work with machines in gyms, aerobic exercises such as the use of elliptical or stationary bicycle and other related activities such as recreational dance. It is also a group that presents more probabilities than the national average of practising running activities, outdoor sports and yoga. In this sense, due to the number of sports practised this group is considered as omnivore. However, unlike the previous class, in this group sports are clearly oriented to fitness and sports traditionally oriented to high culture have little or no presence. Nevertheless, the concept of fitness attributed to this category must be interpreted carefully when comparing this fitness group with those in other contexts. Particularly, in Chile this idea refers predominantly to the care and cultivation of the body using mechanical and repetitive exercises. Yet, in other contexts, for example in England, in addition to this, this category is associated with swimming and cycling (Widdop et al., 2016).

The fourth class corresponds to the Univore class and represents 8.73% of the adult population. This social group is predominantly associated with the practise of Chile’s most popular sport, football (soccer). Concerning the other sports or activities, the Univore class presents low patterns of practice, and none of them is higher than the national average, which is why this class was named Univore.

Finally, with regard to the frequency variables, although the three groups of practitioners stand out for being physically active, not all of them show non-sedentary behaviour. Concretely, the Univore class, despite being the most intensive group regarding session length (the conditional probability of practising more than one hour per session is 88%), stands out for being the only one with probabilities of sedentary behaviour since about 60% of this group practise sports or physical activities only twice or once a week. In contrast, members of omnivorous groups are highlighted for participating three or more times in their physical activities. In this sense, these latter groups show greater sporting commitment.

Social stratification of the types of sports and PA in the adult population of Chile

Social characterization of the types of sports practice

Once the latent classes are identified, the connection between them and the social determinants will be analysed. Table 6 presents the distribution of the social characteristics of the different latent classes.

Table 6.

Descriptive statistics for latent classes according to social stratification variables.

		Class 1	Class 2	Class 3	Class 4
		Non-Participants	Highbrow Omnivore	Fitness Omnivore	Univore
Size		68.23%	12.59%	10.45%	8.73%
Social class	Working class	25.39	12.98	12.63	18.23
	Middle class	55.54	45.37	64.98	57.44
	Service class	19.05	41.64	22.38	24.31
Interviewee education	Primary education or less	35.87	16.16	19.13	23.02
	High school incomplete or equivalent	13.51	13.13	11.82	10.58
	Full high school or equivalent	24.46	25.54	22.26	28.21
	Incomplete university	12.50	17.46	12.17	20.33
	Complete university	13.64	27.70	34.60	17.84
Gender	Male	60.31	59.36	32.52	93.97
	Female	39.68	40.63	67.47	6.02
Age group	18–29	15.71	29.62	22.43	51.35
	30–39	15.44	17.91	26.95	17.46
	40–49	15.84	20.37	12.52	14.13
	50–59	19.86	16.76	14.60	10.18
	60–69	16.61	6.64	13.73	5.61
	>70	16.51	8.67	9.73	1.24
Head of household education	Primary education or less	28.77	11.09	13.54	18.87
	High school incomplete or equivalent	14.04	7.20	8.50	13.07
	Full high school or equivalent	24.78	22.76	26.56	27.59
	Incomplete university	9.65	14.26	10.76	10.78
	Complete university	22.73	44.66	40.62	29.66
Ethnic identification	Chilean	90.29	94.52	96.58	87.96
	Other	9.70	5.47	3.41	12.03
Couple status	Without partner	46.64	59.74	54.51	68.46
	With partner	53.35	40.25	45.48	31.53
Health problems	Yes	19.94	6.77	8.33	1.87
	No	80.05	93.22	91.66	98.12

As for the social class variable, Table 6 indicates that the Highbrow Omnivore class is the one most represented by the class of services, with 41.64% of its members in this category. Likewise, the Fitness Omnivore class and the Univore class stand out for being composed mainly of middle-class individuals. A higher percentage of working-class people and a lower proportion of service-class members characterize the Non-Participants.

