Intergenerational transfer of a sports-related lifestyle within the family

Abstract

The intergenerational transfer of a sports-related lifestyle within the family is a potential way to explain the social differences in sports participation that are displayed by young people. In this article, the importance of a sports-related lifestyle within the family, as well as parents’ educational background and sports participation in childhood, in the sports participation of adolescents and young adults is examined from the perspective of socialisation theory. Structural equation modelling was used to demonstrate that all of the examined predictors have significant positive effects on the current sports participation of individuals between the ages of 15 and 30 years (N = 4028; M = 21.48; SD = 4.64). The most pronounced effect on sports participation was observed for a sports-related lifestyle in the family.

Keywords

intergenerational transfer socialisation sports participation and physical activity sports-related lifestyle youths and young adults

Introduction

The question of whether adolescents are able to relate to sports and integrate physical and sports activities in their everyday lives, as well as the way in which they do it, is determined to a large extent by their families, particularly by their parents. In this context, social differences are often manifested in socially disadvantaged children and young people who are less active in sports (Berger et al., 2008; Scheerder et al., 2005; Vandermeerschen et al., 2016; White and McTeer, 2012).

This raises the question of how the positive linkage between the educational status of the parents and the sports involvement of children and adolescents can be explained. Sports and physical activity are forms of behaviour and habits that can vary in importance from one family to the next, depending on the parents’ attitudes, preferences and lifestyle. Depending on these factors, children and adolescents are offered different opportunities for physical activity and sports experiences. According to Welk et al. (2009), parents affect the formation and shaping of sports-related attitudes and behaviour in adolescents, which are understood as a sports-related lifestyle. This sports-related lifestyle includes sports-related orientations, interests and preferences, as well as sports activities, which are learned through socialisation processes (e.g. Brettschneider, 1992). It becomes clear that a sports-related lifestyle is transferred intergenerationally within the family by means of regular joint sports activities, social and material support or sports-related communication during the course of a family’s daily life (Burrmann, 2005; Downward et al., 2014; Welk et al., 2009).

In the studies that have been conducted thus far, the effects of parents’ sports activities and their social support of the sports participation of their children have been investigated comparatively often; however, overall, the findings are inconsistent (Gustafson and Rhodes, 2006). Furthermore, numerous studies consider the direct effects of social inequality and parental sports behaviour, as well as social support of sports activity, to explain the differences in sports participation among youths (Biddle et al., 2011; Pugliese and Tinsley, 2007). However, it is surprising that the relationship between social inequalities (e.g. parental educational level) and the sports-related lifestyle within the family, as well as sports participation in childhood, and its effect on the sports participation of adolescents and young adults has rarely been investigated.

The present study aims to consider the connection between the sports participation of adolescents and young adults and a range of different factors, specifically the sports-related lifestyle within the family, parental educational status and past sports participation in childhood. Although many studies confirm positive influences, the question arises of the extent to which a sports-related lifestyle is conveyed through socialisation processes within the family, particularly in childhood, and whether it is carried over into adolescence and young adulthood, thereby influencing the sports activities in these phases of life. The aim of this study is to examine in more detail: (a) whether a sports-related lifestyle within the family depends on the social background of the parents; and (b) the extent to which these aspects are relevant for adolescents’ participation in sports.¹

First, theoretical considerations and recent studies on social inequality in sports and family socialisation in sports will be presented. Then, a structural equation model will be designed based on the latest research and theoretical considerations, which will be empirically analysed. Finally, the results will be presented and discussed.

Theoretical considerations

In previous studies, it has been shown that the socioeconomic status of the family influences children’s (Chen et al., 2002; White and McTeer, 2012) and adolescents’ (Reichert et al., 2012; Toftegaard-Stockel et al., 2011) access to sports. Education status is crucial to the financial situation of the family, which may determine the availability of a sports-stimulating environment to children and adolescents (e.g. sports equipment, club memberships). In the present study, the direct effect of parents’ educational status on the sports participation of children and adolescents, as well as young adults, is analysed. It can be assumed that the higher the parents’ educational status is, the more pronounced the sports participation in childhood (H1), as well as in adolescence and adulthood (H2), will be. Because the importance of the family decreases and that of a child’s peers increases with age (Hurrelmann, 2010; Zinnecker, 2000), it is possible that the effect of the parents’ educational status on sports participation changes with age (Chen, 2004; Chen et al., 2002). Therefore, the question arises of how the sports-related socialisation practices within the family depend on the parents’ level of education. It is possible that the parents’ level of education is positively linked with the awareness of the effect of sports on the holistic development of the individual. The higher the educational levels of the parents, the greater the probability is that they will invest time, money and social relationships to provide a sports-stimulating environment for their child. It has been shown that an educationally advantaged home is more able to offer children the necessary temporal, financial and socio-emotional support than an educationally disadvantaged home (Evans and Davies, 2010). Based on these considerations, it may be assumed that the parental level of education has an effect on the accessibility of the relevant contexts for sports activities and the extent to which they are actively used. Therefore, it is to be assumed that parental educational status has positive effects on the sports-related lifestyle within the family (H3).

The organisation of the contexts that are relevant to sports and physical activity of youth is conveyed to a substantial degree through socialisation mechanisms within the family (e.g. attitude towards sports and physical activities, social support, educational style of the parents) (Burrmann, 2005; Duncan et al., 2005; Wheeler, 2011). Differences in sports participation of adolescents and young adults are thus determined by sports-related socialisation strategies and practices within the family and their effectiveness (Baur and Burrmann, 2008).

