Contextual factors influencing voluntary organizations’ inclusiveness and individual participation in leisure activities: The Case of Norway

ABSTRACT

While research on participatory inequalities in organized leisure has mainly focused on individual resources and contextual factors, studies linking the individual, organizational and contextual levels are lacking. This study remedies this by investigating the degree to which contextual, organizational and individual factors enhance or inhibit participation in membership-based voluntary leisure organizations among children and adolescents in Norway. We studied the individual, organizational and contextual factors by examining how these factors affect participation across municipalities. Our analyses based on individual, organizational, and municipal data investigate, using a mixed-effects approach, the extent to which organizational resources, organizational inclusiveness, and local public policy are associated with increased participation in leisure activities. Our findings show that contextual factors have little direct explanatory power but moderate individual and organizational characteristics while public policies addressing the supply of leisure activities are at best ineffective and should be aimed to a greater extent at reducing socio-economic inequalities at the individual level.

KEYWORDS:

• Inclusiveness, voluntary organizations, contextual effects, leisure

Organized leisure activities are often promoted as socially inclusive arenas, especially for children and youth. Empirical studies have underscored the positive influence of structured leisure activities on various developmental outcomes among children and youth (Sichling & Plöger, 2018), including mental health and problem behavior (Bartko & Eccles, 2003), the acquisition of politically relevant skills (Quintelier, 2008), and success in school and education (Badura et al., 2016). Participation in organized leisure activities in early ages has also been shown to contribute to fostering adult active participation and citizenship, as people who engage in volunteering and leisure activities in their youth are likelier to also be engaged in politics and society when entering adulthood (Plutzer, 2002).

Leisure activities for children and youth in the domains of culture, sports, and recreational activities are organized by local voluntary organizations in many countries, including Norway. While active participation in voluntary leisure organizations has positive external effects for both individuals and communities, inequalities characterize participation in these organizations (Schröder & Neumayr, 2023). At the individual level, the lack of individual resources such as financial means, cognitive abilities, or social skills has been considered the main factor that limits the possibilities of individuals’ participation (Schlozman et al., 1999; Verba et al., 1995; Wilson & Musick, 1998). However, opportunities for participation are also contingent on contextual and organizational factors, such as sociodemographic characteristics and the degree of organizational selectivity or inclusiveness, which is, in turn, conditioned on social, economic and political factors characterizing the local communities in which organizations operate (Duncan, 2010; Grønbjerg & Paarlberg, 2001).

While research on participatory inequalities (Hustinx et al., 2022; Schröder & Neumayr, 2023) has mainly focused on individual resources and contextual factors, studies linking the organizational and contextual levels are lacking. In this study, we address this gap in the literature by investigating whether and to what extent contextual factors influence and moderate (a) the degree of inclusiveness of voluntary organizations through the level of their resources and (b) influence and moderate individual socioeconomic characteristics influencing participation in leisure activities supplied by membership-based voluntary organizations. More specifically, we use individual, organizational and contextual data to examine how municipal-level factors impact the demand and supply of leisure activities and, consequently, the participation of children and youth in leisure activities. Demand factors include individual socio-economic characteristics. Supply factors include the level of inclusiveness and the resources of voluntary organizations. Contextual factors relate to local public policies and selected socio-economic characteristics of municipalities. The inclusiveness of voluntary organizations is defined by their efforts to increase the diversity of their members in terms of socioeconomic status and ethnic or cultural background. To establish a connection between individual participation and these individual, organizational, and contextual factors, we have developed a comprehensive conceptual framework that integrates various insights from existing literature. Methodologically, we use mixed-effects regression models clustered by municipalities to investigate the extent to which these contextual factors moderate organizational inclusiveness through their resources and individual participation through the role played by individual socioeconomic characteristics.

We study this within the Norwegian setting, where many leisure activities are organized by local organizations, volunteering is the main driver of civil society, and professional actors with paid staff play only a marginal and supplemental role to the state in providing leisure activities. The relationship between the state and civil society in Norway is characterized by closeness and cooperation, as the organizations turn to the state for cooperation, funding, and legitimacy, but they also have a great amount of autonomy from the state (Enjorlas & Strømsnes, 2018). Norway is a particularly good case for investigating the contextual and organizational factors influencing individual participation in organized leisure activities and voluntary organizations’ inclusiveness due to the crucial role that voluntary organizations play in society and, in this context, for children and youth in particular. Still, the autonomy that the organizations have from the state also means that the state has limited tools for implementing and enforcing inclusiveness policies and measures for leisure activities. Furthermore, in contrast to schools, kindergartens, and health centres for children, participation in organized leisure is not regulated in detail at the national and local levels. This means that the municipalities of Norway also have great autonomy from the state in this policy field. Thus, we expect some variation, at the municipal level when it comes to policies for participation and inclusiveness.

Conceptual Framework

The term ‘social inclusion’ originated in European social policy discourse during the 1980s (Koikkalainen, 2011). Social inclusion, understood as a specific target-oriented policy, was developed in opposition to the concept of social exclusion credited to René Lenoir (Lenoir, 1974) and covers a wide range of economic and social problems (Allman, 2013; Sen, 2000). The concept of social exclusion denotes the inability of some persons or groups facing a combination of linked problems in multiple domains – such as education, employment, health, social participation, and community integration (Levitas et al., 2007) – to participate in the mainstream functions of society (Koikkalainen, 2011). Social inclusion in leisure activities can be understood in terms of four dimensions: A spatial dimension related to proximity and closing of social distances; a relational dimension related to a sense of belonging and social acceptance; a functional dimension related to the development of knowledge, skills and understanding; and a power dimension related to individual control and communal influence (Bailey, 2005).

