Exploring the minority–majority gap in sport participation: different patterns for boys and girls?

Abstract

This paper examines the participation of minority and majority youth in organized sport based on the Young in Oslo 2015 survey (response rate: 72%, N = 9.774). Four possible explanations of the minority–majority gap are introduced: (i) culture, (ii) religion, (iii) discrimination/racism and (iv) class and socio-economic resources. Class and religion are further examined in the paper. Initially, among the boys, there are no differences between the minority and majority groups. When taking socio-economic resources into account, the minority boys have a slightly higher participation rate than the majority boys. Among the girls, the likelihood of participating in sport clubs is considerably lower for those with a minority background than for those with a majority background. The analyses show that socio-economic resources have an effect on both girls and boys and that religious denomination also explains, to some extent, the minority–majority gap that exists among girls.

Introduction

The integration potential of organized sport has been highlighted by sport associations and in governmental papers (Elling, De Knop, and Knoppers 2001; European Commision 2007; Krouwel et al. 2006; Meld St 2011–2012; Spaaij 2015). The refrain is that sport is international, does not require strong language abilities and follows rules that are not difficult to manage.

However, despite the alleged integration benefits of sport, it has been reported that young people with immigrant parents participate in sports organizations to a lesser extent than do young people from the majority population. The purpose of this study is to enable an understanding of the differences in sports participation between young people from minority and majority populations. The first aim is to determine whether religious commitment, class background and financial resources in the home contribute to understanding these differences. The second is to determine whether gender differences exist in such relationships.

The minority–majority gap in sport participation has been illustrated in studies that have been conducted in several European countries: the Netherlands (Elling and Claringbould 2005), the United Kingdom (UK) (Rowe and Champion 2000), Germany (Burrmann, Mutz, and Zender 2015), Switzerland (Adler-Zwahlen, Nagel, and Schlesinger 2017), Finland (Zacheus 2010), Sweden (Elofsson et al. 2014; Fundberg 2012; Riksidrottsförbundet 2010), Denmark (Agergaard, Michelsen la Cour, and Gregersen 2016; Nielsen et al. 2013) and Norway (Friberg 2005; Myrli and Mehus 2015; Seippel, Strandbu, and Sletten 2011; Strandbu and Sletten 2006). These studies have also demonstrated larger gender differences in sports participation in the minority population than among the majority. Minority women are less involved in sport than are women in the majority population. However, the data for men are less straightforward. Some studies have shown that for minority men, the participation rate is slightly higher or at the same level as among the majority population with regard to participation in sport clubs, ethnic-based clubs and unorganized exercise (Elling, De Knop, and Knoppers 2001; Riksidrottsförbundet 2002). In short, the existing studies demonstrate the relevance of examining the role of gender in sports participation among minority youth.

The research on ethnicity and sport has been divided by gender. Racism has primarily been studied as a factor affecting boys – particularly black boys (Massao and Fasting 2014) – while the studies on girls have focused predominantly on culture and religion as barriers to participation (Benn, Pfister, and Jawad 2011; Toffoletti and Palmer 2015). Few studies have addressed both boys and girls in relation to sport and have compared their participation. Based on an extensive survey study titled Young in Oslo 2015, this paper analyses minority and majority youth’s participation in sport and investigates gender differences in the factors that might promote and inhibit their participation.

To provide a background for our analyses, we first outline previous attempts to explain the minority–majority gap, followed by notes on the Norwegian organization of youth sport. We then present the procedure and methods used in the study. The findings are then introduced in a descriptive section on participation rates, followed by an analysis of whether class and socio-economic background contribute to explaining the minority–majority gap in sport participation. The discussion centres primarily on two of the explanations presented in the following section.

Explanations of the minority–majority gap

Explanations for the differences between minority and majority youth in previous research could be grouped into four main categories (Strandbu 2006); (1) culture, (2) religion (primarily Islam), (3) discrimination and racism and (4) class and socio-economic resources. The first two explanations have been well addressed in previous research, especially regarding girls, while the latter two have been investigated to a lesser extent. Some of these explanations apply to both girls and boys, while others are seen as relevant to girls only.