According to Hurrelmann (2006), socialisation is a lifelong process whereby an individual learns the (sports-related) attitudes, values and behaviours that are appropriate to his or her own society. In childhood and youth, individuals are mainly influenced by their family and peers, who are the primary agents of socialisation. For socialisation concerning sports, the previous research has shown that adolescents and young adults are strongly shaped by their parents, whereby the parents’ sports-related attitudes and activities are likely to be relevant to their children’s participation in sports (Downward et al., 2014). Depending on parents’ sports-related orientation and their own sports history and sports-related educational views, children will be offered more or fewer opportunities for sports-related experiences. These findings on the intergenerational transfer of behaviour indicate the importance of parents in their function as positive role models (Bois et al., 2005; Downward et al., 2014). In this scenario, parents can exert an influence on the sports-related lifestyle of their children by serving as passive and unintentional models (Fredricks and Eccles, 2004; Kay, 2004). Aside from this, however, the – active and intentional – support that is given by parents to their children’s sports activities is also important (Beets et al., 2006; Duncan et al., 2005).

Children who grow up in a home with an affinity for sports are, therefore, more likely to become involved in sports than children who are from families that are detached from sports (Downward et al., 2014). It seems that parents’ sports behaviour and attitudes lead to an intergenerational transfer through socialisation processes to the children and youth. Depending on the way in which children are involved in the sports activities of their parents or siblings, sports are assigned high, indifferent or low relevance, which represents the sports-related lifestyle within the family, and this affects the sports involvement of adolescents and young adults accordingly. Therefore, it can be assumed that a sports-related lifestyle within the family leads to a higher sports participation in adolescence and young adulthood (H4).

Sports-related orientation and competencies are communicated to children through family socialisation by involving children in everyday activities such as joint bicycle tours (Baur, 1989). In addition, communicating about sports-specific topics and attending sporting events with the family may shape the sports-related orientation of children. Such health-, physical- or sports-related everyday routines within the family can have a lifelong effect on children while they are growing up. Sports activities in childhood exert a positive influence on sports involvement in adolescence and (young) adulthood (Beunen et al., 2004; Telama et al., 2005) (H5).

However, because social learning experiences within the family are bi-directional, it is also conceivable that an influence may occur in the opposite direction: if a child makes contact with sports-related activities outside of the family home, for example, during physical education (PE) lessons at school, this could also affect the sports-related lifestyle within the family. Children’s involvement in sports increases the relevance of sporting habits in the family’s everyday life, which may lead to parents and/or siblings becoming involved in sports (once again) or, at least, it may lead them to revise their own sports-related lifestyle and their sports-related views on childrearing; this may, in turn, influence the sports participation of the child (Hurrelmann, 2009). A reciprocal relationship may thus be assumed to exist between the sports-related lifestyle within the family and the sports participation of children (H6).

Furthermore, considering the importance of social inequality in sports participation of adolescents and young adults, analyses of the socialisation mechanism should be conducted for different social groups (e.g. gender, age group, nationality). For example, it is possible that the effect of socialisation processes in the family may be less for young women than it is for men. In particular, gender, age group and nationality are of interest because differences in sports participation are often correlated with these sociodemographic factors (European Commission, 2014; Lamprecht et al., 2014). The previous research on the socialisation mechanism in the family neglected the effect of such sociodemographic variables (Dukes and Coakley, 2002). Taking this into consideration, the effect of gender, age group and nationality on the postulated model will be investigated. In particular, the question of whether these three sociodemographic properties moderate the effect of the postulated influencing factors will be studied.

Method

Sample

With the aim of obtaining a sample of approximately 3000 to 4000 adolescents and young adults between the ages of 15 and 30, municipalities were first selected in two Swiss cantons, Bern and Freiburg, which were as heterogeneous as possible in terms of language (German and French), size (number of inhabitants) and extent and quality of municipal sports promotion (e.g. participation in appropriate sport promotion programmes on a national and cantonal level). Of the 100 municipalities that we approached in writing, 36 agreed to take part in the study and gave us access to their population’s postal addresses. Depending on the size of the municipality, either the entire population was used or a random sample was drawn from the entire population of 15- to 30-year-olds. Overall, 24,715 letters were sent to the postal addresses of adolescents and young adults whose names were provided by the municipalities. They were invited to participate in an online survey by following a link or using a QR-code with their smartphone. In the end, the present study included data from 4039 individuals (rate of return: 16.3 %).

Slightly more women (54 %) took part in the study than men (46 %), and the mean age was 21.48 years (SD = 4.64). The youngest age group (15- to 20-year-olds) accounted for 47 %, i.e. almost half, of the respondents (Table 1); the age group of 21- to 25-year-olds accounted for 29 %, and 24 % were in the age group of 26- to 30-year-olds. With respect to the level of education, it is notable that the majority (64%) have either completed compulsory school or secondary school (including lower professional education). Approximately 10% hold a university degree. For 72%, the main language is German, and for 28% the main language is French. Regarding nationality, 82% of the participants are Swiss nationals, 9% are foreign nationals, and 9% have a second nationality in addition to their Swiss citizenship. Furthermore, 83 % of respondents stated that they actively took part in sports, i.e. engaged in gymnastics, fitness or sports at least one to three times a month, whereas 17 % did not participate in sports.