Organizational inclusiveness relates, in this study, to the spatial dimension: The degree of diversity among voluntary organizations’ members in terms of their socio-economic and cultural backgrounds. Voluntary organizations’ resources and membership’s diversity are hypothesized to be related to each other. Indeed, as voluntary organizations are highly dependent on their members, both for financial resources such as membership fees and for voluntary work, they risk being exposed to ‘the paradox of inclusion and exclusion’ (Solebello et al., 2016) according to which ‘attempts to increase the association’s inclusiveness are met with countervailing desires to maintain the membership association’s exclusiveness’ (Solebello et al., 2016). Efforts to increase the degree of organizational inclusiveness may raise challenges related to the organization’s identity and the threat of shrinking membership and resources (Mason, 2020; Solebello et al., 2016).

As Wolpert (1993) and Bielefeld et al. (1997) have shown, voluntary organizations are largely locally supported. Therefore, contextual factors characterizing localities, such as ethnic and cultural heterogeneity and socioeconomic inequalities, are likely to influence both the demand and supply of voluntary organized leisure locally. However, organizational density within a community, and thus the organizational resources available, appears to vary across metropolitan and rural areas (Bielefeld, 2000; Clifford, 2012; Grønbjerg & Paarlberg, 2001; Kim, 2015). While Gilster (2017) found an association between organizational resources (i.e., neighbourhood organizations) and participation across neighbourhoods, Higgs et al. (2021) found that despite volunteering being associated with local organizational resources, this association did not hold after adjusting for other contextual factors, suggesting that much of the area-level variance is due to spatial variations in compositional factors (other than organizational resources).

Previous research has established a relationship between the socioeconomic characteristics of localities (neighbourhoods, municipalities) and participation, but not a consistent pattern of relationships (Wilson, 2000). Indeed, different studies associate affluence, disadvantage, diversity, and homogeneity with decreased participation (Gilster, 2017). One explanation for these contradictory findings is that these factors may be at work simultaneously. Organizational resources can be thought of as mediating contextual factors (Sampson & Graif, 2009) as a range of locally-bound contextual socioeconomic factors affect individual outcomes (such as participation) through local organizational resources. For instance, a community with a diverse cultural background can have a more diverse supply of voluntary leisure activities than less diverse communities, but only if there are organizational resources in the community to facilitate the demand for these activities.

To disentangle how these complex mechanisms influence organizational inclusiveness and individual participation, we propose (i) to distinguish contextual factors from organizational factors influencing participation and (ii) relative to organizational factors, to identify the different pathways through which contextual factors influence both the demand and supply of voluntary leisure activities.

Contextual Effects on Demand and Supply of Voluntary Activities

Contextual factors may influence both the demand for and supply of voluntary activities in a locality. The economic theories of the nonprofit and voluntary sector provide a useful starting point for elucidating how contextual factors may influence the demand and supply of voluntary activities.

The theories of market and government failure (Hansmann, 1987) suggest that voluntary and nonprofit organizations meet the demand for goods and services that are not adequately supplied by the market or government. On the one hand, government failure theory proposes that the demand for the provision of voluntary organized activities will result from the inability of the political process to satisfy the diversity of needs and preferences, since the government will respond to the demand of the majority (the median voter) and will leave unsatisfied the preferences of small or powerless minorities (Weisbrod, 1988). More generally, this theoretical argument suggests the existence of a relationship between the heterogeneity of preferences in the population (expressed in terms of socioeconomic, ethnic, cultural, or religious diversity) and the demand for voluntary organized activities (James, 1989). This implies that communities characterized by high levels of heterogeneity or population diversity and socioeconomic inequalities will express stronger needs for voluntary organized activities because these activities will be undersupplied by the government.

On the other hand, the supply of voluntarily organized activities may be explained in terms of different forms of coordination failure. Voluntary organizations may be conceived of as governance structures adapted to the coordination of action and able to alleviate different forms of coordination failures, such as market failure, government failure, and collective action failure (Enjolras, 2009). Market failure occurs in situations characterized by informational asymmetries between the parties involved in market exchange. Government failure occurs in the presence of public goods or merit goods when the governmental provision of these goods may be nonoptimal because actors have no incentive to reveal their true preferences. Collective action failure happens as a result of free riding in strategic situations where not contributing to a collective action (free riding) is the best strategy, irrespective of whether others contribute (Hardin, 1992).

The governance structure approach to voluntary organizations (Enjolras, 2009) argues that the specific governance structure that characterizes these organizations allows them to mitigate market failures and collective action failures and, therefore, to constitute an alternative to the government provision of collective or informational goods. Yet, as pointed out by Salamon (1987), this capacity is contingent on voluntary organizations’ ability to mobilize the necessary voluntary resources (such as voluntary work, donations, and membership fees) in their environment. From a supply-side perspective, the uneven distribution of voluntary organizations across localities may be the result of ‘voluntary sector failure’ (Salamon, 1987): gaps relative to the coverage of needs or demands may exist because ‘private charitable resources may or may not be available where the need for them is greatest’ (p. 45). In other words, the distribution of local voluntary organized activities may be due not only to variations relative to the demand for these activities but also to spatial variations characterizing the supply of resources needed for these activities.

