Are immigrant youth involved in voluntary organizations more likely than their non-immigrant peers to be engaged in politics? Survey evidence from Norway

The aim of this paper is to examine the link between associational involvement and political participation among youth in Norway and how this link varies between boys and girls of an immigrant or non-immigrant background. We examine the relationship between activities in instrumental and expressive types of voluntary organizations. The paper draws on the survey Ung i Oslo 2015 (Young in Oslo 2015), which was carried out among students in almost all high schools (30 schools) in Oslo, Norway. The results indicate that those who are actively involved in socially oriented organizations are most likely to participate in politics. However, the relationship between organizational activity and political participation is often stronger among youth of immigrant background and is visible even among those who are active in less socially oriented organizations. Most notably, the results suggest that being active in a sports organization is positively related to political participation among immigrant youth. We found no such effect among non-immigrant youth. In accordance with Robert Putnam’s theoretical framework on social capital, the paper discusses why youth of immigrant background seem to benefit more, politically speaking, from associational involvement.


Introduction
Citizens of immigrant descent constitute a growing share of the populations in most western democracies, thereby increasing their political significance. Compared to the majority population, however, immigrants coming from Asia and Africa, in particular, are less likely to engage in politics (Bergh et al., 2014;Bhatti et al., 2014;Heath and Khan, 2012;Qvist, 2018;Svedberg et al., 2010). As far back as the 1960s, empirical studies have shown that members of voluntary associations are more politically active and interested than non-members (Almond and Verba, 1989). Voluntary organizations may work as both 'schools' and 'pools' of democracy (cf. Tocqueville, 2000); Van Ingen and Van der Meer, 2016) in two ways. First, participation in voluntary organizations may help individuals build 'bridging' social capital (Putnam, 2000) as well as enable a link to governance structures, thereby producing democratic outcomes by facilitating members' political engagement (Inglehart, 1977;Kuhnle and Selle, 1992). Second, politically engaged citizens are more likely to participate in general, and both voluntary organizations and political participation offer pools of opportunities for these citizens.
The question addressed here is the extent to which the relationship between voluntary and political engagement varies between youth of immigrant and majority background and between boys and girls. By differentiating between several types of voluntary organizations, we examine the association between political engagement and participation in more or less politically oriented voluntary organizations and how this varies between different groups of youth. In particular, as youth of immigrant background lack the necessary network and family resources, we are interested in whether their engagement in voluntary organizations is more strongly related to their political participation. Previous research has primarily been concerned with the adult majority population, and we have limited knowledge on participation patterns among the youth. As citizens of immigrant descent become a more significant group in most Western democracies, there is a need for knowledge about their participation patterns in the early phases of political socialization and how these patterns potentially deviate from those of the majority population. The article draws on survey data from Ung i Oslo 2015 (hereafter Young in Oslo 2015) (Andersen and Bakken, 2015), conducted among 16-19-year-old students in almost all high schools in Oslo, Norway. 2009;Putnam, 2000;Stolle and Rochon, 1998;Torpe, 2003). Other scholars have questioned such a general relationship between voluntary engagement and political participation, arguing that the relationship is mainly visible among volunteers engaged in instrumental organizations, such as interest and activist organizations (Dekker, 2014;Hanks, 1981;Van der Meer and Van Ingen, 2009).
In relation to migrant populations, Putnam (2000: 362) argues that participation in voluntary organizations can be 'the glue that bonds those who would otherwise be divided along racial and ethnic lines'. In the last decades, there has been a growth in voluntary organizations based on ethnic and religious identity, and several scholars have discussed the relevance of the social capital hypothesis based on ethnic minority associations. Fennema and Tillie (1999) and Togeby (2004) found that participation in these types of organizations builds social trust and tolerance, which in turn creates the basis for political participation. However, the social capital effect varies between different ethnic groups.