Table 1.

Description of the sample (N = 4028).

Criterion	n	%
Gender	4022
Female	2158	53.7
Male	1864	46.3
Age group	4028
15–20 years old (born 1999–1994)	1903	47.2
21–25 years old (born 1993–1989)	1180	29.3
26–30 years old (born 1988–1984)	945	23.5
Level of education	3208
Compulsory school	1211	37.7
Secondary school/lower professional education/10th school year	850	26.5
Higher professional education leaving certificate	479	14.9
Technical college	336	10.5
University	332	10.3
Language	4028
German	2915	72.4
French	1113	27.6
Nationality	3566
Swiss	2928	82.1
Dual nationality	312	8.7
Foreigner	326	9.1
Current sports activity	4028
Actively doing sports	3336	82.8
Not actively doing sports	692	17.2

Compared with the available data for the Swiss population (Federal Statistical Office, 2016a, 2016b), the population is satisfactorily represented by the sample with respect to gender (female: 51%) and language (German: 63%, French: 23%). Regarding sports activity, the sample appears to be slightly more active than the Swiss population, in which 80% of 15- to 24-year-old adolescents and young adults and 76% of 25- to 34-year-old adolescents and young adults take part in sports (Lamprecht et al., 2014). In the Swiss population, almost a quarter of the participants are from a foreign country (24%), whereas 18% are identified as foreign citizens in the sample. With regard to education level, it is difficult to compare the sample with the overall data for the Swiss population because the adolescents and young adults in that age range are still completing their studies. In conclusion, the sample is not only adequately large, but it also reflects the Swiss population to an acceptable, although certainly not ideal, extent. Some bias may be introduced by the fact that the questionnaire was available only in German and French, so that adolescents and young adults with a different main language are underrepresented in this survey.

Operationalisation of the constructs

The main focus of the standardised questionnaire was on the past and present sports participation of adolescents and young adults, as well as their sports-related lifestyle. The present study draws on the group that the adolescents and young adults designate as their family.

Sports participation of adolescents and young adults

The questions about the sports participation of adolescents and young adults were developed based on the questionnaire for the national survey of sports behaviour of the Swiss population (Lamprecht et al., 2014). In the interest of a broad interpretation of sports, the central dependent variable ‘sports participation of adolescents and young adults’ comprises all of the sports that can be pursued in a sports club, at a private or commercial sports service provider or informally with one’s family, friends or alone. However, it does not include obligatory PE lessons at school or mobility activities (e.g. going shopping on a bicycle). In concrete terms, the sports participation of adolescents and young adults was operationalised by the time that is spent engaging in sports activities (number of hours per week, not including obligatory PE lessons) and the frequency with which the two preferred sports were performed, distinguishing between summer and winter (1 = never, 2 = less than once a month, 3 = approximately one to three times a month, 4 = approximately once a week, 5 = several times a week, 6 = (almost) every day).

Sports participation in childhood

Sports participation in childhood was determined retrospectively by asking ‘How would you describe your sports involvement as a child (up to the age of 14)?’ and offering two response options for continuity of the sports activity (1 = sporadic sports activity or 2 = constant sports activity without interruption) and four possible answers for the degree of the sports activity (1 = no sports activity, 2 = leisure time spent with physical activity, 3 = doing sports, 4 = competitive sports).

Sports-related lifestyle within the family

The items for a sports-related lifestyle within the family were based on a questionnaire to measure the health-related family climate that was developed by Niermann et al. (2014). Six items were used, for which the adolescents and young adults had to state on a five-point Likert scale (1 = does not apply to 5 = applies) to what extent the statement applied to them. In my family: (a) ‘sports are/were very important’; (b) ‘it is/was taken for granted that one does sports regularly’; (c) ‘we support/supported each other in our respective sports activities’; (d) ‘we enjoy/enjoyed actively doing sports together’; (e) ‘it is/was taken for granted that you behave in ways that are beneficial to your health in everyday life’; and (f) ‘we regularly speak/spoke about sports topics’.

Educational status of the parents

The educational status of the parents was determined by the highest level of education that was completed by the father and the mother (‘What level of education has your father/mother completed?’). The possible answers were 1 = no education, 2 = obligatory school, 3 = 10th grade, 4 = secondary school, 5 = lower professional education, 6 = higher professional education leaving certificate, 7 = technical college, 8 = university.

Sociodemographic details of the adolescents and young adults

The adolescents and young adults were also asked to indicate their gender (1 = female, 2 = male), age group (1 = 15- to 20-year-olds, 2 = 21- to 25-year-olds, 3 = 26- to 30-year-olds) and their nationality. It was possible to indicate multiple nationalities. A foreign nationality or dual Swiss and foreign nationality was counted as a foreign nationality, and a (sole) Swiss citizenship was not (1 = Swiss nationality, 2 = foreign nationality).

Data analysis

The data were processed and analysed using the statistical program packages IBM SPSS 23 and R (Version 3.2.2). The key analyses were performed using structural equation models (R package lavaan, Rosseel, 2012), which have the advantage over path models that the postulated effects can not only be estimated simultaneously but also more precisely and with greater validity (correction of measuring errors) (Brown, 2015).