In education terms, there are interesting differences between the Highbrow Omnivore group and the Fitness Omnivore group. While the former gathers the largest proportion of individuals where the head of household has a university education or more, the latter is characterized by the largest proportion where the individuals themselves have a university education or more. The Non-Participants is the least favoured in educational terms.

When considering gender, both the Non-Participants and the Highbrow Omnivore can be described as predominantly masculine, while about 60.00% of the members in both classes are men. This situation is more extreme in the case of the Univore Class, where the male composition is 93.97%. The only group that is predominantly female is the Fitness Omnivore (67.47%).

Concerning the age group, there are also significant differences between classes. Specifically, the youngest group is the Univore class, with the majority of its members in the 18–29 age group (51.35%). Likewise, the Fitness Omnivore and the Non-Participants, stand out for having an equitable age distribution and the Highbrow gathers individuals from 18 to 59 years old. Elderly age groups have a low presence in all three groups of participants.

To strengthen the analysis of the types of sports practice, two additional types of models will be estimated. On the one hand, to analyse the classes globally, logistic regression models will be used. On the other hand, in order to compare the classes, and thus understand in a more adequate way how the theory of cultural consumption and lifestyles fits with the Chilean panorama, a multinomial regression model will be estimated.

Global analysis of the types of sports practice

Table 7 shows four logistic models (Models 1–4) that predict belonging to the four latent classes found.

Table 7.

Logistic models for latent class membership according to social stratification variables (logits).

	Non-Participants	Highbrow Omnivore	Fitness Omnivore	Univore
	(1)	(2)	(3)	(4)
Social class
Reference category: Working class
Middle class	−0.117	−0.086	0.279*	0.148
	(0.100)	(0.148)	(0.156)	(0.164)
Service class	−0.390***	0.623***	−0.230	0.311
	(0.127)	(0.174)	(0.195)	(0.216)
Interviewee education
Reference category: Primary education or less
High school incomplete or equivalent	−0.435***	0.792***	0.377*	−0.425*
	(0.124)	(0.172)	(0.193)	(0.218)
Full high school or equivalent	−0.224**	0.292*	0.115	0.123
	(0.107)	(0.154)	(0.165)	(0.177)
Incomplete university	−0.117	0.138	0.084	0.107
	(0.126)	(0.170)	(0.198)	(0.196)
Complete university	−0.581***	0.219	1.136***	−0.277
	(0.119)	(0.157)	(0.172)	(0.201)
Gender
Reference category: Male
Female	0.837***	−0.654***	0.850***	−3.109***
	(0.065)	(0.089)	(0.100)	(0.197)
Age group
Reference category: 18–29
30–39	0.485***	−0.330**	0.393***	−0.945***
	(0.099)	(0.137)	(0.138)	(0.154)
40–49	0.532***	0.115	−0.409**	−0.793***
	(0.103)	(0.133)	(0.166)	(0.167)
50–59	0.762***	−0.060	−0.330**	−1.265***
	(0.106)	(0.141)	(0.159)	(0.184)
60–69	1.064***	−0.652***	−0.154	−1.697***
	(0.124)	(0.184)	(0.171)	(0.248)
>70	1.123***	−0.283	−0.154	−3.085***
	(0.132)	(0.177)	(0.192)	(0.422)
Head of household education
Reference category: Primary education or less
High school incomplete or equivalent	0.067	−0.238	0.067	0.121
	(0.139)	(0.216)	(0.221)	(0.220)
Full high school or equivalent	−0.366***	0.527***	0.484***	−0.188
	(0.121)	(0.181)	(0.186)	(0.198)
Incomplete university	−0.482***	0.866***	0.497**	−0.446*
	(0.150)	(0.212)	(0.229)	(0.248)
Complete university	−0.549***	0.877***	0.451**	−0.250
	(0.136)	(0.196)	(0.204)	(0.227)
Ethnic identification
Reference category: Other
Chilean	−0.345***	0.297	1.001***	−0.284
	(0.121)	(0.183)	(0.243)	(0.175)
Couple status
Reference category: Without partner
With partner	0.367***	−0.328***	−0.102	−0.279**
	(0.068)	(0.093)	(0.098)	(0.123)
Health problems
Reference category: No
Yes	1.131***	−0.916***	−0.593***	−1.888***
	(0.120)	(0.174)	(0.169)	(0.346)
Constant	0.534***	−2.441***	−4.260***	−0.123
	(0.150)	(0.226)	(0.297)	(0.217)
Observations	5425	5425	5425	5425
Log-likelihood	−2874.497	−1814.889	−1583.065	−1168.876
Akaike information criterion	5788.995	3669.778	3206.130	2377.752

Note: ^***p < 0.01; ^**p < 0.05; ^*p < 0.1.