In sum, from a demand perspective, contextual factors characterizing local communities, such as lower socioeconomic resources (non-satisfied needs), ethnic, cultural, and religious diversity, will lead to increased demand for voluntary organized activities, whereas from a supply side perspective, socioeconomic and cultural heterogeneity will lead to less available resources (such as voluntary work, donations, and membership fees) available to voluntary organizations, since socioeconomic inequalities and cultural heterogeneity are associated with lower levels of trust, social capital, and economic resources. Yet, voluntary organizations are not necessarily a substitute for government, but they might also complement government and get public funding by being part of the implementation of local public policies (Young, 2006). Thus, one yet-to-be-resolved research puzzle in the literature is the actual combined effect of the supply-and-demand side of contextual factors on participation and inclusiveness in voluntary organizations.

To summarize, we can distinguish between three locally embedded contextual mechanisms that interact with each other and affect both the supply and demand of voluntary organizations and, by extension, their degree of inclusiveness: the degree of cultural heterogeneity of the community, the level of resource inequalities, and the nature of public policies.

Hypotheses: Social Heterogeneity, Inequalities and Local Public Policies

At the organizational level, we posit that the degree of organizational inclusiveness is likely to be influenced by their dependency on resources. Organizations need resources to carry out their activities (McCarthy & Zald, 1977). Resource dependence theory (Pfeffer, 1978) considers organizations as strategic agents that are strongly influenced by their outside environment (Scott, 2003). An organization’s dependence on the environment is determined by how rare the resources are in the environment, and how important or critical they are to the organization’s survival. Resources are ‘anything of value, tangible, or intangible, that can be exchanged between organizations’ (Saidel, 1991). Voluntary organizations rely heavily on members, donors, philanthropic and public funders, and on ancillary commercial activities for their resources (Froelich, 1999), a type of dependence that can affect their strategy and generate critical issues of multiple principle – agent problems (Johnson & Prakash, 2007). Still, organizations with higher levels of resources will be able to pursue more inclusiveness than less resourceful organizations. More specifically, it is expected that the more volunteers and paid employees an organization has, the more inclusive it is, as these reflect its human and financial resources.

Community diversity or heterogeneity, because it produces benefits but also imposes coordination costs (Costa & Kahn, 2003), has contrasting consequences for participation in voluntary organizations as well as for organizational resources, and as a result, on organizational inclusiveness. On the one hand, at the individual level, the demand for voluntary activities is likely to increase with the degree of heterogeneity of the community (James, 1989). On the other hand, the empirical regularity of lower participation in more heterogeneous communities has been repeatedly documented (Costa & Kahn, 2003).

Alesina and La Ferrara (2000) have proposed a model in which individuals prefer to interact with others who are like themselves in terms of income, race, or ethnicity. In such a setting, the preference for homogeneity is expected to decrease total participation when community heterogeneity increases. However, if people choose to sort themselves into homogeneous groups, their participation will increase. Their empirical results in US localities (Alesina & La Ferrara, 2000) suggest that income inequality and racial and ethnic heterogeneity reduce the propensity to participate in voluntary organizations, especially for groups in which direct contact among members is important (such as churches and youth clubs), with racial fragmentation displaying the strongest negative effect on participation.

Community heterogeneity reduces participation in voluntary organizations because such participation is transmitted by social networks and is homopilous (McPherson, 1983; McPherson et al., 1992; McPherson et al., 2001). Social homophily entails that individuals occupying similar social positions are likelier to interact with one another. As a result, heterogeneity is negatively related to participation because the potential for homophilous social network ties decreases with heterogeneity (Rotolo, 2000).

This should pertain to both volunteering in an organization and for individual participation in the activity. The relative scarcity of volunteering is expected to have a negative impact on the relationship between volunteers and the degree of organizational inclusiveness, because organizations that succeed in recruiting volunteers will tend to be more homophilous. Still, organizational tangible resources (such as paid employees often made possible by public subsidies) can remedy the negative association between heterogeneity and organizational inclusiveness by making it possible to facilitate costly efforts for increased inclusiveness. Our first hypotheses are thus:

H1-Community-level ethnic diversity hypothesis: High levels of ethnic and cultural diversity at the community level will be associated with lower levels of organizational inclusiveness.

We expect also the following:

H1a: High levels of ethnic and cultural diversity will negatively moderate the influence of organizational volunteers on organizational inclusiveness.

However, if organizations have the tangible resources necessary to facilitate costly efforts for increased inclusiveness, for instance by hiring paid employees, it may lessen the negative impact of heterogeneity. In fact, financial and professional resources may matter more in a high diversity context because of the lower participation and greater need for inclusive measures. Furthermore, organizations in communities with high levels of ethnic and cultural diversity may be more likely to be supplied those resources, for instance through public funding aimed at promoting inclusion in such communities. Thus, it may be expected that paid employees, as a proxy for financial resources, will have a strengthened effect on organizational inclusiveness in high diversity communities:

H1b: High levels of ethnic and cultural diversity will positively moderate the influence of paid employees on organizational inclusiveness.

Levels of community heterogeneity, especially ethnic and cultural heterogeneity, have been repeatedly found to be associated, in empirical studies, with lower levels of social capital, as well as with varying measures of participation in voluntary organizations, including membership and volunteering (see e.g., Alesina & La Ferrara, 2000). Yet, community cultural and ethnic heterogeneity, in addition to exercising an effect on participation propensity through social networks and trust, appears to interact with socioeconomic inequalities. Extant literature (e.g., Wiepking & Bekkers, 2012) has explored how individual-level resources, such as income and education, influence membership and volunteering in voluntary organizations. However, socioeconomic inequalities at the community or societal level appear to impact participation (Schröder & Neumayr, 2023) and to influence participation directly and indirectly through trust and social capital.

