Regular participation in moderate to vigorous PA is imperative for good health. It is therefore important to understand how to facilitate participation. There is much research on the determinants of general levels of PA. However there are many different forms and contexts of PA participation and we need to understand what influences these behaviours so that strategies can be developed to increase participation in each of these specific types of PA. Club sport participation during adolescence is one context of PA participation that can positively influence PA in adulthood [59, 60], and lay the foundation for fundamental motor skills  and competence  as well as providing the benefits of social interactions  and enhancing self-esteem . This study investigated adolescent girls in two age ranges living in two types of geographical location, and found that club sport participation was a popular form of leisure time PA, particularly among younger girls and particularly in rural and regional areas. The study further investigated the relationship between club sport and household- and area-level characteristics including indicators of SES, access to facilities, and support by family and peers.
Little is known about how family structure influences PA behaviours. It has been suggested that the relationship between family structure (single v double v stepfamily) and children’s activities appears to be stronger for sedentary activities than for physical activities . Furthermore, children in two-parent families reportedly have more opportunities to engage not only in activities as individuals, but also in joint activities with their parents . In our study, sports club members were more likely to live with two parents, with or without siblings, but this applied predominantly to non-metropolitan adolescents and to the Year 7 cohort, and not to metropolitan adolescents or to the Year 11 cohort.
In this study ethnicity and religion were significantly associated with sports club membership. Highest rates of club membership were reported amongst those with family born in Australia and who spoke English only. Furthermore, sports club members were most likely to indicate being Christian or of no religious faith. A systematic review of correlates of physical activity reported that for adolescents ethnicity was consistently related, whereby those of ethnic minority groups were less active than others . It may be that different ethnic groups are associated with different levels of family support, which is a key determinant of PA behaviour amongst adolescents . Broadly, ethnicity and religion are social determinants of health that affect adolescents . Ethnic identity is a known risk factor for health, however the relationship between health and religion or spirituality is not well defined .
SES is often reported as a key determinant of PA behaviour and health in general [7, 14, 47]. We found that sports club membership was predominantly associated with relatively affluent Australian-born and acculturated families living in higher SES localities. A systematic review summarised the findings of people with lower SES backgrounds being less active . They concluded that as long as the activities require financial payment there is a barrier for persons with lower SES, but also that there may be many other related factors within the community, including: having fewer and poorer quality recreational areas and longer distance to PA facilities; communities perceived as being less safe; a general lack of time; and lower educational levels .
We also found that the evidence of association between SES and club membership was stronger in metropolitan than regional/rural settings. This was particularly so at an area level, with club membership not significantly associated with SEIFA IRSAD for regional/rural adolescent girls. This may reflect the smaller scale of rural and regional communities, which makes access to facilities less dependent on SES of home postcode, as well as lower costs of participation in PA and more limited alternative recreational options in rural and regional areas . Stalsberg and Pedersen (2010) also found that almost half of the papers they reviewed reported either a negative or no relationship between SES and level of PA, which may indicate that the issue of SES differences is only partly or periodically relevant . Two recent studies found that people with low SES were just as likely to participate in sport and PA [16, 67]. Ball and colleagues (2009) suggested that, given the well documented inverse relationship of SES with PA levels in adult samples, such disparities may only emerge after adolescence .
Two other commonly reported demographic measures of SES are the employment status and educational levels of parents or caregivers. We found that these factors were significantly associated with club membership for metropolitan adolescents and for the Year 7 cohort, but not for regional/rural students or the Year 11 cohort. It has been similarly reported that adolescents with one or two unemployed parents were less likely to participate in sports that those with two employed parents . A review of participation in PA for children and adolescents reported that mothers’ education level was a factor affecting participation for their children, as was family income . However Ball and colleagues (2009) did not find significant evidence of influence of maternal education either cross-sectionally or longitudinally predictive of children’s PA behaviours .
Proximity to facilities is commonly reported to be positively associated with participation in PA [14, 31, 33, 69]. However Grow and colleagues (2008) investigated facility location in more detail and concluded that for adolescents, proximity was not related to frequency of use of indoor recreation facilities, basketball courts and other playing fields/courts . Perceived availability of sports facilities and parks has been shown to be significantly associated with engagement of adolescents in sports , however the agreement between objective and perceived availability of parks, sports facilities, bicycle lanes and sidewalks was low . No associations were found between objectively assessed availability of sports facilities and parks and PA . In a US study of adolescent girls, perceptions about both individual facilities and the total number of facilities were associated with increased PA . In an Australian study, the environment – presence of public open spaces in the neighbourhood – played a smaller role than did psychosocial factors . However, a recent systematic review of reviews of environmental correlates of PA amongst children and adolescents found that physical activity was more consistently related to school and neighbourhood characteristics than to interpersonal and societal environments . As defined in the present study, access encompassed both knowledge regarding the existence and location of facilities and perceptions of their ability to get there, either by themselves or with the assistance of an adult for transport. We found that in general, club members reported significantly better access to club based activities, but this was not the case for geographically widespread facilities/venues such as beach, bike lanes, lakes, rivers and creeks, parks and playgrounds. Level of access was a more significant determinant for the Year 7 cohort.