Considering the number of missing values, large differences were found between the individual variables (age: 0.0%, gender: 0.2%, sports participation in adolescence: max. 6.1%, parental education: 12.3%, sports-related lifestyle within the family: max. 12.9%). To detect a possible bias, patterns of missing data were analysed. In particular, correlations were calculated between indicator variables for missingness (0 = no valid answer, 1 = valid answer) and all 21 variables that were used in this study, i.e. all of the model variables and all of the variables describing the sample. By and large, the effects were small (r_max = .18). The most consistent pattern could be observed with respect to sports activity, in the sense that, for example, willingness to answer item i21 (‘In my family (a) sports are/were very important’) was higher among individuals who actively participated in sports (│r│ =.15) in comparison to inactive individuals. However, a slight bias in this direction is probably ubiquitous in the sports science research. As a result, it was assumed that the missing values were random (Enders, 2013). Following Enders (2013), they were estimated using the full information maximum likelihood (FIML) option in the R package lavaan.

In the face of some skewed and ordered categorical variables, two robust methods of estimation were used for the structural equation modelling process (MLR, DWLS). These revealed no differences that are worth mentioning in the results, so only those results obtained using the robust MLR estimate will be presented.

To determine the effect of the three sociodemographic variables of gender, age and nationality on the postulated model, multiple group comparisons were performed. Following a step-up approach (Brown, 2015), different aspects of measurement invariance were checked for the three variables: configurational (equivalence of factor structure), metric (equivalence of factor loadings), scalar (equivalence of the means of manifest variables) and strict factorial invariance (equivalence of error covariances). These analyses are a necessary prerequisite to make valid statements about the structural effects in the various different subsamples, more specifically about the moderator effects and the differences between the latent constructs (equivalence of covariances or the means of latent variables). The models that were examined in this context were compared with one another using χ² difference tests. Based on Chen (2007), who considers this procedure to be too restrictive, we primarily focused on the differences in the comparative fit index (CFI) of the models, and we rejected the null hypothesis (measurement invariance) for differences > .01.

Results of structural equation modelling

Analysis of the overall group

The means, standard deviations and intercorrelations of all the variables used for all of the model estimates are displayed in Table 2. The key results of the structural equation analyses are presented in Figure 1.

Table 2.

Zero-order correlations (r), means and standard deviations for study variables.

	Variable	n	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19
	Sports participation in young adulthood
1	Number of hours of sports activity per week (not including school sports)	3955	3.96	3.73	–
2	Frequency of sports activity in summer (1st sport)	3832	2.88	1.67	.62	–
3	Frequency of sports activity in winter (1st sport)	3784	2.75	1.63	.63	.80	–
4	Frequency of sports activity in summer (2nd sport)	3907	1.66	1.67	.39	.43	.46	–
5	Frequency of sports activity in winter (2nd sport)	3840	1.56	1.62	.39	.45	.40	.73	–
	Sports participation in childhood or adolescence
6	Sports activity in childhood (active/inactive in sports)	3892	2.74	0.70	.31	.19	.21	.15	.13	–
7	Continuity of sports activity in childhood (sporadic sports activity/constant sports activity without interruption)	3853	2.67	0.54	.21	.14	.15	.13	.12	.61	–
	Family attitude towards sports
8	Sports are/were very important in my family	3578	3.20	1.31	.33	.26	.28	.21	.20	.34	.29	–
9	In my family, it is/was taken for granted that one does sports regularly	3570	3.41	1.33	.32	.28	.30	.20	.18	.34	.33	.76	–
10	In my family, we support/supported each other in our sports activities	3546	3.41	1.33	.32	.27	.27	.17	.16	.31	.28	.63	.64	–
11	In my family, we enjoy/enjoyed actively doing sports together	3509	3.04	1.36	.23	.18	.18	.16	.16	.21	.22	.61	.61	.61	–
12	In my family, it is/was taken for granted that you behave in ways that are beneficial to your health in everyday life	3527	3.81	1.09	.18	.18	.17	.14	.12	.19	.22	.49	.54	.50	.46	–
13	In my family, we regularly speak/spoke about sport topics	3583	3.09	1.36	.34	.23	.25	.16	.15	.31	.26	.64	.57	.54	.50	.38	–
	Education
14	Highest level of education (father)	3532	3.97	1.38	.11	.12	.10	.11	.09	.10	.13	.12	.16	.12	.12	.18	.09	–
15	Highest level of education (mother)	3495	3.59	1.29	.09	.12	.09	.08	.07	.07	.10	.08	.12	.10	.07	.15	.04	.49	–
16	Gender (0 = female 1, 2 = male)	4022	0.46	0.50	.18	.08	.04	.01	.02	.10	.03	.08	.07	.01	.03	−.02	.11	.02	.01	–
17	Age group (1 = 15 to 20 years old, 2 = 21 to 25 years old, 3 = 26 to 30 years old)	4028	1.76	0.81	−.09	−.10	−.09	.05	.04	−.04	−.05	−.09	−.10	−.13	−.03	−.06	−.07	−.06	−.09	−.02	–
18	Nationality (1 = Swiss, 2 = foreigner)	3566	1.27	0.62	−.07	−.07	−.11	−.10	−.10	−.07	−.10	−.05	−.06	−.06	−.06	−.08	−.07	−.13	−.06	.01	−.02	–
19	Language-speaking region (0 = French-speaking community, 1 = German-speaking community)	4028	0.73	0.45	.10	.08	.11	.06	.06	−.04	−.03	.04	−.02	.05	.14	.10	.04	.07	.01	−.01	−.03	−.17	–

Notes: 3495 ⩽ n ⩽ 4028. Correlations greater than .07 are significant at p <. 05.