Socioeconomic inequalities at the community level exert a direct influence since, according to the ‘resource hypothesis’ (Schröder & Neumayr, 2023), ‘with an increase in inequality, groups, and organizations in communities with lower endowments have comparatively fewer resources – such as financial support, links to centers of power and leadership skills – to rely upon to foster other civic engagement’ (Schröder & Neumayr, 2023). Yet, studies that control for the effects of contextual-and individual-level resources show that differences in resources do not explain all the variance across contexts, indicating that other mechanisms such as cultural and ethnic heterogeneity and differences in social capital are at play (Alesina & La Ferrara, 2000; Gesthuizen et al., 2009; Lancee & Van de Werfhorst, 2012).

Indeed, contextual socioeconomic inequalities appear to indirectly affect participation by depressing social trust and the capacity for collective action (Uslaner & Brown, 2005). Inasmuch as trust is likelier to develop between people who share the same social and cultural background – thus being likelier to develop in more homogeneous and equalitarian groups (Schröder & Neumayr, 2023) – increased inequalities are likely to lead to lower levels of social trust, and therefore to lower levels of participation (Rothstein & Stolle, 2003; Uslaner & Brown, 2005). Additionally, according to the ‘resource hypothesis’ (Schröder & Neumayr, 2023, p. 684), communities levels of inequality tend to reduce individual participation opportunities, impacting negatively on participation. At the organizational level, socioeconomic inequalities impact also the pool of voluntary resources (membership fees, voluntary work, donations, and ancillary resources) that organizations can mobilize within a community and, thus, tend to lower levels of organizational inclusiveness. Still, we expect that organizational tangible resources (such as paid employees often made possible by public subsidies) will remedy the negative association between community level inequality andorganizational inclusiveness. Thus, our second set of hypotheses is as follows:

H2-Community-level inequalities hypothesis: High community-level inequalities will be associated with lower levels of organizational inclusiveness.

H2a: High community-level inequalities will negatively moderate the influence of organizational volunteers on organizational inclusiveness.

Again, however, if organizations in communities are supplied the necessary resources to facilitate costly inclusive measures, for instance through public funding that enable them to hire paid employees, it may negate the lack of voluntary resources. In a high inequality community, being supplied those resources may also matter more for inclusion due to more children and youth being at risk than in low inequality contexts. In other words, the effect of paid employees on inclusion is expected to be strengthened in high inequality communities:

H2b: High community-level inequalities will positively moderate the influence of paid employees on organizational inclusiveness.

H2c: High community-level inequalities will positively moderate the influence of individual socioeconomic resources on individual participation.

Still, the direct and indirect effects of social inequalities on participation in voluntary organizations can be mitigated by active public policies promoting inclusiveness. Young’s (2006) approach to government – nonprofit relations conceptualizes that nonprofits can develop three types of relationships with government that can coexist and are not mutually exclusive: Nonprofits can supplement government, complement government, through collaboration and/or be adversarial to government. Kim (2015) underscores that these types of relationships fulfil different functions. Within a supplementary relationship, nonprofits mobilize voluntary resources to satisfy needs that are unmet by the market and the government. In contrast, within a complementary relationship, nonprofits receive government funding because voluntarily mobilized resources are not sufficient to satisfy unmet needs. The complementary perspective emphasizes nonprofit – government collaboration (Salamon, 1995) and suggests that government, by providing financial resources to voluntary organizations through local public policies, may support organizations experiencing difficulties in mobilizing sufficient voluntary resources, especially in communities with culturally and ethnically diverse populations that have heterogenous demands and low levels of socioeconomic resources.

‘The paradox of inclusion and exclusion’ implies that if members join a voluntary organization association based on a set of shared values and identities, shifting these values and identities as a result of the inclusion of new groups risks compromising what makes the organization special and may cause members to feel socially alienated (Solebello et al., 2016). The threat that shrinking membership could jeopardize the resource basis of the organization may discourage the organization’s leadership from implementing actions oriented toward increasing the degree of organizational inclusiveness. This tendency can, however, be countered by institutional pressures exerted by public authorities through public policies and programmes promoting social inclusion, since organizations that increase their inclusiveness because of such public action would be rewarded by legitimacy and resources. One way of signalling such a public action would be to adopt an active strategy for inclusion. Organizational strategies, programmes, and projects on inclusion would also serve to anchor such initiatives among the organization’s members and/or constituency, which may counteract the possible alienation of existing members. Based on this, we hypothesize the following:

H3-Local public policy hypothesis: Local public policy toward the inclusion of children and youth will be positively associated with organizational inclusiveness.

H3a: Local public policy toward the inclusion of children and youth will positively moderate the influence of organizational volunteers on organizational inclusiveness.

H3b: Local public policy toward the inclusion of children and youth will positively moderate the influence of paid employees on organizational inclusiveness.

H3c: Local public policy toward the inclusion of children and inequalities will negatively moderate the influence of individual socioeconomic resources on individual participation.

In sum, in localities that are most diverse (characterized by a higher degree of cultural and ethnic heterogeneity and lower levels of social capital) and with higher levels of socioeconomic inequalities (in income and education), the need for inclusiveness in voluntary leisure organizations will be higher. At the same time, the environment of these organizations will be characterized by lower levels of available resources (different forms of resources, capital, and network ties). However, policies (or the lack thereof) toward voluntary organizations implemented by local authorities (municipalities) are likely to influence the resources available to voluntary organizations and foster their inclusiveness.