Political socialization among teenagers
The above-presented research focused primarily on adult populations. Regarding political socialization among young people in general, recent contributions have examined the democratic effect of special programmes, such as educational and local community programmes (see Hart et al., 2007). However, Quintelier (2008) summarized the three main findings from a review of studies on the relationship between voluntary engagement and political participation among young people. First, associational involvement helps socialize young people into political decision-making; second, they develop politically relevant attitudes; and third, they acquire politically relevant skills. More specifically, voluntary engagement broadens the sphere of interest and concerns among young people (Knoke, 1986), socializes them into political participation (Hanks and Eckland, 1978) and makes them feel more efficacious through participation in youth councils, environmental groups, youth fora, etc. (Eden and Roker, 2002;Verba et al., 1995). In her own study of members in different organizations, Quintelier (2008) found that cultural, deliberative and help organizations are more successful than religious, ethnic and youth groups in socializing young people into politics. This finding might be explained by the organizations' aims, as cultural, deliberative and help organizations aim to aid society, which is in line with political participation. In contrast, expressive organizations are more entertainment-oriented (Quintelier, 2008: 365) and, in line with previous research, seem to be more weakly related to political participation (Bowler et al., 2003;Hanks, 1981).
Previous research on political socialization has focused on the family's role, showing that children are more likely to be civically and politically active if their parents are also involved (Plutzer, 2002;Verba et al., 1995). Even though education is one of the most important predictors of voter turnout (Smets and Van Ham, 2013), a recent Finnish study of parental influence on young people's turnout showed that parents' voting behaviour appears to be more important for young people's turnout than parents' educational level (Gidengil et al., 2016). The same pattern was found among young adults approaching their thirties and when controlling for the young adults' and their parents' education and income. This is in line with previous research indicating that parents who believe that political involvement is important often directly pass this belief onto their children through political discussions within the family (Andolina et al., 2003;Verba et al., 1995).
Lacking childhood experiences from the host country, many immigrant parents will have a disadvantage, will be less familiar with the political system and organizational landscape and, therefore, less likely to pass on behaviours that encourage political and associational participation (Eimhjellen et al., 2018;Voicu, 2014;Ødegård et al., 2014). In addition, the lack of strong and significant social networks among immigrants has been deemed a substantial barrier to involvement in secular volunteering (Qvist, 2018). In a nationally representative longitudinal survey of adolescents in the US, Humphries et al. (2013) argued that traditional models of adolescent political socialization may not accurately describe the experiences and diversity of children of immigrants. In effect, other institutions, such as voluntary associations, may supplement parental socialization processes and become more important as an arena for political socialization among children of immigrants compared to young people of a majority background.
Several scholars have found country of origin to be a significant predictor of multiple factors regarding political participation in the country of settlement (Buecker, 2005;Fennema and Tillie, 1999;Martinello, 2005;Togeby, 2004). For example, McIntosh and Youniss (2010) maintain that immigrants transmit social norms of civic and political practices and engagement from their country of origin to their descendants. Kim and Amnå (2015) argue that parents serve as a hub through which immigrant youth connect with their country of origin in terms of civic engagement. Using Swedish data, they found that the level of civic competence was significantly higher among Kurdish immigrants than among Iraqi immigrants and even among the native Swedish population. The Kurdish youth also reported significantly higher levels of perceived social norms of civic/political engagement from their parents than the Iraqi and Swedish youth. This is in line with Takle and Ødegård's (2017) finding that the heritage of the country of origin among immigrant youth is a promising factor from which to interpret civic and political engagement in the national and transnational contexts.

The classical Norwegian model of voluntary organizations
Historically, Norway has experienced high levels of participation in voluntary organizations (Eimhjellen et al., 2018). This has been linked to a specific 'Scandinavian' hierarchical institutional model of participation, whereby the local association, a self-governed, membership-based democratic entity, is the locus of activity (Enjolras and Strømsnes, 2018). In recent years, however, the voluntary sector has experienced a marked decrease in local organizations with a political agenda and a corresponding increase in cultural and recreational associations (Arnesen et al., 2016). Despite the dominance of expressive organizations, voluntary organizations in Norway are still viewed as playing an important political role and as representing a mediating structure between individuals and the state. Compared with many other countries, and in contrast to much Anglo-American citizenship theory (Trägård, 2007), Norway -and other Nordic countries -have large public and voluntary sectors.
Traditionally, voluntary organizations in Norway have recruited broadly and have been more socially inclusive than in many other countries (Enjolras and Strømsnes, 2018). This strategy of recruitment has reduced widespread political marginalization, as one of the aims has been to recruit members from different social classes as well as mobilizing residents of rural areas -not only from large cities and higher social classes (Inglehart, 1977). This way, different social groups have had an opportunity to make their voices heard. However, the extent to which the various voices were heeded depended on the resources that the organizations controlled -or the sympathy that they were able to mobilize among the general public (Tranvik and Selle, 2007). In the last decade, several new organizations and networks have been established for the purpose of mobilizing immigrants and their descendants (Eimhjellen et al., 2018;Ødegård et al., 2014). Local and national initiatives have been developed to help these networks formalize their structures in line with the classical Scandinavian model of voluntary organizations, to expand the opportunities for participation in local and national political processes and to receive government funding (Takle, 2015;Takle and Ødegård, 2017). Several scholars have concluded that members are socialized into a democratic culture through the immigrant organizations' internal democracy procedures (Bay et al., 2010;Bengtsson, 2010;Predelli, 2008;Takle, 2015).
In our empirical analyses, we distinguish between activities in instrumental and expressive organizations and examine the link between voluntary engagement and political participation (see data and method section). The instrumental category includes outward-looking organizations with a political agenda; for example, humanitarian and environmental organizations. Expressive organizations are less outward-oriented and focused on producing leisure activities (e.g. music and theatre, scouting, sports) or create a meeting place for youth with the same faith and/or ethnical background. Even though most voluntary organizations in Norway follow the hierarchical and democratic structure of the 'Norwegian model', we expect the relationship between organizational engagement and political participation to be stronger among youth who are active in instrumental organizations, which are more likely to attract politically interested youth. However, following the perspective that voluntary organizations may also be 'schools of democracy', we expect the relationship to be stronger among youth of immigrant background, who often have fewer initial family and network resources compared to the majority population. These expectations are formulated into the following hypotheses: H1: Youth who are active in instrumental organizations are more likely to participate in politics than those who are active in expressive organizations.
H2: The relationship between organizational engagement and political participation is generally stronger among youth of immigrant background than those of non-immigrant background.
It is important to note that the data analysed cannot determine whether voluntary organizations work as 'pools' or 'schools' of democracy. This will be discussed more thoroughly in the discussion section