Culture and traditions

Cultural differences in the evaluation of sport participation in childhood and adolescence, as well as the gendered character of sport, are among the most notable concerns in the European literature on the sport participation of girls with an immigrant background. In various studies of sports participation among Asian minorities in the UK, the girl’s familial responsibilities (such as household chores, caring for younger siblings, and expectations of spending time with the family) have been described as one cause of poor sport participation (Kay 2006; Rowe and Champion 2000). The family of the immigrant child has primarily been considered a restricting factor (Dagkas, Benn, and Jawad 2011; Hamzeh and Oliver 2012; Kay 2006). A study of young Norwegians with an immigrant background concluded that some of the parents held the view that ‘young women should ideally be family-oriented and spend a lot of time helping their mothers cook, looking after younger siblings or studying’ (Walseth 2006, 91). It has been suggested that family life is more restrictive for girls; however, in line with other studies showing a strong emphasis on family life in minority populations (Østberg 2003; Prieur 2002), this might be a barrier for boys as well. Time spent on sport could minimize the amount of time spent with the family or on schoolwork. In addition, the parents of minority youth seldom have limited childhood experiences of organized sport.

The perception of gendered identities as being in conflict with sport participation is another issue that arises in studies on the cultural differences in sport participation. Sport has traditionally been a male activity in many countries (Hargreaves 1994), and especially in several of the countries in which the immigrant parents spent their childhood. In her study of minority girls with Turkish backgrounds in Germany, Pfister (2000) claims that ‘traditional norms and ideals and their appropriation by individual girls and women lead to a situation in which sport and femininity are incompatible from a Turkish perspective’ (Pfister 2000, 510). It is often suggested that gender roles are more traditional among non-Western immigrants than among the majority population (Prieur 2002), with parents’ restrictions being highlighted as an explanation for lower participation in sport among girls with immigrant parents. The refrain is that ‘girls are very restricted in their behaviour mainly because parents fear that their daughters will be badly influenced through the western way of life’ (De Knop et al. 1996, 151). It follows that cultural differences should have greater relevance for girls than for boys and that domestic work, time spent with the family and parental restrictions are relevant factors in research on ethnic differences in sports participation.

The cultures of minority populations have received the most attention in studies on the cultural explanations of the minority–majority gap in sport participation. We would like to emphasize that the majority populations culture of sport needs to be presented to provide a comprehensive picture. We have included a presentation of the Norwegian culture of sport in this article.

Religion

Religion is a part of culture and is one of the most notable concerns in the European literature on the sport participation of girls with immigrant backgrounds. It has been suggested that the Muslim ideal of gender segregation is a main factor restricting young Muslim women’s participation in sports (De Knop et al. 1996; Pfister 2000). Segregation of the sexes is legitimized by reference to the Koran and Hadiths. It is required among some Muslims based on the explanation that women can create fitna (chaos or temptation) by engaging in sports that could be sexually exciting for men to watch (Walseth and Fasting 2003).

A general feature of most research on Islam and sport is that it does not discuss in depth the various additional factors contributing to sport participation – a problem that is not confined to sport research. As Mabro and El-Sohl (1994, 19) have pointed out, there is a tendency to choose Islam as the only ‘cultural determinant’ and, thus, to overlook other important factors, such as class, national differences and cultural traditions. In the case of sports participation (as in other arenas), there may be restrictions on women that do not relate directly to religion. Because religion occupies various degrees of centrality in people’s lives, it is reasonable to explore the different relevance of prescribed rules and, therefore, the varying degrees and forms of religious involvement. In addition, the role of Islam should be addressed in relation to other factors.

Walseth and Strandbu (2014) have underlined the need to specify the notions of culture and religions used in studies on sport participation of minorities, and they distinguish between culture and religion as (i) restricting factors, (ii) as embodied dispositions for action and (iii) as the basis for reflexive praxis. Hence, the relevance of both religion and culture could work in different ways.

Racism and exclusion

For ethnic minorities, exposure to racism, discrimination and marginalization is a problem that has been revealed in research conducted in several European countries (Andersson 2007; Hylton 2009; Massao and Fasting 2014; van Sterkenburg and Knoppers 2004, 2012). Direct and violent acts, such as the use of racial slurs, are described as predominantly affecting men, while more subtle forms of discrimination are experienced by women and men (Massao and Fasting 2014). South Asian women, among others, tend to be associated with the stereotypes of the Asian woman as passive and are, consequently, not encouraged to participate in sports (Samie 2013).

An interview study of athletes, leaders and coaches in Norway found that the forms of racialization varied between sports (Andersson 2007). In a study of black Norwegian elite athletes’ experiences of marginalization and racism, few examples of direct exposure to racism were described, and those that existed were directed towards men (Massao and Fasting 2010). The female athletes experienced evaluations as secondary athletes. Massao underlines that her informants described these messages as conveyed in subtle ways, and this made it difficult for the female athletes to pinpoint the messages as direct discrimination (Massao and Fasting 2014, 345).