Figure 1.

Standardised coefficients of the proposed model for intergenerational transfer of a sports-related lifestyle within the family. Latent constructs are enclosed in ellipses and observed variables are enclosed in boxes. * p < .001.

Based on Brown (2015), several parameters were used to assess the goodness of the structural equation model: the CFI, the Tucker Lewis index (TLI), the root mean square error of approximation (RMSEA) and the standardized root mean residual (SRMR). Values of TLI ⩾ .95, CFI ⩾ .95, SRMR ⩽ .08 and RMSEA ⩽ .06 were considered to indicate an acceptable model fit.

The measures of goodness of fit show that the postulated model fits the data well (MLR-χ²(82, N = 4028) = 752.747, p < .001; CFI = .970; TLI = .961; SRMR = .026; RMSEA = .045 [.042–.048]). With regard to the estimated model parameters of the base model B, Figure 1 shows that the current sports participation in adolescence and young adulthood was predicted by the sports-related lifestyle within the family (γ = .35, p < .001), whereas the sports participation in childhood (γ = .17, p < .001), as well as the level of education of the parents (γ = .09, p < .001), had a smaller, albeit significant, direct effect on the current sports participation. Furthermore, sports-related lifestyle is also relevant to the current sports participation since it appears as a mediator for parental educational status (β = .21, p < .001). In addition, the parental educational status had an effect on the current sports participation of the adolescents and young adults, which was mediated by sports participation in childhood (β = .18, p < .001). A pronounced correlation (r = .46, p < .001) was found between the sports-related lifestyle within the family and the past sports participation in childhood. The explained variance for the latent variable of the current sports participation of adolescents and young adults is R² = 23.6%. Thus, on the whole, the data provide evidence for the six postulated hypotheses.

Moderator analyses

The results that have been presented thus far are based on the overall group. To determine whether the three sociodemographic features of gender, age group and nationality have an effect on the postulated model, a series of multi-group comparisons was conducted (Table 3). The results of the moderator analyses indicate that there are no moderating effects for gender, age group and nationality. Thus, the central result of these supplementary multi-group comparisons suggests that we must indeed confine ourselves to discussing the parameters of the model that was estimated on the basis of the overall group. This is because it can be assumed that the corresponding model parameters, which were estimated in the gender and age groups, as well as the group with a Swiss or foreign nationality, are identical apart from chance differences.

Table 3.

Tests for invariance across gender, age and nationality: Summary of model fit and χ² difference test statistics.

	Model	χ²	MLR-χ²	df	CFI	TLI	SRMR	RMSEA [95% CI]	CFit	Model comparison	ΔMLR-χ²	Δdf	p
	Gender (G)
A_G1	Women (n = 2159, 53.7%)	439.162	420.538***	82	.973	.965	.027	.044 [.040–.048]	.994
A_G2	Men (n = 1863, 46.3%)	387.875	368.717***	82	.970	.962	.027	.043 [.039–.048]	.994
B_G	Base model (n = 4022): No constraints	827.037	789.065***	164	.972	.964	.027	.044 [.041–.047]	1.000
	Measurement parameters
C_G	Metric invariance: All factor loadings invariant	891.969	857.822***	179	.969	.964	.032	.043 [.041–.046]	1.000	C_G vs. B_G	68.429	15	<.001
D_G1	Scalar invariance: All factor loadings and intercepts of observed variables invariant	1094.547	1055.186***	190	.961	.957	.035	.048 [.045–.050]	.924	D_G1 vs. C_G	203.281	11	<.001
D_G2	Partial scalar invariance: All factor loadings and intercepts of observed variables invariant except for the item ‘hours of sports activity per week’	998.746	962.653***	189	.965	.961	.034	.045 [042–.048]	.998	D_G2 vs. C_G	107.199	10	<.001
E_G	Strict factorial invariance: All factor loadings and intercepts of observed variables invariant except for the item ‘hours of sports activity per week’; two common residual covariances invariant	1018.906	983.495***	191	.964	.961	.034	.045 [.043–.048]	.997	E_G vs. D_G2	22.514	2	<.001
	Structural parameters
F_G	Structural invariance: Equal factor means	1052.435	1015.883***	195	.963	.960	.038	.046 [.043–.049]	.995	F_G vs. E_G	32.401	4	<.001
G_G	Structural invariance: Equal regression weights and factor covariance	1065.985	1027.461***	201	.963	.961	.039	.045 [.043–.048]	.998	G_G vs. F_G	12.469	6	.052
	Nationality (N)
A_N1	Swiss nationality (n = 2927, 82.1%)	667.249	639.732***	82	.966	.956	.027	.048 [.045–.052]	.802
A_N2	Other nationality (n = 639, 17.9%)	205.504	191.110***	82	.973	.965	.035	.046 [.038–.044]	.804
B_N	Base model (n = 3566): No constraints	872.754	824.000***	164	.967	.958	.028	.048 [.044–.051]	.902
	Measurement parameters
C_N	Metric invariance: All factor loadings invariant	1010.451	965.352***	179	.961	.954	.039	.050 [.047–.053]	.573	C_M vs. B_M	151.216	15	<.001
D_N	Scalar invariance: All factor loadings and intercepts of observed variables invariant	1060.090	1016.757***	190	.959	.954	.039	.049 [.047–.052]	.626	D_M vs. C_M	50.864	11	<.001
E_N	Strict factorial invariance: All factor loadings and intercepts of observed variables invariant; two common residual covariances invariant	1086.747	1045.120***	192	.957	.953	.041	.050 [.047–.053]	.512	E_M vs. D_M	34.401	2	<.001
	Structural parameters
F_N	Structural invariance: Equal factor means	1134.296	1089.189***	196	.955	.952	.044	.051 [.048–.053]		F_M vs. E_M	42.552	4	<.001
G_N	Structural invariance: Equal regression weights and factor covariance	1169.197	1115.744***	202	.954	.952	.049	.050 [.048–.053]	.409	G_M vs. F_M	27.698	6	<.001
	Age group (A)
A_A1	15–20 years (n = 1903, 47.2%)	398.656	383.104***	82	.971	.962	.027	.044 [.040–.048]	.989
A_A2	21–25 years (n = 1180, 29.3%)	316.707	296.384***	82	.966	.956	.032	.047 [.042–.053]	.799
A_A3	26–30 years (n = 945, 23.5%)	231.836	223.402***	82	.974	.967	.029	.043 [.036–.049]	.965
B_A	Base model (n = 4028): No constraints	947.199	902.978***	246	.970	.962	.029	.045 [.042–.048]	.998
	Measurement parameters
C_A	Metric invariance: All factor loadings invariant	1077.296	1036.027***	276	.965	.961	.045	.045 [.042–.048]	.996	C_A vs. B_A	134.830	30	<.001
D_A	Scalar invariance: All factor loadings and intercepts of observed variables invariant	1217.819	1173.555***	298	.960	.958	.048	.047 [.044–.050]	.972	D_A vs. C_A	138.971	22	<.001
E_A	Strict factorial invariance: All factor loadings and intercepts of observed variables invariant; two common residual covariances invariant	1267.138	1222.837***	302	.958	.956	.046	.048 [.045–.050]	.919	E_A vs. D_A	53.306	4	<.001
	Structural parameters
F_A	Structural invariance: Equal factor means	1340.965	1293.853***	310	.955	.955	.051	.049 [.046–.051]		F_A vs. E_A	70.756	8	<.001
G_A	Structural invariance: Equal regression weights and factor covariance	1383.466	1331.028***	322	.954	.955	.054	.048 [.046–.051]		G_A vs. F_A	38.067	12	<.001