Model 1 predicts that working-class individuals with a low level of education have a greater chance of belonging to Non-Participants. This finding is relevant insofar as the simple practise of sports (in any of its forms) seems to be the greatest social distinction for the Chilean case. Likewise, women and people over 60 are more likely to belong to Non-Participants, favouring the approaches of Bourdieu (1978) about gender and age variables when dealing with the most disadvantaged social groups. Moreover, Non-Participants also stand out as the only group in which those who identify themselves as indigenous have a 40% greater chance of belonging than those identified as Chileans.

Regarding the Highbrow Omnivore, Model 2 shows that the class of origin (of services) and the education of the head of household are the key determinants of social position in belonging to this group. In this sense, given the social and sporting characteristics of this class, evidence can be attributed in favour of the omnivorous–univorous argument for this social group (Peterson and Kern, 1996). This finding can also be interpreted as a reinterpretation of the Bourdieusian argument proposed by Lizardo and Skiles (2012), insofar as the most relevant variables in this class are those of origin.

As for the Fitness Omnivore, Model 3 predicts that education is the most determining factor when explaining membership in this class. However, in terms of social class, unlike the previous group the Fitness Omnivore is predominantly a middle-class group since those who belong to this category are 1.32 (exp(0.279)) times more likely to be in this class than those in the working class. Likewise, Model 3 emphasizes that the Fitness Omnivore is the only class that is predominantly female. In terms of age participation, regression coefficients show that the individuals with the highest chances of belonging to this group are those between 18 and 39 years old.

For Model 4, it is shown that belonging to the Univore class is better explained by being male and belonging to the youngest group, than by the variables of social position. However, this finding must be interpreted with care because, as will be seen below, there are differences in terms of social position when this class is contrasted with the omnivorous groups.

Comparative analysis of the types of sports practice

To complete the analysis, it is necessary to compare and relate the types of sport present in the Chilean adult population. For this purpose, a multinomial regression model was estimated. Table 8 shows Model 5 that relates the types of sports practice found with the social determinants. In this case, the Non-Participants are chosen as the reference category since it is relevant to compare the practising groups with the most socially disadvantaged group.

Table 8.

Model 5 – Multinomial model for the latent classes with respect to the Non-Participants class according to social stratification variables (relative logits).

Reference category: Non-Participants	Highbrow Omnivore	Fitness Omnivore	Univore
Social class
Reference category: Working class
Middle class	−0.05(0.15)	0.28(0.16) *	0.15(0.17)
Service class	0.67(0.18) ***	−0.04(0.20)	0.51(0.22) **
Interviewee education
Reference category: Primary education or less
High school incomplete or equivalent	0.78(0.18) ***	0.44(0.20) **	−0.26(0.22)
Full high school or equivalent	0.30(0.16) *	0.13(0.17)	0.17(0.18)
Incomplete university	0.15(0.18)	0.07(0.20)	0.13(0.20)
Complete university	0.36(0.16) **	1.15(0.18) ***	−0.06(0.21)
Gender
Reference category: Male
Female	−0.91(0.09) ***	0.45(0.10) ***	−3.23(0.20) ***
Age group
Reference category: 18–29
30–39	−0.54(0.14) ***	0.18(0.14)	−1.04(0.16) ***
40–49	−0.17(0.14)	−0.55(0.17) ***	−0.88(0.17) ***
50–59	−0.40(0.15) ***	−0.54(0.16) ***	−1.41(0.19) ***
60–69	−1.01(0.19) ***	−0.46(0.17) ***	−1.91(0.25) ***
>70	−0.71(0.18) ***	−0.46(0.19) **	−3.26(0.42) ***
Head of household education
Reference category: Primary education or less
High school incomplete or equivalent	−0.23(0.22)	0.03(0.22)	0.07(0.22)
Full high school or equivalent	0.57(0.18) ***	0.52(0.19) ***	−0.06(0.20)
Incomplete university	0.88(0.22) ***	0.56(0.23) **	−0.23(0.25)
Complete university	0.93(0.20) ***	0.54(0.21) ***	−0.02(0.23)
Ethnic identification
Reference category: Other
Chilean	0.37(0.19) **	1.05(0.24) ***	−0.14(0.18)
Couple status
Reference category: Without partner
With partner	−0.43(0.10) ***	−0.22(0.10) **	−0.40(0.13) ***
Health problems
Reference category: No
Yes	−1.13(0.17) ***	−0.77(0.17) ***	−2.12(0.35) ***
Constant	−1.85(0.23) ***	−3.60(0.30) ***	0.08(0.22)
Akaike information criterion	8821.48
Bayesian information criterion	9217.09
Log-likelihood	−4350.74
Deviance	8701.48
Number observed	5425