These findings point away from openly racist insults being the most pressing factors keeping minority youngsters out of sport. The more important factors are probably the various forms of racialization and differing degrees of welcoming and inclusive attitudes among other participants and leaders (Lidén 2016). Based on these studies, one can assume that discrimination and exclusion can influence the rate of organized sport participation and that exposure to racism and racialization intersects with gender.

Social class background and socio-economic resources

Sports participation tends to be more common in the upper- and middle-classes, as well as among young people with highly educated parents (Ohl 2000; Wilson 2002). To some extent, this finding also holds for young Norwegians (Strandbu, Gulløy et al. 2017) and probably more so in the larger cities in which more marked social inequality exists (Andersen and Bakken 2015).

Donnelly suggests that even though sport has often been considered the great social leveller, it is more appropriate to view it as ‘contested terrain, providing fertile ground for the production and reproduction of social inequality’ (Donnelly 1996, 237). Class can influence participation in organized sport in at least two ways. First, the economic costs associated with sport participation make it more easily available for upper- and middle-class teenagers. Second, in class cultures, sport might be more highly valued among some class factions (Aarseth 2014). Both of these factors might contribute to explaining the minority–majority gap. The minority populations in Norway and other European countries are over-represented in the lower socio-economic groups (Hermansen 2016). In their study, Dagkas and Hunter (2015) found that compared to working-class immigrants, middle-class immigrants put more weight on their children’s sport participation. Research on the role of class and socio-economic resources in explaining the differences in the participation rates of minority and majority groups has, with some exceptions (Dagkas and Hunter 2015; Nielsen et al. 2013; Strandbu and Sletten 2006), been sparse. With regard to both minority and majority youth, there remains a need to examine the relationship between sports participation and socio-economic factors.

Research questions

The purpose of this study is to examine the differences between young people from the minority and majority populations with regard to their participation in organized sport. In the analyses, we primarily address two of the four explanations presented above: class and religion. With a particular focus on gender, our aim is to examine the possibility that socio-economic background and religious commitment and affiliation can contribute to explaining the observed differences. Thus, the main question raised in this article is the following: To what degree are socio-economic background and religion relevant factors for explaining the minority–majority gap in sport participation? In addition, we want to explore whether minority background and socio-economic and religious factors are equally relevant for girls and boys in explaining participation in club-organized sport.

Norwegian organization of sport

To put the study data in context, it is necessary to describe the Norwegian situation. In Norway, competitive sports for children and youth are organized by voluntary sports clubs, and exercise takes place during leisure time. The catchment area of most sports clubs is the local community, which means that children and youth often participate in the same sports clubs as their peers from the same geographical areas and the same schools. There are approximately 12,000 Norwegian voluntary sports organizations, all of which are under the umbrella of the Norwegian Olympic and Paralympic Committee and Confederation of Sports (NIF). In a country with a population of 5 million, the NIF had more than 2 million registered memberships in 2015 (one person could be registered as a member of more than one club) (NIF 2016).

With regard to the development of sport and its voluntary organizations, Norway shares numerous characteristics with other Western nations (Goksøyr 1992; Olstad 1987; Seippel 2002). However, according to Seippel (2002), there are two distinctive features: There is a close relationship between voluntary sport organizations and the state, and the geography and climate have made certain sports, such skiing and skating, particularly popular in Norway. Football (soccer) is, nevertheless, the most popular sport among Norwegian teenagers, with more than 120,000 registered memberships in the 13–19 age group. Handball follows football in popularity (32,000 memberships) and skiing ranks third (25,000 memberships). Football is the most popular sport among both boys and girls. The second most popular sport among girls is handball, followed by skiing, whereas the order is the opposite among boys (NIF 2015).

Sport is, without doubt, the most popular organized leisure activity among young people in Norway. A recent national representative study showed that 93% of Norwegian teenagers are, or have previously been, active in a sports club (Bakken 2017). The participation declines during the teenage years and more so among girls than boys. Among the eighth graders (13-14 year olds), 72% of the boys and 69% of the girls reported that they had trained in a sports club at least once in the last month. Among the pupils in their final year of upper secondary school (18–19-year-olds), the numbers were 42% for the boys and 28% for the girls.