Design and Data

Data

To test our three hypotheses, which build on theoretically grounded mechanisms of influence within and between contextual, organizational and individual factors and children and youths’ participation in organized leisure activities, we use a combination of data sources. More specifically, we combine original survey data collected at the organizational level with existing data sources at the individual and municipal levels.

Individual Data

At the individual level, we use Ungdata, a series of high quality nationally representative youth surveys (Bakken, 2019),¹ containing information about different types of participation (e.g., sport clubs, religious organizations and cultural activities) and background information. Ungdata has been implemented in most Norwegian municipalities gathering answers from more than 630,000 adolescents aged 13–19 years since 2010. A great advantage of Ungdata is that it is distributed to virtually all schools, endorsed by school authorities, and carried out during school hours, allowing to secure a high response rate, also in groups that often are underrepresented in surveys. We use Ungdata surveys from 218 municipalities during the period 2017–2019.

Organizational Data

Organizational data are provided by a population survey, fielded and completed in spring 2022, of local voluntary organizations that are part of the Norwegian Register of Non-Profit Organizations. Voluntary organizations are required by public authorities to register to apply for public subsidies and Value Added Tax (VAT) compensation. There were 62,150 units in the register as of January 1, 2022. The survey was carried out electronically, with a questionnaire sent out by e-mail. Of the 62,150 units in the register, 35,892 had registered e-mail addresses for contact persons. A total of 5133 responses were received, comprising 16.1% of a corrected online sample of 31,911. If we compare the net sample with the register, we find an overrepresentation of associations within art and culture, hobby and leisure, and rights and social work, and an underrepresentation of those within sports, housing, and the local environment. Within the sample of 5,133 organizations, we retained for the analyses presented here only those organizations that answered having organized activities directed at children and youth, leaving us with a sample of 2255 organizations.

Municipal Data

Indicators of socioeconomic inequalities, cultural/ethnic heterogeneity and municipal expenses were taken from the Statistics Norway KOSTRA database.² The KOSTRA database is a national information system providing information about local government. All data in the KOSTRA database is electronically reported by municipalities and county municipalities to Statistics Norway each year.

Design

We analyzed these data at two levels. We examined the relationship between (a) contextual and organizational variables and (b) contextual and individual variables. Using a mixed-effect model, we conducted a two-level regression analysis, for respectively organizations and individuals. In the first model, organizations constitute level 1 (fixed effects) of the analysis and local communities (municipalities or urban districts in Oslo) constitute level 2 (random effects). In the second model, individuals constitute the level-1 while local communities constitute the level-2. We included interaction effects of those contextual or level-2 factors with individual and organizational variables in the models in order to examine cross-levels effects i.e., the extent to which contextual variables moderate organizational or individual factors. We present a the null-model to assess the percentage of variance attributable to the local community (level-2) factors.

Variables

The dependent variables are first, a measure capturing organizational inclusiveness (the inclusion index) and second, a measure of children and youths’ participation in organized leisure activities.

The dependent variable at the organizational level, the degree of inclusiveness of the organization is measured through an inclusion index that aggregated the answers to a 7-item question battery on a 10-point scale (ranging from 1 to 10), which were answered by respondents to the organization survey. This index is thus based on self-reported data, but by surveying organizations, we receive information about their activities that it would be impossible to uncover if only using text data, expert surveys, or a (naturally) limited number of qualitative interviews. Still, the respondents may have normative or strategic reasons to distort their answers or may have problems understanding how to answer accurately. We implemented measures to remedy such biases by carefully designing the survey and piloting the questionnaire among selected practitioners. The question battery was comprised of the following items that captured the extent to which organizations are inclusive in terms of being active in recruiting adolescents who do not participate, reducing the costs of participation, and opening activities to children and youth with diverse backgrounds (see Table 1).

Table 1. Summary of statistics items in the inclusion index (sorted).

Item	Mean	Sd
v5 Prefer to give a good offer to those already active (recoded)	7.71	2.36
v4 We do not have resources to prioritize the recruitment of children and youth (recoded)	7.60	2.51
v3 Come from homes with good economy (recoded)	6.63	2.37
v7 Participate even if the parents do not pay member fees	6.84	3.33
v2 Willing to adjust activity	6.17	2.76
v1 Work toward recruiting children and youth that do not participate	5.42	3.06
v6 Work toward recruiting children and youth with minority background	4.29	2.69

Note that the word inclusion is not actually used in the seven statements, and the question battery was asked in the context of recruitment. We thus argue that the statements together capture organizational inclusiveness without the organization actually or necessarily thinking about it in that way. Table 1 shows the distribution of answers to these items. The dependent variable, the inclusion index, at the organizational level, was computed as the additive mean of these seven items.

The independent variables at the organizational level are indicators of resources measured for each organization that participated in the survey in the respective municipality. Resources were measured in terms of the number of volunteers who work regularly for the organization (count variable) and whether the organization has paid employees (dummy variable, yes = 1).

The dependent variable at the individual level is one dichotomous variable indicating adolescents (13-19) participation in three types of activities organized by voluntary organizations: sports, cultural activities, and religious activities.

The independent variables at the individual level are indicators for Gender (girls = 2), Grade level (5 grades) as a proxy for age, and Socio-economic status. The Socio-Economic Status (SES) index, capturing both economic and cultural resources, is computed based on several variables related to family affluence and parents’ education. The index ranges from 0 to 3, with 0 representing the lowest socio-economic status. The variables used to calculate the SES index are described in Table 2.

Table 2. Indicators in the Socio-Economic Status (SES) index.