Data and method
We drew on the survey Young in Oslo 2015 among high school students (aged 16-19) in Oslo (Norway), carried out by Norwegian Social Research (NOVA) in the winter/early spring of 2015. The survey was unique in the Norwegian context, as it was distributed to virtually all high schools in Oslo. It was endorsed by school authorities and was carried out during school hours. Consequently, it was one of very few studies to include a large share of people of immigrant background and is the most representative survey of the student population in Oslo. Of the total student population in Oslo, 72% participated in the survey, returning a total of 10,932 answers. The response rate was highest among first graders (VG1) and poorest among third graders (VG3). Declining response rates throughout high school are attributable to at least three factors. First, while most students attend the same mandatory classes in VG1, in VG2 and VG3 the students choose between several different classes, making it more difficult to carry out surveys among all students simultaneously. Second, starting from VG2, students attending vocational education have fewer school hours and more practice. Third, school drop-out rates increase throughout high school. Due to variations in the response rates between the grades, descriptive analyses are weighted according to real grade sizes.
The survey respondents were randomly divided into three sub-samples, one of which was asked questions about political participation and organizational activities. These questions came late in the survey, and it was explicitly stated that they were 'voluntary', resulting in some degree of survey dropout. The final target sample consisted of those who answered questions about both political participation and organizational activity (n ¼ 2575). Compared to the full sample of high school students participating in Ungdata, the sub-sample deviated in terms of immigrant background. The respondents of immigrant background from Africa (-2.7 percentage points (pp)) and Asia (-3.1 pp) were somewhat underrepresented, while those of a non-immigrant background (þ5.7 pp) were overrepresented. It is not unlikely that those who did not complete the survey also participated less in political and organizational activities. As such, we must be cautious when interpreting absolute numbers of participation. This may, however, not be critical, as we are mainly interested in the differences between those who participate in politics and those who do not. As we will see, the majority of students did not participate at all. Still, we cannot leave out the possibility that non-response bias among immigrant youth might have affected the results. Deviations regarding other central background characteristics were small. Compared to the full survey sample, boys (-2.4 pp) and VG2 students (-2.1 pp) were somewhat underrepresented.

Dependent variable
We constructed an additive index (0-6) of political participation based on six single items (alpha ¼ 0.71): participation in the youth wing of a political party, participation in another political organization, involvement in a youth council, taking part in signature campaigns, attending political events (public meeting, demonstrations, etc.) and boycotting certain products (Table 1 below contains descriptive statistics on participation in each activity). The value '0' means no participation, while '6' means participation in all six activities. 1 Since the index is skewed (very few pupils have the maximum value), we performed additional analyses with the natural logarithm of political participation (results retrievable upon request). This yielded more or less identical results as in the use of the original variable, and since standard ordinary least squares (OLS) regression is more easily interpreted, we present this in the results section.