In the case of Norway, three characteristics of the culture of sport – which is, to a large extent, shared among Scandinavian countries – should be noted as especially relevant when studying the relationship between minority and majority populations and participation in sport. First, a large proportion of Norway’s children and youth participate in sports clubs (Bakken 2017; Seippel, Strandbu, and Sletten 2011), and there are relatively minor class differences in sport participation compared to many of the countries from which the parents of the minority participants originated (Strandbu, Gulløy, et al. 2017). Membership in a sports club might not be taken for granted among parents who were raised in countries in which ‘sport in clubs and centres is something for rich people’.¹

A second point is that sport participation could be described as an extension of family life in the Norwegian context (Archetti 2003), in which parents’ voluntary work is not only expected but is also necessary for the existence of the clubs. Introducing children to sports at an early age is perceived as a sign of good parenting (Stefansen, Smette, and Strandbu 2016). The important roles played by family in introducing children to sport has been illustrated in studies from other European and North American contexts (Fredricks and Eccles 2005; Wheeler 2012). Although it is not sufficiently documented, some studies point in the direction of sport as a family project being especially widespread in Scandinavian countries and not solely in the middle-class (Hertting 2007; Stefansen, Smette, and Strandbu 2016; Toftegaard-Støckel et al. 2011). Hence, families are important for Scandinavian children’s sport participation.

The third relevant cultural factor is the gendered character of sport. In Norway, as well as in many other countries, sport has traditionally been a male activity (Hargreaves 1994). Gender differences in terms of the number of members in sport organizations have been relatively stable since the mid-1980s. However, during this period, more girls have taken up sports that were previously dominated by boys (Fasting et al. 2008). One possible explanation for the high participation rate among women in sport is the gender equality tradition in Norway and the other Nordic countries, and another may be the development of policies that aim to encourage female sports participation (Fasting 2003). The high participation rates among girls, the fact that football is the most popular sport among girls, and the results of some studies on gender identity and sport (Broch 2016; Strandbu and Hegna 2006) indicate that the tradition of sport as a masculine activity is less clear cut for today’s Norwegian teenagers. Women’s participation in some traditional male-dominated sports such as ski-jumping and wrestling is nevertheless often trivialized (Lippe 2001; Sisjord 2009).

Methods and measures

We used data from the Young in Oslo 2015 study, the details of which have previously been reported (Andersen and Bakken 2015). The purpose of this study was to map the living conditions of adolescents in Oslo. It included self-reported information about young people’s lives, including their participation in sports, family backgrounds, religious affiliations and religious beliefs. The information was collected using a school-based survey administered during 2015. The survey was administered in 30 upper secondary schools, including all the public schools and eight of the 11 private schools. The response rate was 72%, and the sample represents approximately 62% of the total Oslo population of 16–18-year-olds. Parents and students were informed in advance about the purpose of the study, including the fact that participation was voluntary. The ethical aspects of the study were approved by the Norwegian Social Science Data Services. We excluded individuals who were older than 19 years and those with missing values on at least one of the variables used in the analyses. The total sample was 9774, of which approximately a third had minority backgrounds. The average age was 17.06 years, and girls made up 53% of the sample.

Dependent variable

We measured participation in sports clubs by asking ‘How often do you train or take part in competitions in a sports club?’ We provided the following six response options: ‘never’, ‘seldom’, ‘1–2 times a month’, ‘1–2 times a week’, ‘3–4 times a week’ and ‘more than five times a week’. We defined ‘active participants’ as those who exercise in a sports club at least ‘1–2 times a week’.

Independent variables

Minority/majority status was measured by asking where the respondent’s parents were born. Those who answered that both parents were born abroad were classified as having minority status, and of these, 63% were born in Norway. More than half of the minority youth had parents who were born in Pakistan, Somalia or Sri Lanka. The remaining parents were, for the most part, born in Asia, Africa or Eastern Europe. Those with one or both parents born in Noway were classified as having majority background.

The response options for our question about religious affiliation were ‘Islam’, ‘Christianity’, ‘other religions’ and ‘non-religious’. We also asked how much impact religion had on the respondents’ daily lives, for which the response options ranged from ‘Religion has no importance for how I live my life’ (0) to ‘Religion is very important’ (3) (see: Cotton, McGrady, and Rosenthal 2010).