Question	Items	Points
Does your family own a car?	No Yes, one Yes, two or more	0 2 3
Do you have your own bedroom?	Yes No	3 0
In the last year, how many times have you gone on vacation to a place with your family?	None Once Twice More than twice	0 1 2 3
How many computers does your family own?	None One Two More than two	0 1 2 3
Number of books in household	No books Between 100 and 500 More than 1000 books	0 1,5 3
Parental education	Neither parents have higher education One parent has higher education Both parents have higher education	0 1,5 3

The calculation of the SES index involves recoding the variables to assign points according to certain guidelines. The points for each variable are then summed up and divided by 6 (the total number of variables) to obtain the SES index.

The independent variables at the municipal level (level 2) are indicators of socioeconomic inequalities, ethnic and cultural heterogeneity, and local public policies. Regarding socioeconomic inequalities at the community level, we used the Gini index. Ethnic and cultural heterogeneity were measured by the percentage of the population with an immigrant background. Furthermore, we included an indicator of public policies oriented toward the inclusion of children and youth in leisure activities: the share of municipal public expenses that are dedicated to children. The descriptive statistics of all variables are provided in Table 3.

Table 3. Summary of statistics for the variables in the analyses.

Variable	N**	Mean	Std. Dev.	Min	Pctl. 25	Pctl. 75	Max
Organizational level
Inclusion index (DV)	1906	6.41	1.39	1.571	5.429	7.429	10
Number of volunteers*	1991	54.355	176.639	0	10	40	4500
Has paid employees	2240	0.264	0.441	0	0	1	1
Individual level Participation Leisure Activities(0/1, DV) Gender Socio-economic Status Grade level	628 678 628 678 628 678 628 678	0.75 1.49 2.26 3.02	0.43 0.49 0.42 1.53	0 1 0 1	1 1 2.01 2	1 2 2.58 4	1 2 3 6
Municipal level
Share population immigr-ant background	282	16.15	6.225	2.78	11.4	18.64	34.71
Gini index Post-tax income	282	23.484	2.73	18.3	21.5	24.3	32
Municipal expenditures children	282	0.821	0.258	0	0.6	1	2.3

Notes: *Due to the skewed distribution, we took the log of this variable when included in the regression analyses.

** In the analyses N varies depending on the variables and levels involved. Here we describe the variables in each data set.

Results

We present the results of a mixed-effects model with interactions that allow us to investigate individual, organizational and contextual factors influencing participation in organized leisure activities. We estimate two different models with organizations as level-1 variables and with individual variables as level-1 variables. Such fixed-effects models enable to distinguish between (a) between-level-2 or group variables effects and within-level-2 or groups effects. Contextual effects are the difference between between-effects and within-effects. One way of thinking of this is that the between effect shows the overall, average relationship, but using a mixed-effects model we can decompose it into a within (cluster) effect, and a between (cluster) effect, the contextual effect being the difference between within and between effects (Hamaker & Muthén, 2020).

Table 4 displays the results of the model with organizational variables as level-1 variables, with the organizational inclusiveness index as the dependent variable, municipal factors as level-2 variables, and organizational characteristics and their interactions with municipal factors as independent variables. We present the null model in column 1, the model with fixed-effects and organizational variables in column 2, the model with contextual factors in column 3, and the full model including level-2 variables in the fixed-effects portion, cross-levels interactions, and contextual factors in column 4.

Table 4. Mixed effect model: organizational inclusiveness index and contextual factors across municipalities.

	(1) Null model	(2) Fixed	(3) Fixed	(4) Fixed
		Effects	Effects	Effects
			Level-1	Level-1 Level-2
			Random	Predictors
			Effects	Cross-levels
				Interactions
				Random
				Effects
	Inclusion Index	Inclusion Index	Inclusion Index	Inclusion Index
Fixed-effects
Number of volunteers (ln)		0.106***	0.098***	0.284
		(0.028)	(0.029)	(0.303)
Paid employees (1)		0.406***	0.415***	0.787
		(0.077)	(0.079)	(0.827)
Municipal expenditures children				0.091
				(0.347)
Gini Index				-0.001
				(0.049)
Share population immigrant background				-0.001
				(0.022)
Number of volunteers (ln)# Gini index				-0.001
				(0.015)
Number of volunteers (ln)# Share population immigrant background				-0.005
				(0.006)
Number of volunteers (ln)# Municipal expenditures children				-0.026
				(0.108)
Paid employees (1) # Gini Index				-0.048
				(0.041)
				0.032*
Paid employees(1) # Share population immigrant background				(0.018)
Paid employees (1) # Municipal expenditures children				0.327
				(0.324)
cons	6.420***	6.021***	6.050***	5.997***
	(0.034)	(0.089)	(0.091)	(0.959)
Random Effects
Var(residuals)	1.916	1.828	1.808	1.802
	(0.064)	(0.064)	(0.064)	(0.063)
Var(constant)	0.134	0.011	0.000	0.000
	(0.019)	(0.017)	(0.000)	(0.000)
Municipal expenditures children			0.000	0.000
			(0.000)	(0.000)
Gini Index			0.000	0.000
			(0.001)	(0.001)
Share population immigrant			0.000	0.000
			(0.000)	(0.000)
N organizations	1906	1906	1906	1906
N Municipalities	291	282	282	282
ICC	0.007	0.006	0.000	0.00
	(0.009)	(0.009)	(0.000)	(0.000)

Sd. in parentheses.

*p < 0.1, ** p < 0.05, *** p < 0.01.