Notes: MLR = maximum likelihood estimation with robust (Huber-White) standard errors and a scaled test statistic that is (asymptotically) equal to the Yuan-Bentler test statistic; CFI = comparative fit index; TLI = Tucker-Lewis index; SRMR = standardized root-mean-square residual; RMSEA = root-mean-square error of approximation; CFit = p-value of close fit (p RMSEA < .05); ΔMLR-χ² = chi-square difference testing using the Satorra-Bentler scaled chi-square difference.

***

p < .001

To control for a possible method effect, the above calculations were repeated using complete case analysis for the missing data instead of FIML, and all of the relevant estimates (loadings, covariances and regressions) were correlated. The correlation of r = .99 indicated that the results were almost identical.

Discussion

Considering the social inequalities that are often mentioned in the literature in connection with the sports participation of adolescents and young adults (e.g. gender, age group and nationality) and the further explanations for sports participation that are typically drawn upon (including parental educational status and sports participation in childhood) (Nagel, 2003), the central goal of the present study was to focus on the family as an agent of socialisation. We were interested in determining whether the sports-related lifestyle within the family contributes towards explaining individual differences in the sports participation of adolescents and young adults over and beyond the known factors that are mentioned above. In addition, our aim was to examine whether this effect, along with the others, differed depending on gender, age group or nationality.

Overall, the results underscore the fact that the family and its sports-related lifestyle are noteworthy agents of socialisation for the sports participation of adolescents and young adults and that they serve both as a direct and indirect agent of socialisation, as a mediator between the parental educational status and the sports activities in childhood and the sports participation in adolescence and young adulthood. The results that have been obtained thus support the hypothesis that the sports-related lifestyle within the family influences the sports participation of children when they are growing up (Baur, 1989). However, it is possible that the psychosocial support between family members can compensate for a lack of parental education (see Stecher, 2001; Zinnecker and Stecher, 1996). It thus seems appropriate to pay more attention to the construct of a sports-related lifestyle within the family in developing future studies.

To determine the extent to which sociodemographic properties (gender, age group and nationality) moderate the relationships that are postulated in the model, particularly the effect of sports-related lifestyle within the family on the sports participation in youth and adulthood, multi-group comparisons were conducted. These revealed no signs of any moderator effects. It follows that the effect of a sports-related lifestyle within the family on the sports participation of adolescents and young adults (as well as all other effects) can be assumed to be the same – irrespective of which factors of gender, age group or nationality are considered. This finding is remarkable in that the previous studies have shown typical gender transmissions of so called ‘sports capital’ in the family context (e.g. Fredricks and Eccles, 2004; Heinze et al., 2014) and that these three sociodemographic features are associated with substantial social inequalities in sports (Nagel, 2003).

Regarding public sports and health promotion, the results of the present study show that efforts based on family-specific sports promotional concepts may be promising in terms of sports participation and that this is true irrespective of gender, age group or nationality.