Note: ^***p < 0.01; ^**p < 0.05; ^*p < 0.1.

As for social class, it is possible to observe that the relative risk (RR) of belonging to both the Highbrow Omnivore class and the Univore class is significantly higher for individuals of the service class as compared to individuals of the working class. In particular, this is 1.9 (exp(0.67)) times higher in the case of the Highbrow Omnivore class and 1.6 times higher in the case of the Unviore class. On the contrary, the differences between the Fitness Omnivore and Non-Participants are in the middle-class category. In particular, the RR of belonging to the Fitness Omnivore class compared to the reference category is 32% higher for middle-class people compared to people belonging to the working class.

The situation is different when considering the variables of the educational level. Specifically, it can be noted that when considering the Univore class with respect to Non-Participants, there are no statistically significant differences in the RR in belonging to all educational categories. However, this result changes when considering the other two sports groups where individuals with a higher level of education are more likely to belong.

Lastly, as for age stratification, it is shown that for the three groups of participants the RR of belonging decreases abruptly when passing from the younger age groups to the older ones.

The omnivorous–univorous argument in the Chilean context

The findings presented so far have made it possible to compare participants with Non-Participants. However, considering the category of Non-Participants as a reference hides the differences between the Omnivore class and the Univore class, a fundamental aspect of this research. Table 9 shows Model 5 when the Univore class is the reference category. It is observed that the RR of belonging to the Highbrow Omnivore group and the Univore group is explained better by differences in educational levels than by social class of origin. In particular, the education of the head of household is the variable with the most significant effect on belonging to the Highbrow Omnivore class. Likewise, the effect of the interviewee’s education is in the same direction, although to a minor extent.

Table 9.

Model 5 – Multinomial model for the latent classes with respect to the Univore class according to social stratification variables (relative logits).

Reference category: Univore	Highbrow Omnivore	Fitness Omnivore
Social class
Reference category: Working class
Middle class	−0.20(0.21)	0.12(0.22)
Service class	0.16(0.26)	−0.55(0.28) **
Interviewee education
Reference category: Primary education or less
High school incomplete or equivalent	1.04(0.26) ***	0.70(0.28) **
Full high school or equivalent	0.13(0.22)	−0.04(0.23)
Incomplete university	0.02(0.23)	−0.06(0.27)
Complete university	0.43(0.23) *	1.21(0.25) ***
Gender
Reference category: Male
Female	2.32(0.21) ***	3.68(0.22) ***
Age group
Reference category: 18–29
30–39	0.50(0.19) ***	1.22(0.20) ***
40–49	0.71(0.19) ***	0.33(0.23)
50–59	1.01(0.21) ***	0.87(0.24) ***
60–69	0.90(0.29) ***	1.45(0.29) ***
>70	2.55(0.45) ***	2.80(0.46) ***
Head of household education
Reference category: Primary education or less
High school incomplete or equivalent	−0.30(0.29)	−0.04(0.30)
Full high school or equivalent	0.63(0.25) **	0.58(0.26) **
Incomplete university	1.11(0.30) ***	0.79(0.33) **
Complete university	0.95(0.28) ***	0.56(0.29) *
Ethnic identification
Reference category: Other
Chilean	0.51(0.23) **	1.19(0.29) ***
Couple status
Reference category: Without partner
With partner	−0.03(0.14)	0.19(0.15)
Health problems
Reference category: No
Yes	0.98(0.38) ***	1.34(0.38) ***
Constant	−1.93(0.29) ***	−3.68(0.36) ***
Akaike information criterion	8821.48
Bayesian information criterion	9217.09
Log-likelihood	−4350.74
Deviance	8701.48
Number observed	5425