Many studies confirm the importance of social support from parents regarding adolescents’ PA [22, 23, 72]. This is especially so in younger years and regardless of the child’s gender , but when a relationship is found between the mother’s support and their child’s PA, it is more often associated with girls . However this correlation is not always reported. In a study of adolescents and young adults there was no significant differences in perceived parental support relating to PA behaviour [21, 24, 26]. Paternal care for staying fit and exercising was, however, related to increased MVPA for older females . In our study family support was a significant predictor of club membership. This support included travel, encouragement, playing with their girls, watching, and praising their involvement in sport and physical activity. Family support was a significant factor for both Year 7 and Year 11 girls. This is in contrast to another Australian study which found family support was a positive influence for Year 10’s but not for the younger Year 7’s . It has been stated that this support differs across adolescence, with peers having greater influence on older adolescents when they are more autonomous and not as reliant on parents for travel . These researchers also found that younger adolescents receive more praise and joint participation from parents .
Many studies have found the importance of peer influence on out-of school PA behaviours , or general MVPA levels . In our study, peer support was significantly related to club membership. Furthermore, we found that negative peer interaction i.e. teasing for not being good at the activity was not a significant barrier to club membership.
With regard to mediation of the relationships between measures of SES (SEIFA IRSAD of home postcode, education level and employment status of parents/caregivers) and club membership, we found that access to facilities, family support and peer support were all mediators of the associations between particular aspects of SES and club membership. The most consistent and by far the strongest mediator, across the three dimensions of SES we examined, was the interpersonal factor of family support. That is, the level of family support is the most crucial channel through which the effects of SES influenced the adolescent girls’ participation in club sport. With regard to the other interpersonal factor we examined, peer support, we found that the adolescent girls’ perceptions of peer support were more influenced by the employment status of their parents/caregivers than by either parental education level or neighbourhood SES, as measured by postcode SEIFA IRSAD. This may be because of the more pervasive and immediate effects of parental employment status on adolescents’ daily lives. Be that as it may, peer support partially mediated only the link between parental employment status and club membership. Access to facilities, whilst it has an interpersonal aspect to the extent that adolescents depend on parents/caregivers for transport, is primarily an environmental determinant in terms of geography and proximity to facilities. As such, it was found to be an important mediator for neighbourhood SES, as measured by postcode SEIFA IRSAD, but was not a significant mediator for household-based measures of SES. In comparison, Cerin and Leslie (2008) also reported that differences in social support, as well as perceived benefits and self-efficacy, were largely responsible for the observed individual- and area-level SES differences in regular participation in leisure-time PA . In contrast, in the Trial of Activity in Adolescent Girls (TAAG) study, environmental factors appeared to be the strongest mediators (compared to behavioural and psychosocial) between the intervention and the activity level . The environmental factors mainly related to issues around transport to and from activities . The focus of the TAAG study more on the school and community than the home environment may explain some of the differences between study results. Furthermore, the issues reported around access may also be related to lack of support from family in relation to travel.
With regard to moderation of effects, we found stronger relationships in metropolitan areas than in rural/regional areas between SES (both area- and household-level indicators) and sports club membership. It is known that postcodes in rural areas have more heterogeneous SES profiles than those in urban areas and it may be that rural/regional communities have more heterogeneous SES profiles than do metropolitan (suburban) communities at a similar ‘social scale’ (i.e. the scale over which interactions between adolescents take place), and hence that in rural/regional communities behaviours such as sports club membership are likely to be influenced by groups (peers, peers’ parents, school communities) with a broader range of SES characteristics. There was also some evidence that the relationships are stronger at Year 7 level than at Year 11 level. This too may be explained by increased autonomy and mobility, and consequential exposure to a wider range of influences, in later adolescence.
In summary, we found stronger relationships between sports club membership and SES for the metropolitan rather than rural females. These relationships were mediated mostly by family support and less by access and peer support. These results concur with very recent literature which reports that the family environment, including provision of equipment, financial, logistic and emotional support and parental modelling are positively associated with sport participation .
Assumptions and limitations
The mediation models in this study, as with all such models, are framed in causal terms, with the intervening variables — access to facilities, family support and peer support — mediating the effect of SES on sports club membership. In general one should be cautious about attributing causation in cross-sectional studies. However, in this study the converse — that sports club membership influences SES — is considered unlikely.
The relatively high prevalence of sports club membership in the sample suggests self-selection bias, with the more active girls more likely to participate in the study. However, even if such bias were present, this would not of itself threaten the validity of the mediation analysis.