Limitations and future research

From a methodological perspective, it is critical to note that we conducted a cross-sectional study. Hence, the information about sports participation in childhood had to be ascertained retrospectively. The operationalisation of a sports-related lifestyle must also be viewed critically, since the chosen timeframe included both the present and the past. This is because it was only possible to use one questionnaire with a single question format. In the case of adolescents and young adults who no longer live with their parents, we were interested in the sports-related lifestyle in their past, so a unidirectional effect has been assumed between the sports-related lifestyle within the family and the sports participation in adolescence and young adulthood. However, in the case of children who are still living with their parents, we wanted to know about the sports-related lifestyle in the present. Due to the way in which the question was asked, including both the present and the past, a certain temporal fuzziness occurred; this led to our specifying a reciprocal influence rather than a unidirectional effect of the sports participation in childhood on the sports-related lifestyle within the family and a unidirectional effect of the sports-related lifestyle within the family on the sports participation of adolescents and young adults. It is left to future studies to analyse more in depth the relationship between sports participation at different stages of life and a sports-related lifestyle within the family.

In the future, longitudinal studies should be conducted to mitigate these methodological difficulties. Beyond this, it would be advisable to draw on the manifold options of structural equation modelling in the analysis. Although complex, this method of analysis is excellently suited to account for error variance in the data and to detect differential item functioning which occurs when a questionnaire has different measurement properties for members of different subgroups (e.g., Brown, 2015). Considering this, future studies should pay more attention to the measurement equivalence of the instruments that are used. However, a prerequisite for this is the use of relatively large samples.

Of course, the structural equation model that is presented does not capture all of the factors that affect sports participation in adolescents and young adults. Because peers have an increasing importance with age (Hurrelmann, 2010; Zinnecker, 2000), it would be advisable to involve the influence of peers on adolescents’ and young adults’ sports participation in future studies to improve the model fit and to reduce the unexplained variance. Furthermore, structural circumstances such as local sports infrastructure and the number and variety of sports that are offered at the place of residence, education or work are presumably important for the sports participation of young people when they are growing up (Davison and Lawson, 2006; Wicker et al., 2012), which is why it is recommended that future analyses should take into account not only the social but also the structural factors that influence the sports participation of adolescents and young adults.

From the perspective of sports and health promotion, further research questions on the intergenerational transfer of a sports-related lifestyle within the family also need to be answered, specifically, the extent to which the parents’ knowledge of the effects of sports activities on the development of their children affects sports behaviour in the everyday lives of the family, and how one might influence this (e.g. intervention studies), with the aim of maintaining the long-term sports participation of young people as they are growing up and reducing the social inequality that (still) exists.

Footnotes

Declaration of Conflicting Interests

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

Funding

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

Notes

References

Baur

(1989) Körper- und Bewegungskarrieren. Schorndorf: Hofmann.

Baur

Burrmann

(2008) Sozialisation zum und durch Sport. In: Weis

Gugutzer

(eds) Handbuch Sportsoziologie. Schorndorf: Hofmann, pp. 230–238.

Beets

Vogel

Forlaw

. (2006) Social support and youth physical activity: The role of provider and type. American Journal of Health Behavior 30(3): 278–289.

Berger

O’Reilly

Parent

. (2008) Determinants of sport participation among Canadian adolescents. Sport Management Review 11: 277–307.

Beunen

Lefevre

Philippaerts

. (2004) Adolescent correlates of adult physical activity: A 26-year follow-up. Medicine & Science in Sports & Exercise 36(11): 1930–1936.

Biddle

Atkin

Cavill

. (2011) Correlates of physical activity in youth: A review of quantitative systematic reviews. International Review of Sport and Exercise Psychology 4(1): 25–49.

Bois

Sarrazin

Brustad

. (2005) Elementary schoolchildren’s perceived competence and physical activity involvement: The influence of parents’ role modelling behaviours and perceptions of their child’s competence. Psychology of Sport and Exercise 6: 381–397.

Brettschneider

(1992) Adolescents, leisure, sport and lifestyle. In: Williams

Almond

Sparkes

(eds) Sport and Physical Activity: Moving Towards Excellence. London: E & FN Spon, pp. 536–550.

Brown

(2015) Confirmatory Factor Analysis for Applied Research. New York: Guilford Press.

10.

Burrmann

(2005) Zur Vermittlung und intergenerationalen ‘Vererbung’ von Sport(vereins)engagements in der Herkunftsfamilie. Sport und Gesellschaft 2: 125–154.

11.

Chen

(2004) Why socioeconomic status affects the health of children: A psychosocial perspective. American Psychological Society 13(3): 112–115.

12.

Chen

Matthews

Boyce

(2002) Socioeconomic differences in children’s health: How and why do these relationships change with age? Psychological Bulletin 128(2): 295–329.

13.

Chen

(2007) Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal 14(3): 464–504.

14.

Davison

Lawson

(2006) Do attributes in the physical environment influence children’s physical activity? A review of the literature. International Journal of Behavioral Nutrition and Physical Activity 3(1): 19 (accessed 21 May 2013).

15.

Downward

Hallmann

Pawlowski

(2014) Assessing parental impact on the sports participation of children: A socio-economic analysis of the UK. European Journal of Sport Science 14(1): 84–90.

16.

Dukes

Coakley

(2002) Parental commitment to competitive swimming. Free Inquiry in Creative Sociology 30(2): 185–197.

17.