Note: ^***p < 0.01; ^**p < 0.05; ^*p < 0.1.

Similarly, the chances of belonging to the Fitness Omnivore class are better explained by education than by social classes. However, in this case, the most significant effect was found in the education variable of the interviewee and not in that of the head of household as in the previous case. Concretely, the chances of belonging to the Fitness Omnivore class with respect to the Univore class are 230% higher for those with a university education or more, with respect to the reference category. Finally, Model 5 also highlights the above results: the Univore class is the most masculine and youngest of all latent classes found, since the gender (female) and age group (30 to 70 and older) coefficients are positive and significant at 99% confidence.

Considering this, omnivorous classes are in a better social position than the Univore, with education being the most relevant variable, although the social class also plays a determining role. Nevertheless, it is necessary to emphasize from the previously exposed results that the Univore group is predominantly young and middle class, which means that the omnivorous/univorous distinction is not necessarily a difference between high and low culture, especially when comparing the Fitness Omnivore group with the Univore. This finding contrasts with the international literature, in which the Univore class has been predominantly linked to the working class (Widdop and Cutts, 2013), showing a greater social distance between the omnivore and the univore, which in this case exists but is minor.

Discussion and conclusion

This research has shown that the practice of sports in Chile is strongly stratified. Thus, the promise of “sport for all” is far from being the reality faced by Chile. In this case, however, the inequalities are even higher. In Chile, the most significant social distinction, as understood by Bourdieu (1989), is the simple practice of PA. Indeed, the most disadvantaged social groups, who have the lowest levels of education and come from a home where the main occupation is unskilled and generally manual, do not participate in any sporting activity on a regular basis. On the contrary, participation in these activities is more common in the middle class and service class, and in the more educated groups.

However, despite this, important social differences have been found between the practitioners of physical activities and sports. Particularly, evidence has been found for both the homology argument and the omnivorous–univorous argument. On the one hand, the homology argument exists for at least two reasons. First, there is a strong difference between participants and non-participants that is determined by social position. Second, there is a proportion of the population that practises sports traditionally associated with the upper strata and another that practises sports usually associated with the middle and lower-middle strata, as is the case of soccer in Chile. Likewise, these differences are explained better by institutionalized cultural capital than by differences in the social class of origin.

On the other hand, the omnivorous–univorous argument coexists with the previous one, since there are groups in the adult population with a tendency to practise different types of sport, and other groups that only predominantly practise a single sport. The coexistence of this argument with the criterion of homology lies in the fact that precisely the Highbrow Omnivore class is, at the same time, the most favoured in terms of social position and the Univore class is the least favoured (without considering the Non-Participants). In this sense, evidence has been found in favour of a reinterpretation of the Bourdieusian theory, such as those suggested by Lizardo and Skiles (2012) and other researchers in the particular cases of the sociology of sport (Gemar, 2018a, 2018c; Lefèvre and Ohl, 2012; Stempel, 2005; Widdop and Cutts, 2013; Widdop et al., 2016).

However, the omnivorous–univorous argument is re-configured in this field because the figure of the cultural omnivore is not always associated with the most distinguished classes. Specifically, this research showed a diversity of omnivorism in Chile, with a more middle-class one. This is the case of the Fitness Omnivore that, without being a distinguished class regarding social position, has the chance to participate in different fitness sports. Thus, the multiplicity of omnivore classes found in this and other studies makes the omnivore–univore dichotomy break when observing the sports phenomenon, giving way to diverse forms of cultural omnivores.