To measure socio-economic background, we asked about parental education, the number of books at home and four items from the Family Affluence Scale (Currie et al. 2008). Based on the average scores of these variables, we constructed a single socio-economic score and each respondent was placed on a scale of 0–3 (Bakken, Sletten, and Frøyland 2016). In addition, we asked about each parent’s participation in the labour market, for which the response options were ‘not working’ (0), ‘part-time’ (1) and ‘full-time’ (2). We constructed an average score using information on both parents (scale 0–2).

Statistics

First, we describe the extent to which young people participate in sports clubs and whether this varies according to gender and minority status. Using chi-square tests, we tested whether the minority–majority differences within each gender were statistically significant. To examine whether the independent variables could explain the possible minority–majority gap, we conducted a series of logistic regression analyses using the xtlogit model (random intercepts only) in Stata version 13.1. A multilevel framework was used to capture the hierarchical structure of the sample (individuals at level one and schools at level two) to give more realistic standard errors (Raudenbush and Bryk 1992).

We calculated the bivariate regressions between each of the independent variables and the dependent variable. Due to the relatively large sample size, the significance level was set at p < 0.01. In a second model, we included socio-economic factors and family structure. In a third model, we controlled for religiosity and religious affiliation. Based on Model 3, we used a z-test to examine the interaction effects between all independent variables and gender. All regression analyses were conducted separately by gender. When comparing the bivariate and multivariate models, interpretation of the ORs or B coefficients in the logistic regression can be problematic because they can reflect the degree of unobserved heterogeneity in the models (Mood 2010). To overcome this problem, we rescaled the results of the xtlogit model to the same scale as the intercept-only model, using the ‘meresc’ command in Stata (cf. Hox 2010 chapter 6.5). We used binary logistic regression for all the dependent variables.

Results

Table 1 shows the descriptive statistics for all the variables broken down by gender and minority status. A total of 29.4% of all the respondents exercised in a sports club on a weekly basis. There was no statistically significant difference between minority boys (38%) and majority boys (40%), and both groups were much more active in sports than girls. The minority girls (12%) had significantly lower levels of activity compared to the majority girls (25%). The minority–majority gap among the girls can be considered large.

Table 1. Descriptive statistics for all variables used in the analysis.

	Majority boys	Minority boys	Sig	Majority girls	Minority girls	Sig	Total	(Std.dev)
Dependent variable
Participating in a sports club (%)	39.8	38.3	NS	24.5	12.4	*	29.4
Individual-level independent variables
Age	17.03	17.09	NS	17.05	17.15	*	17.06	(.91)
Socio-economic family background (0–3)	2.36	1.71	*	2.37	1.70	*	2.16	(.61)
Parental labour market status (0–2)	1.75	1.30	*	1.73	1.22	*	1.60	(.58)
Daily life impact of religion (0–3)	.47	1.51	*	.47	1.65	*	.82	(.98)
Religious affiliation
Non-religious (%)	55.8	15.5	*	47.6	12.0	*	40.2
Christianity (%)	37.2	17.1	*	47.8	17.7	*	35.3
Islam (%)	2.5	52.2	*	2.1	56.1	*	17.8
Other (%)	4.6	15.3	*	2.4	14.2	*	6.8
N=	3.240	1.338		3.490	1.676		9.774

Note:

NS = non-significant (p > 0.01).

* p < .01

Table 1 also shows that most independent variables vary between the minority and majority students (but not between girls and boys). The minority students have considerably fewer socio-economic resources than the majority, and their parents have become integrated into the labour market to a lesser degree than those of the majority students. The minority students are considerably more religious than the majority students, and more than half of the minority population consider themselves Muslims. Thus, these variables are relevant to include in the multivariate models.

Table 2 addresses the minority–majority gap in sports participation more systematically by presenting the results from the multilevel logistic regression models conducted separately by each gender. In Model 1, minority status was included as a dummy variable together with the respondent’s age. The results confirm the findings in Table 1 – no minority–majority gap among the boys and a large gap among the girls. In Model 2, we also controlled for socio-economic family background. Compared with Model 1, we observe in Model 2 that the minority–majority gap among the boys becomes positive, meaning that the minority boys have a slightly higher probability of attending a sports club than the majority boys with the same family background. Among the girls, the minority–majority gap becomes weaker; however, there is a large gap that cannot be explained by the inclusion of family background factors. When religion is introduced into the equation in Model 3, the gap among the girls is somewhat further reduced. Altogether, the included variables explain approximately half of the initial minority–majority gap in sports participation among the girls. Among the boys, the minority–majority gap again becomes close to zero when controlling for religion.