Duncan

Strycker

(2005) Sources and types of social support in youth physical activity. Health Psychology 24(1): 3–10.

18.

Enders

(2013) Dealing with missing data in developmental research. Child Development Perspectives 7(1): 27–31.

19.

European Commission (2014) Sport and Physical Activity: Report. Brussels: European Commission.

20.

Evans

Davies

(2010) Family, class and embodiment: Why school physical education makes so little difference to post-school participation patterns in physical activity. International Journal of Qualitative Studies in Education 23(7): 765–784.

21.

Federal Statistical Office (2016a) Sprachliche Praktiken in der Schweiz: Erste Ergebnisse der Erhebung zur Sprache, Religion und Kultur 2014. Neuchâtel: Federal Statistical Office.

22.

Federal Statistical Office (2016b) Switzerland’s Population 2014. Neuchâtel: Federal Statistical Office.

23.

Fredricks

Eccles

(2004) Parental influences on youth involvement in sports. In: Weiss

(ed.) Developmental Sport and Exercise Psychology. Mornington, WV: Fitness Information Technology, pp. 144–164.

24.

Gustafson

Rhodes

(2006) Parental correlates of physical activity in children and early adolescents. Sports Medicine 36(1): 79–97.

25.

Heinze

Davis

. (2014) Gender role beliefs and parents’ support for athletic participation. Youth & Society. DOI: 10.1177/0044118X14553580

26.

Hurrelmann

(2006) Einführung in die Sozialisationstheorie. Weinheim: Beltz.

27.

Hurrelmann

(2009) Social Structure and Personality Development: The Individual as a Productive Processor of Reality. Cambridge: Cambridge University Press.

28.

Hurrelmann

(2010) Lebensphase Jugend: Eine Einführung in die sozialwissenschaftliche Jugendforschung. Weinheim: Juventa.

29.

Kay

(2004) The family factor in sport: A review of family factors affecting sports participation. In: Sport England (ed.) Driving up Participation: The Challenge for Sport: Academic Review Papers Commissioned by Sport England as Contextual Analysis to Inform the Preparation of the Framework for Sport in England. London: Sport England, pp. 39–60.

30.

Lamprecht

Fischer

Stamm

(2014) Sport Schweiz 2014: Sportaktivität und Sportinteresse der Schweizer Bevölkerung. Magglingen: Bundesamt für Sport.

31.

Nagel

(2003) Soziale Ungleichheiten im Sport. Aachen: Meyer & Meyer.

32.

Niermann

Krapf

Renner

. (2014) Family health climate scale (FHC-scale): Development and validation. International Journal of Behavioral Nutrition and Physical Activity 11(1): 30.

33.

Pugliese

Tinsley

(2007) Parental socialization of child and adolescent physical activity: A meta-analysis. Journal of Family Psychology 21(3): 331–343.

34.

Reichert

Monteiro

Freitas Junior

. (2012) Characteristics of family nucleus as correlates of regular participation in sports among adolescents. International Journal of Public Health 57(2): 431–435.

35.

Rosseel

(2012) lavaan: An R package for structural equation modeling. Journal of Statistical Software 48(2).

36.

Scheerder

Vanreusel

Taks

(2005) Stratification patterns of active sport involvement among adults: Social change and persistence. International Review for the Sociology of Sport 40(2): 139–162.

37.

Stecher

(2001) Die Wirkung sozialer Beziehungen: empirische Ergebnisse zur Bedeutung sozialen Kapitals für die Entwicklung von Kindern und Jugendlichen. Weinheim: Juventa.

38.

Telama

Yang

Viikari

. (2005) Physical activity from childhood to adulthood: A 21-year tracking study. American Journal of Preventive Medicine 28(3): 267–273.

39.

Toftegaard-Stockel

Nielsen

Ibsen

. (2011) Parental, socio and cultural factors associated with adolescents’ sports participation in four Danish municipalities. Scandinavian Journal of Medicine & Science in Sports 21(4): 606–611.

40.

Vandermeerschen

Vos

Scheerder

(2016) Towards level playing fields? A time trend analysis of young people’s participation in club-organised sports. International Review for the Sociology of Sport 51(4): 468–484.

41.

Welk

Babkes

Schaben

(2009) Parental influences on youth sport participation. In: Coelho Silva

Figueiredo

Elferink-Gemser

. (eds) Youth Sports: Participation, Trainability and Readiness. Coimbra: Imprensa da universidade de Coimbra, pp. 75–102.

42.

Wheeler

(2011) The significance of family culture for sports participation. International Review for the Sociology of Sport 47(2): 235–252.

43.

White

McTeer

(2012) Socioeconomic status and sport participation at different developmental stages during childhood and youth: Multivariate analyses using Canadian National Survey Data. Sociology of Sport Journal 29: 186–209.

44.

Wicker

Hallmann

Breuer

(2012) Micro and macro level determinants of sport participation. Sport, Business and Management: An International Journal 2: 51–68.

45.

Zinnecker

(2000) Selbstsozialisation. Essay über ein aktuelles Konzept. Zeitschrift für Soziologie der Erziehung und Sozialisation 20(3): 272–290.

46.

Zinnecker

Stecher

(1996) Zwischen Lernarbeit und Erwerbsarbeit: Wandel und soziale Differenzierung im Bildungsmoratorium. In: Zinnecker

Silbereisen

(eds) Kindheit in Deutschland. Weinheim: Juventa, pp. 165–184.