Likewise, the omnivorous–univorous argument is also transformed in this field because the univore group is not necessarily the most disadvantaged in society. The Univore class, as has been shown, although it has the highest percentage of the working class within the practising classes, is also a predominantly middle-class group. Consequently, the close association between univorism and low culture is questioned in this research.

With the above, it has been shown empirically how the argument of the homology and the omnivorous–univorous argument are re-configured in the field of sport, in three fundamental aspects, namely:

The coexistence of the homology argument with the omnivorous–univorous argument.

The multiplicity of forms of cultural omnivores.

The cultural univore as middle class or absent typology.

I call these new social categories in the field of sport, present both in this and in previous research on the sociology of sport (Gemar, 2018c; Lefèvre and Ohl, 2012; Lefèvre et al., 2020; Stempel, 2005; Widdop and Cutts, 2013; Widdop et al., 2016), sports omnivorism. I use this concept because although the frame of reference of the classic theories of the sociology of culture serves to explain much of the social stratification of sport, they are not sufficient. In this way, sports omnivorism is a dynamic concept, since it involves a multiplicity of ways in which society makes sport its own, breaking with the dichotomous schemes used until now to explain the practices of sport in society.

In addition to the above, two of the most important findings in this research cannot be overlooked: the gender gap; and the age gap. The fact that sports groups are predominantly young and dominated by men shows that there is still some way to go in promoting sport as a public policy in Chile.

As for the limitations of this study, it is necessary to mention three fundamental aspects. First, the lack of an income variable made it impossible to test hypotheses of economic capital. Second, the high rate of inactivity present in Chile leads to tiny proportions in specific sports. Thus, it was not possible to observe the behaviour of sub-classes, such as the elite class, related to golf or other club sports. In this sense, qualitative research could better characterize minority sports groups such as these. Third, as anticipated in Bourdieu (1989), the declaration of sporting practices in surveys does not allow us to fully understand their social heterogeneity. The practice of the same sport can mean something very different depending on the contexts in which it is embedded (Bourdieu, 1989, 1993; Sánchez García, 2008). For example, in the Chilean case, declaring to play soccer and practising it in a sector with poor resources could be associated with different values and social benefits in comparison with practising the same sport in sectors with high resources.

Finally, although this study has outlined the general patterns of stratification of sports practice in Chile, there are still questions to be answered. On the one hand, because the data source used only considers adults as a population, it becomes relevant to investigate the stratification patterns of the younger groups. On the other hand, the task of investigating sports consumption, that is, the consumption of sports through different media as a spectator, both in Chile and in other Latin American countries, remains pending. Lastly, the importance of the social study of sport in Latin America must be emphasized, where sociological literature has been primarily disconnected from this subject, but surely where its understanding is most needed.

Footnotes

Acknowledgements

The author thanks the anonymous reviewers of this article for their contributions and the Chilean Ministry of Sport for providing the data for this research.

Declaration of Conflicting Interests

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

Funding

The author disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by the Fondo Nacional de Desarrollo Científico y Tecnológico [Grant number 1170558]; and the San Andres Grant from the Pontificia Universidad Católica de Chile.

ORCID iD

Javier Reyes

Notes

References

Andersen

Bakken

(2019) Social class differences in youths’ participation in organized sports: What are the mechanisms? International Review for the Sociology of Sport 54(8): 921–937.

Bauman

(2000) Trabajo, Consumismo y Nuevos Pobres [Work, Consumerism and New Poors]. Barcelona: Editorial Gedisa. [In Spanish.]

Becker

Hadjar

(2013) “Individualisation” and class structure: How individual lives are still affected by social inequalities. International Social Science Journal 64(213–214): 211–223.

Bourdieu

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

Bourdieu

(1989) La Distinción. Crítica Social del Juicio [The Distinction. Social Criticism of the Trial]. Madrid: Taurus. [In Spanish.]