Table 2. Multilevel regression analysis of participation in sports clubs among upper secondary school students in Oslo, Norway N_individuals = 9.774, N_schools = 30.

	Boys									Girls									z-test of gender differences in Model 3 estimates
	Model 1			Model 2			Model 3			Model 1			Model 2			Model 3
	B	Se B	Sig p	B	Se B	Sig p	B	Se B	Sig p	B	Se B	Sig p	B	Se B	Sig p	B	Se B	Sig p	z-test	Sig
Minority status (1 = minority)	.012	.075	.878	.183	.082	*	.014	.095	NS	−.700	.092	***	−.418	.099	***	−.297	.118	*	2.05	*
Age	−.259	.035	***	−.245	.035	***	−.242	.035	***	−.327	.041	***	−.317	.040	***	−.319	.040	***	1.45	NS
Socio-economic background (0–3)				.364	.066	***	.380	.065	***				.419	.080	***	.402	.080	***	0.21	NS
Parental labour market status (0–2)				.011	.067	NS	.057	.069	NS				.207	.008	**	.178	.077	*	1.17	NS
Daily life impact of religion (0–3)							.204	.045	***							.100	.054	NS	1.48	NS
Religious affiliation (ref=non-religious)
Christianity							.333	.083	***							.045	.088	NS	2.38	*
Islam							.185	.141	NS							−.501	.180	**	3.00	**
Other							−.266	.147	NS							−.189	.186	NS	.32	NS
Intercept	3.911	.613	***	2.800	.639	***	2.399	.647	***	3.633	.701	***	2.835	.726	***	2.888	.735	***	.50	NS
N individuals	4.578			4.578			4.578			5.166			5.166			5.166
N schools		30			30			30			30			30			30
Fit statistics
Log likelihood	−2938			−2920			−2885			−2453			−2428			−2420

Notes:

b: rescaled logistic regression coefficients using the ‘meresc’ command in Stata. SE: standard error.

p: significance level:

* p < .05

** p < .01

*** p < .001

NS = non-significant.

Table 2 shows that the family’s socio-economic status has a major impact on sports participation among both girls and boys. Parental labour market status plays a minor role. Religion also seems to play an important role, but in a different way for boys and girls. Among the boys, those who are most active in sports are the ones who consider themselves either Muslims or Christians and among those for whom religion plays an important role in their daily lives. Among the girls, the importance of religion in their daily lives is not independently related to sports participation; however, Muslims have significantly lower levels of participation in sports compared to girls from other religions and those who are non-religious.

The tests of interaction effects based on Model 3 showed that minority status has a significantly higher impact among the girls and that the effects of religious affiliation differ between the boys and girls. The coefficients for family background and the daily impact of religion do not differ significantly between the boys and girls.

Discussion

We set out to explore the differences in sports participation between young people from minority and majority populations, aiming specifically to examine whether socio-economic background and religion explained the gender differences in the minority–majority gap. The analyses showed that, for minority girls, the likelihood of participating in sport clubs was considerably lower than the likelihood among majority youth in general and minority boys. Controlling for parental education, family financial situation and religious denomination reduced the difference in probability of sports participation between the minority and majority girls. However, as a significant effect of minority background remained after the control, these variables were not able to offer a comprehensive explanation of the minority–majority gap among the girls.

However, among the boys, there were initially no differences between the minority and the majority groups. When taking social background into account, the minority boys had a slightly higher participation rate than the majority boys. This indicates that the minority boys actually participate at a higher rate relative to their social backgrounds (ie the higher probability of low socio-economic status among minority boys creates a suppressor effect). However, when further controlling for religious affiliation and the daily life impact of religion, the difference between the minority and minority boys was no longer significant.

In general, the analyses supported the assumption that minority background affects boys’ and girls’ sports participation differently. Moreover, the analysis indicated that a lower level of socio-economic resources in minority families acts as a barrier to the participation of both boys and girls. Nevertheless, as minority boys generally participate at the same rate as majority boys, the financial barrier per se may be surmountable (at least in the sports in which the boys participate). The finding that socio-economic background is able to explain the minority–majority gap among girls to only a limited extent points towards more complex explanations. The family’s financial resources may, for example, add up with or intersect with the explanations related to cultural differences or discrimination/racism. The analyses in this article indicated that being Muslim, in particular, intersects with gender in explaining sports participation among minority youth. Whereas affiliation with Islam reduces the likelihood of sports participation among minority girls, this is not the case for boys. Contrary to our expectations, the level of centrality that religion occupied in the girls’ lives had no independent effect on their sports participation.