For mixed-effects models, the Interclass Coefficient (ICC) can be interpreted as the proportion of the total variance that is due to the variance between groups or level-2 differences (as opposed to within groups). The ICC of 0.007 for the null model decreases to 0.006 when introducing level-1 variables and to 0.000 when introducing contextual variables, indicating that most of the observed variability is occurring within municipalities and not between municipalities.

When looking at the coefficients of the contextual variables (level-2) in the third column of Table 4, the coefficients of the variables – the indicators of inequalities (Gini Index) and of ethnic/cultural heterogeneity (share of the population with immigrant background), as well as local public expenditures devoted to children – are very close to zero and display little explanatory power. Considering the fixed-effects part of the model, among the two organizational resource indicators (number of volunteers and paid employees) both variables are significantly and positively related to the inclusiveness index in models 2 and 3,

The interactions between the organizational (level-1) and contextual (level-2) variables (Gini index, Share population immigrant background) are used to test the community-level heterogeneity hypothesis (H1a) and the community-level inequality hypothesis (H2a). The share of population with immigrant background, contrarily to what is expected according to (H1a) has no moderating effect on the influence of the number of volunteers in the organization on organizational inclusiveness. However, it has a positive moderating effect on the influence of the presence of paid employees in organizations on organizational inclusiveness, indicating, in support of (H1b), that when ethnic/cultural heterogeneity is high, the presence of paid employees increase organizational inclusiveness. The level of inequality measured by the Gini Index does not moderate significantly the influence of the number of volunteers (H2a) and of paid employee (H2b) on organizational inclusiveness – contradicting (H2a) and (H2b). The interactions between Levels of volunteering and Paid employees with the level of Municipal expenditures for children are used to investigate the local public policy hypotheses (H3a) and (H3b). The results show no significant coefficient for these interactions and indicate that local public expenditures do not moderate organizational resources, in contradiction to what was expected according to (H3a) and (H3b).

In sum, we do not find any direct contextual effects of indicators of inequalities (Gini Index) or of indicators of ethnic/cultural heterogeneity (share of the population with immigrant background), or of local public policies on organizational inclusiveness. We do not find, either, that these contextual factors moderate the influence of organizational resources (volunteers and paid employees) on organizational inclusiveness, with the exception of the moderating effect of ethnic/cultural heterogeneity on the influence of having paid employees on organizational inclusiveness. In context of high cultural heterogeneity, increasing the number of paid employees has a positive influence on organizational inclusiveness.

Table 5 displays the results of the model with individual variables as level-1 variables, participation in leisure activities (sport, cultural and religious activities) as the dependent variable and individual characteristics as well as their interactions with aggregated factors at the municipal level as independent variables. We present the null model in column 1, the model with fixed-effects and organizational variables in column 2, the model with contextual factors in column 3, and the full model including level-2 variables in the fixed-effects portion, cross-levels interactions, and contextual factors in column 4.

Table 5. Mixed effect model: children and young’s levels of participation in leisure activities and contextual factors across municipalities.

		(1) Null Model	(2) Fixed	(3) Fixed	(4) Fixed
			Effects	Effects	Effects
			Level-1	Level-1	Level-1
				Random	Level-2
				Effects	Predictors
					Cross-levels
					Interactions
					Random
					Effects
		Participation	Participation	Participation	Participation
		Organized	Organized	Organized	Organized
		Leisure	Leisure	Leisure	Leisure
Fixed Effects			0.148***	0.154***	0.213***
Socio-economic status (SES)			(0.002)	(0.002)	(0.029)
Gender (girls)			0.000	0.003	0.003*
			(0.001)	(0.002)	(0.002)
Grade Level^i
9. Grade			0.007**	0.010**	0.010***
			(0.003)	(0.003)	(0.003)
10. Grade			-0.084***	0.084***	-0.084***
			(0.003)	(0.003)	(0.003)
First year High School			-0.186***	-0.186***	-0.186***
			(0.0003)	(0.0003)	(0.003)
Second year High School			-0.253***	-0.253***	-0.253***
			(0-003)	(0-003)	(0.003)
Third year High School			-0.314	-0.314	-0.314***
			(0.004)	(0.004)	(0.004)
Municipal expenditures children					0.017
					(0.027)
Gini Index					0.010***
					(0.003)
Share population immigrant background					-0.004***
					(0.001)
Share population immigrant background#SES					0.000
					(0.000)
Municipal expenditures children#SES					0.008
					(0.011)
Gini Index#SES					-0.003**
					(0.001)
Cons		0.572***	0.531***	0.502***	0.321***
		(0.196)	(0.006)	(0.008)	(0.078)
Random Effects
Var(cons)		0.101	0.001	0.000	0.000
		(0.010)	(0.000)	(0.000)	(0.000)
Var(residuals)		0.183	0.171	0.171	0.171
		(0.000)	(0.000)	(0.000)	(0.000)
Municipal expenditures children				0.000	0.000
				(0.000)	(0.000)
Gini Index				0.000	0.000
				(0.000)	(0.000)
Share population immigrant				0.000	0.000
				(0.000)	(0.000)
Inclusion Index Aggregated				0.000	0.000
				(0.000)	(0.000)
N Individuals		175,007	175,007	175,007	175,007
N Municipalities		160	160	160	160
ICC		0.355 (0.235)	0.009 (0.001)	0.006 (0.001)	0.006 (0.002)

Sd. in parentheses.

*p < 0.1, **p < 0.05, ***p < 0.01.

ⁱ 8. Grade (age 13–14) is the reference category.

The ICC of 0.355 for the null model decreases to 0.009 when introducing level-1 variables and to 0.006 when introducing contextual variables, indicating that most of the observed variability is occurring within municipalities and not between municipalities.