Bourdieu

(1993) How can one be a sports fan? Empiria 30: 169.

Breuer

Hallmann

Wicker

(2011) Determinants of sport participation in different sports. Managing Leisure 16(4): 269–286.

Chan

(2019) Understanding cultural omnivores: Social and political attitudes. British Journal of Sociology 70(3): 784–806.

Chan

Goldthorpe

(2005) The social stratification of theatre, dance and cinema attendance. Cultural Trends 14(3): 193–212.

10.

Chan

Goldthorpe

(2006) Social stratification and cultural consumption: Music in England. European Sociological Review 23(1): 1–19.

11.

Chan

Goldthorpe

(2007) Social stratification and cultural consumption: The visual arts in England. Poetics 35(2–3): 168–190.

12.

Collins

Lanza

(2010) Latent Class and Latent Transition Analysis: With Applications in the Social, Behavioral, and Health Sciences. Hoboken: John Wiley & Sons.

13.

Cooky

Begovic

Sabo

, et al. (2016) Gender and sport participation in Montenegro. International Review for the Sociology of Sport 51(8): 917–939.

14.

Dziak

Coffman

Lanza

, et al. (2012) Sensitivity and specificity of information criteria. The Methodology Center and Department of Statistics, Penn State, The Pennsylvania State University 16(30): 140.

15.

Erickson

Goldthorpe

Portocarrero

(1979) Intergenerational class mobility in three western European societies. British Journal of Sociology 30(4): 415–441.

16.

Gayo

Teitelboim

Luisa Méndez

, et al. (2009) Patrones culturales de uso del tiempo libre en Chile: Una aproximación desde la teoría bourdieuana [Cultural patterns of use of free time in Chile: An approach from the Bourdieuian theory]. Universum (Talca) 24(2): 42–72. [In Spanish.]

17.

Gayo

Teitelboim

Méndez

(2013) Exclusividad y fragmentación: Los perfiles culturales de la clase media en Chile [Exclusivity and fragmentation: The cultural profiles of the middle class in Chile]. Universum (Talca) 28(1): 97–128. [In Spanish.]

18.

Gemar

(2018a) Sport as culture: Social class, styles of cultural consumption and sports participation in Canada. International Review for the Sociology of Sport 55(2): 186–208.

19.

Gemar

(2018b) Sport in broader leisure lifestyles: An analysis of the professional sport consumer’s cultural engagement. International Review for the Sociology of Sport 55(3): 291–309.

20.

Gemar

(2018c) Which sports do you like? Testing intra-domain omnivorousness in Canadian following of professional sport. International Review for the Sociology of Sport 54(7): 813–836.

21.

Guëll Villanueva

Morales Olivares

Núñez

(2011) Tipología de prácticas de consumo cultural en Chile a inicios del siglo XXI: mismas desigualdades, prácticas emergentes, nuevos desafíos [Typology of cultural consumption practices in Chile at the beginning of the 21st century: same inequalities, emerging practices, new challenges]. Universum (Talca) 26(2): 121–141. [In Spanish.]

22.

Herrera-Usagre

(2011) El consumo cultural en españa. Una aproximación al análisis de la estratificación social de los consumos culturales y sus dificultades metodológicas [Cultural consumption in Spain. An approach to the analysis of the social stratification of cultural consumption and its methodological difficulties]. EMPIRIA. Revista de Metodología de las Ciencias Sociales 22: 141–172. [In Spanish.]

23.

Kahma

(2010) Sport and social class: The case of Finland. International Review for the Sociology of Sport 47(1): 113–130.

24.

Kamphuis

CBM

Van Lenthe

Giskes

, et al. (2008) Socioeconomicstatus, environmental and individual factors, and sports participation. Medicine & Science in Sports & Exercise 40(1): 71–81.

25.

Lefèvre

Ohl

(2012) Consuming sports: Distinction, univorism and omnivorism. Sport in Society 15(1): 44–63.

26.

Lefèvre

Routier

Llopis-Goig

(2020) Sports participation in France and Spain: An international comparison of voraciousness for sport. Poetics. Epub ahead of print 19 February. DOI: 10.1016/j.poetic.2019.101429.