Limitations

Although the present study has several strengths – namely, that it was based on a general population study with a high response rate – some limitations should be mentioned before the discussion of the results. Whenever young people are the subjects of a study, the discussion often turns to characteristics of minority youth compared to majority youth. We could just as well have addressed the institutional level by examining, for instance, the characteristics of clubs that do manage to recruit young people with immigrant backgrounds compared to those that do not. In our discussion, we have taken into account available information about the Norwegian culture and organization of sport to compensate for the lack of information about organizations in our data.

To address another limitation, we have studied characteristics that were suggested as relevant to sports participation; however, we do not know whether it is the measured characteristic themselves that causes the effects. This might be problematic when interpreting the analyses of religion. Agreeing with Dagkas and Hunter (2015, 548), we see ‘the problem of deciphering religious requirements from pseudo-religious culturally embedded practices’. We do not know whether it is religion in itself or other unmeasured characteristics of the Muslims in the study that have generated the differences. This possibility has been tested to some extent, but other factors that were not included in the analyses might also contribute to explaining the differences. This should be kept in mind during the following discussion.

Class and financial resources

Among both boys and girls, the introduction of class and financial resources in the second model of analyses changed the initial connections between majority/minority background and sport participation. To some extent, the under-representation of minority girls was found to be accounted for by socio-economic resources in the family, which are conditions that also influence the majority girls’ probability of joining sport clubs. When taking social background into account in the second model, the minority boys had a slightly higher participation rate than the majority boys.

Class and socio-economic resources have in some other studies explained differences in participation rates between minority and majority groups (Dagkas and Hunter 2015; Nielsen et al. 2013; Strandbu and Sletten 2006). The intersection of class and minority status could be relevant both because the economic demands of participation make it more easily available for upper- and middle-class teenagers and also because sport might hold various positions in class cultures among different class fractions. Organized leisure activities may be less highly valued among working-class than among middle-class minorities (Dagkas and Hunter 2015).

The relevance of socio-economic factors resonates well with an interview study of coaches and leaders from Oslo, who suggested that the participation rates of minority youth in organized sports could be affected by their families’ poor finances combined with the lack of a tradition that would allow the families to place a high value on expensive leisure activities for children and especially girls (Strandbu 2006). Given that these factors were found to be of importance in the Norwegian context, in which class differences with regard to participation rates are still relatively low (Strandbu, Gulløy, et al. 2017), they should be taken into account in studies conducted in other European countries in which the class-related participation gap may be larger.

However, socio-economic resources cannot entirely explain the lower probability of minority girls’ participation in sports and contributed to only a small change among boys, thereby indicating that other explanations should be taken into account.

Religion

It has been suggested that Islam is one of the main explanations for the under-representation of girls with immigrant backgrounds in sports organizations. We did, in fact, find that Muslim girls had a significantly lower probability of being members of sport clubs. Among the boys, being Muslim had no significance.

Among the boys, the somewhat surprising result was the significance of having a Christian affiliation. The boys who regarded themselves as Christian had a significantly higher probability of participating in sports than the boys who regarded themselves as non-religious. Regarding themselves as Muslims or having other religious affiliations had no such significance. This finding, which shows that the relevance of religion is not restricted solely to minority youth, should be highlighted in future research.

The influence of Islam among girls was in line with the results of other studies (Pfister 2011; Walseth and Strandbu 2014). For strictly practising Muslim girls, gender segregation is a precondition for their participation in sport (Strandbu 2005). Although Islam is described as a religion with a positive evaluation of female sport participation (Walseth and Fasting 2003), the way in which sport is organized in Western societies is in conflict with the gender segregation norms to which some Muslims adhere.

We also expected degrees of religious involvement (the impact of religion on daily life) to be correlated with sports participation. However, the variables measuring religious commitment contributed to only a small reduction in the effect of minority background. Hence, being Muslim did not affect only the devout Muslim girls, whom we would assume to have a higher probability of following strict religious requirements and presumably avoiding participation in a sports club, where their activities could be seen by men. It should be mentioned that being a devout Muslim does not necessarily include adherence to segregation rules. Conversely, strict modesty requirements do not necessarily follow from strong religious devoutness (Strandbu 2005). A report based on the same data-set as the one used in this article showed that large differences in sport participation depended on parents’ countries of origin, and varied among predominantly Muslim countries, such as Pakistan and Turkey (Ødegård, Bakken, and Strandbu 2016). As was found in this article, after taking religion into account, a large unexplained difference between minority and majority girls remains, indicating that other aspects of minority girls’ life situations may be as relevant as religion.