Individual variables, in the models, are significantly related to participation. The socio-economic status index is positively related to participation. Participation is negatively associated with age and decreases as the Grade level increases, indicating that adolescents tend to drop-out from organized leisure activities as they get older. Introducing level-2 variable in the fixed-effects portion of the model shows that public local expenditures devoted to children and inequalities (Gini Index) have a positive influence on individual participation.

At the contextual level, the contextual variables have a coefficient very close to zero and, given the low coefficient of the ICC, the results indicate that much of the variability is due to individual factors. Considering the cross-level interactions, only the interactions between Gini index and socioeconomic status (SES) has a significant negative coefficient, indicating, in support of (H2c) that contextual inequalities negatively moderate the influence of SES on participation i.e., reduce individual participation for all levels of SES. The non-significant coefficients of the remaining interactions indicate that (H3c) is not supported.

Discussion and Conclusion

In this study, we investigated the extent to which contextual factors at the municipal level – such as higher degree of cultural and ethnic heterogeneity and higher levels of socioeconomic inequalities – impact the demand and supply of leisure activities, considered in terms of individual participation and organizations’ degree of inclusiveness. We also considered the impact of local public policies aimed at fostering inclusion in organized leisure activities, in terms of municipal expenditures. Overall, our results show that at the organizational level, while direct effects of inequality and cultural heterogeneity on organizational inclusiveness are not observed, they do influence how organizational resources (such as paid employees) impact inclusiveness. Municipal expenditures dedicated to children do not directly enhance organizational inclusiveness. At the individual level, while individual characteristics, especially SES, play a significant role in determining participation in leisure activities, municipal-level factors do not appear to directly influence participation in organized leisure activities. Still, in high inequality contexts, individual participation is reduced for all levels of SES.

The results from our mixed-effects models do not show support for the community-level heterogeneity hypotheses (H1) nor the community-level inequalities hypothesis (H2), according to which contextual factors negatively affect the degree of inclusiveness in leisure activities by voluntary organizations – i.e., the diversity of their members in terms of socioeconomic resources and cultural background, or individual participation in these activities. Yet, the results show that these contextual factors moderate organizational inclusiveness (contextual ethnic and cultural heterogeneity positively moderates the effect of paid employees on inclusiveness) as well as individual participation (contextual inequalities negatively moderate the effect of SES on participation). Local public policies did not appear to have neither a direct nor a moderating effect on both organizational inclusiveness and participation.

Thus, the results show some evidence that higher levels of inequalities affect organizational resources and inclusiveness or moderate the influence of socioeconomic status. The results show evidence that cultural and religious heterogeneity moderate the impact of organizational resources (paid employees). A possible explanation for this finding is the fact that Norwegian society is, in comparative standards, quite equalitarian when it comes to income inequalities after taxes, as well as being a welfare society with an extensive welfare state. Despite some localities being more deprived than others, redistributive and welfare policies may contribute to securing the necessary minimum level of resources in communities for enabling voluntary activities.

Additionally, local public policies do not appear to have any relationships with individual participation and organizational inclusiveness. Due to the decentralization to the municipal level of the responsibilities for many welfare policies characterizing the Norwegian welfare state (some have suggested that it would be more correct to talk about welfare municipalities than of the welfare state in Norway), there exists a diversity of local policies aimed at children and young people. Yet, this diversity of local policies is not reflected in our data, in spite of these local policies being aimed at reinforcing the supply of voluntary organized leisure activities by subsidizing voluntary organizations.

Our findings, indicate that contextual factors have little explanatory power, and that most of the variance characterizing participation in leisure activities across municipalities in Norway is related to individual factors, most importantly differences in socio-economic resources (income, education and cultural capital). These findings have implications for both public policies and voluntary organizations. First, our results emphasize that public policies supporting inclusive organized leisure activities for children and young people do not seem to be effective in levelling down socioeconomic inequalities that influence participation. Second, differences in individual resources appear to explain most of the differences in levels of participation across municipalities both directly (individuals with higher socio-economic status are more likely to participate) and indirectly (the interaction between socio-economic inequalities and SES decreases participation). This indicates that public policies aimed at the inclusion of children and young in leisure activities should, to a greater extent, focus on reducing socio-economic inequalities at the individual level and community levels. Our results do not find evidence that local policies aimed at supporting the supply of voluntary organized leisure activities impact on these inequalities. Achieving the objectives of broad inclusiveness stated in Norwegian policies for leisure activities for children and young would probably entail, in addition to supporting the supply of activities, to address the deep sociological mechanisms that generate individual disparities of socioeconomic resources and lead to non-participation.

Our study has, however, limitations. First, we do not employ direct measures of organizational resources, such as income and turnover, in our analyses, since too few organizations reported these measures. Second, our inclusiveness index consists only of self-reported items that may be biased, in contrast to more objective measures of inclusiveness. Finally, the community level of analysis is delimited by administrative borders, such as the municipality or the urban district. Yet, community-level effects may be present at lower levels of social organizations, such as neighbourhoods. On the other hand, people live in fuzzy spaces that are not necessarily delimited by their neighborhood (Petrović et al., 2019). Future scholarship would need to further explore linkages and effects at the individual, organizational, and contextual levels and explore different (lower) boundaries of the local context. Future research needs also to explain variations in participation rates between municipalities by inquiring into possible factors accounting for differences across municipalities.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work was supported by Norges Forskningsråd [grant number 320702].

Notes

1 See also https://www.ungdata.no/english/

2 See https://www.ssb.no/en/statbank/list/kostrahoved