27.

Linzer

Lewis

(2011) polca: An R package for polytomous variable latent class analysis. Journal of Statistical Software 42(10): 1–29.

28.

Lizardo

Skiles

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

29.

Llopis-Goig

(2016) Participación Deportiva en Europa: Políticas, Culturas y Prácticas [Sports Participation in Europe: Policies, Cultures and Practices]. Barcelona: Editorial UOC. [In Spanish.]

30.

Lusmägi

Einasto

Roosmaa

(2016) Leisure-time physical activity among different social groups of Estonia: Results of the national physical activity survey. Physical Culture and Sport. Studies and Research 69(1): 43–52.

31.

Marshall

Jones

Ainsworth

, et al. (2007) Race/ethnicity, social class, and leisure-time physical inactivity. Me. Dicine & Science in Sports & Exercise 39(1): 44–51.

32.

Ministerio del Deporte de Chile (2015) Informe Encuesta Nacional de Hábitos y Actividad Física en la población de más de 18 años [Report on the National Survey of Habits and Physical Activity in the population over 18 years of age]. Technical report, Ministerio del Deporte de Chile [In Spanish]. Available at: http://www.mindep.cl/wp-content/uploads/2016/07/PRESENTACION-ENCUESTA-HABITOS-2015.pdf

33.

Ohl

(2000) Are social classes still relevant to analyse sports groupings in “postmodern” society? An analysis referring to p. bourdieu’s theory. Scandinavian Journal of Medicine & Science in Sports 10(3): 146–155.

34.

Ollivier

(2008) Modes of openness to cultural diversity: Humanist, populist, practical, and indifferent. Poetics 36(2–3): 120–147.

35.

Peterson

Kern

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

36.

Sánchez García

(2008) Habitus y clase social en Bourdieu: una aplicación empírica en el campo de los deportes de combate [Habitus and social class in Bourdieu: an empirical application in the field of combat sports]. Papers: Revista de Sociologia 89: 103–125. [In Spanish.]

37.

Sánchez García

Moscoso Sánchez

(2015) “How can one be a sports fan?” La contribución de Pierrue Bourdieu al estudio social del deporte [“How can one be a sports fan?” Pierrue Bourdieu’s contribution to the social study of sport]. EMPIRIA. Revista de Metodología de las Ciencias Sociales 30: 161–180. [In Spanish.]

38.

Stempel

(2005) Adult participation sports as cultural capital: A test of Bourdieu’s theory of the field of sports. International Review for the Sociology of Sport 40(4): 411–432.

39.

Stempel

(2019) Sport as high culture in the USA. International Review for the Sociology of Sport 55(8): 1167–1191.

40.

Van Tuyckom

Scheerder

(2010) Sport for all? Insight into stratification and compensation mechanisms of sporting activity in the 27 European Union member states. Sport, Education and Society 15(4): 495–512.

41.

Van Tuyckom

Van de Velde

Bracke

(2012) Does country-context matter? A cross-national analysis of gender and leisure time physical inactivity in Europe. The European Journal of Public Health 23(3): 452–457.

42.

Veenstra

(2007) Who the heck is don bradman? Sport culture and social class in British Columbia, Canada. Canadian Review of Sociology/Revue Canadienne de Sociologie 44(3): 319–343.

43.

Vermunt

(2010) Latent class modeling with covariates: Two improved three-step approaches. Political Analysis 18(4): 450–469.

44.

Warde

Wright

Gayo-Cal

(2007) Understanding cultural omnivorousness: Or, the myth of the cultural omnivore. Cultural Sociology 1(2): 143–164.

45.

Widdop

Cutts

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

46.

Widdop

Cutts

Jarvie

(2016) Omnivorousness in sport: The importance of social capital and networks. International Review for the Sociology of Sport 51(5): 596–616.

47.

Willey

Paik

Sacco

, et al. (2010) Social determinants of physical inactivity in the northern manhattan study (nomas). Journal of Community Health 35(6): 602–608.