With regard to girls, our discussions so far reveal that other explanations need to be taken into account in future studies. We suggest some of the characteristics of the young girls with immigrant backgrounds and recruitment channels that should be addressed.

Gendered identities

Two qualitative interview studies on young girls with immigrant parents in Oslo suggest that bashfulness and embodied identities may affect sport participation among young girls with immigrant parents (Strandbu 2005; Walseth and Strandbu 2014).

As opposed to girls from the majority population, who are strongly encouraged to take up sport (Stefansen, Smette, and Strandbu 2016), immigrant parents may tend to be indifferent to, or have a negative attitude towards, their daughters’ sports participation. In other words, aspects of gender culture that are not necessarily related to Islam could imply that many girls with immigrant parents are simply not interested in sport. A search for more subtle forms of control or simply differences in motivations to engage in organized sports related to gender identities may prove to be a productive area of focus in further research.

Some general thoughts about gender and exile may be relevant to such an attempt. The consideration of women as the bearer of tradition seems to be widespread among groups other than Muslim exiles. Several studies on gender differences in immigrant adaptation have reported stronger expectations of traditional behaviour for daughters (Dion and Dion 2001; Prieur 2002). According to Moghadam (1994), women in many exile situations serve as the markers of boundaries for ethnic, religious and national groups, as well as families. The central role that women play in these groups’ further existence is suggested as the rationale behind this praxis. Women reproduce a group biologically, and more importantly, they reproduce it socially and culturally by having the primary responsibility for the socialization of children (Thorbjørnsrud 2002). In line with this argument, it is more important for young girls to be the bearers of the ethnic group’s traditions than it is for young boys. Sport could be one of the arenas that minority girls avoid to assure others that they are not transgressors but are passing on the notions of acceptable femininity (Strandbu 2005).

Recruitment channels

Like other leisure activities, sports participation has certain identity stamps (Donnelly and Young 1988), one of which is its gendered character. An American study of high school students concluded that with regard to peer relations and youth culture, sport holds a much stronger position among boys than among girls (Shakib, et al. 2011). Even if sport has a somewhat weaker gendered character in the Norwegian context (Broch 2014; Strandbu and Hegna 2006), it is reasonable to suggest that sport plays a more central role in relationships among boys than it does among girls. Hence, friends may not be as strong as recruitment channel for girls as they are for boys.

The other important recruitment channel is parents. A Norwegian study revealed that recruitment into organized sport was primarily family based, whereas recruitment into unorganized sport was more likely to be based upon friends (Skille 2005). Several other studies have shown parents to be initiators of sport participation and motivators for further activity (Stefansen, Smette, and Strandbu 2016; Strandbu, Stefansen, et al. 2017; Wheeler 2012).

It seems reasonable to suggest, therefore, that boys in both the majority and minority populations are recruited into sport through friends to a greater extent than girls are. Hence, the lack of a tradition of organized sport for children (especially girls) within the parents’ country of origin would likely affect girls more than boys. Moreover, if women in many exile situations are markers of cultural traditions (Moghadam 1994), parents may not be enthusiastic about transgressing daughters and may not be as keen about girls’ sport participation as many parents in the majority population would be.

Concluding remarks

Our study showed the importance of taking gender into account when analysing differences in the sports participation of minority and majority youth. The minority girls were significantly less likely to participate in a sports club than the majority girls; however, no such difference existed between the minority and majority boys. The other major finding in this study is the importance of class and socio-economic resources in explaining the majority–minority gap in sport participation among both the boys and girls in the minority populations.

The minority girls were significantly less active than the majority girls, even after controlling for parental education, the family’s financial situation and religious denomination. This indicates that other factors should be taken into account. We suggest that the recruitment channels, subtle barriers related to identity issues, cultural stamps of the organization and the role of sport in family cultures should be addressed in future studies.

In addition, we suggest that future studies address physical activity in other arenas. Focusing on organized sports could ‘disguise [the] great diversity within the group of ethnic minorities’ (Dagkas and Hunter 2015, 548). Conversely, organized sport is often ascribed several valuable outcomes for young people and receives immense public support. Hence, the disproportionate rates of participation in organized sports are an intrusive challenge for sport organizations.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Lale Agün (cited in Pfister 2000, 505) sums up the situation in Turkey.