Independent variables
Our main explanatory variables were organizational activity and immigrant background. The respondents were asked whether they were active or had previously been active in 12 different types of organizations. Present activity was coded as '1', and past or no activity was coded as '0'. The different types of organizations where grouped into four: 'socially oriented' (e.g. environmental, humanitarian), 'religious/ethnic' (e.g. Christian, Muslim, other non-Christian and 'ethnically based'), 'leisure' (e.g. hobby, culture) and 'sports'. Socially oriented organizations are examples of 'instrumental' organizations, which are outward-looking and with a political agenda. The three other types are examples of expressive organizations, mainly concerned with facilitating activities and a meeting place for its members. With the exception of some religious or ethnic organizations, most voluntary organizations for children and youth in Oslo consist of members of different ethnic and religious backgrounds. Following previous studies in Norway on youth civic participation, which emphasized the intersectional relationship between gender and immigrant background (Ødegård et al., 2016;Ødegård and Fladmoe, 2017), we constructed four dummy variables, combining gender and immigrant background: (a) nonimmigrant boys; (b) non-immigrant girls; (c) immigrant boys; and (d) immigrant girls. 'Immigrant' denotes immigrant background, which was constructed on the basis of parents' immigrant status. Thus, 'non-immigrant' comprises both immigrants and youth born in Norway of immigrant parents. Altogether, 25% of the respondents reported that both their parents were immigrants. According to official statistics, approximately 35% of the population aged 16-19 in Oslo is of immigrant background (either immigrants or children of immigrants), meaning that respondents of immigrant background participating in the survey were somewhat underrepresented compared to the total population of immigrant descent in Oslo. In the multivariate analyses, we added interaction terms between gender/immigrant background and organizational activity.
The parents originated from more than 70 different countries, and the largest groups were from Pakistan (n ¼ 110), Sri Lanka (n ¼ 71) and Somalia (n ¼ 51). Due to the limited statistical power of the single countries, we differentiated between non-immigrant and immigrant youth in the analyses.
In the multivariate analyses, we also controlled for a set of other relevant characteristics. Youth in their formative years are influenced by both their peers and parents. We therefore included the variable 'peer effect' based on the item: 'Do the following things affect your social status within your group of friends? Being interested in politics and societal issues (1 ¼ decreases status a lot -5 ¼ increases status a lot)'. To control for parental influence, we included two single items and socioeconomic status (SES). The single items measured how often the respondents discussed politics with their parents (1 ¼ Never -4 ¼ Daily) and whether their parents were or had been active in voluntary organizations (1 ¼ Yes). SES was constructed on the basis of three questions: (a) parents' level of education; (b) number of books at home; and (c) a battery of items measuring material goods in one's family (family has at least two cars; respondent has his/her own bedroom; respondent has gone on at least two travel holidays with family during the past year; family has at least two computers at home). The SES scores were then standardized in five equal-sized categories (quintiles). Finally, due to decreasing response rates throughout high school, we controlled for grade. Table 1 summarizes the descriptive statistics for all the variables used in the analysis.

Results
We present our results in two steps. First, we look at the descriptive statistics on political participation and organizational activity among non-immigrant boys and girls and immigrant boys and girls. Second, we estimate the relative relationship between organizational activities and political participation. The purpose is to determine whether the organizational 'effect' on political participation is different between non-immigrant boys/girls and immigrant boys/girls. Table 2 displays the various forms of political participation by gender and immigrant background. The table shows that participation in signature campaigns (31%) is the most common political activity among high school students in Oslo, followed by participation in political events and boycotting certain products (both 19%), and participation in the youth wing of a political party (12%), youth council (9%) and other political organizations (8%). Furthermore, the numbers suggest that girls (both non-immigrant  and immigrant) had a higher average participation rate than boys, particularly in signature campaigns and political events. Immigrant youth had somewhat higher participation rates than non-immigrant youth in youth councils. Non-immigrant boys were least likely to participate in the youth wing of a political party. This runs contrary to established evidence among adult immigrants, who are less likely to be party members (Bjørklund and Bergh, 2013) and less likely to vote in elections (Bergh et al., 2014). As described above, the response rate was somewhat lower among immigrants. Nevertheless, the data suggest that the children of immigrants in Oslo did not deviate much from non-immigrants in average participation rates. Moreover, overall, gender differences were larger than differences relating to immigrant background. Next, we look at our main explanatory variable, organizational activity. Table 3 displays the share of youth reporting that they have taken part in four types of organizations.
In contrast to political participation, differences relating to immigrant background were much more profound in the realm of organizational activity. Overall, about 50% of the non-immigrant youth and 30% of the immigrant youth were active in at least one organization, which is in line with previous studies on the adult population in Norway (Eimhjellen, 2016;Wollebaek and Sivesind, 2010). There were, however, notable differences between the various organizational categories. With a share of 35%, non-immigrant boys had a particularly high participation rate in sports organizations. In contrast, only 7% of immigrant girls participated in such an organization. Non-immigrant girls (18%) were much more likely than any other group to be active in socially oriented organizations. In these organizations, only 6% of immigrant boys were active. Furthermore, given that the least active students may have dropped out of the survey and that immigrant students had a higher dropout rate, the real differences between immigrants and non-immigrants might be even larger.

The relationship between organizational activity and political participation
The main question is whether there is a relationship between organizational activities and political participation and whether this relationship varies between youth of immigrant and non-immigrant backgrounds. Table 4 summarizes the results from a range of regression models, which estimated the relationship between organizational activity and gender and immigrant background and political participation, controlling for relevant background characteristics.
For each of the four organizational categories, we included organizational activity, gender and immigrant background and the interaction terms between these variables. The interaction terms tell us whether the organizational 'effect' on political participation was different among the four groups (non-immigrant/immigrant boys/girls). In all the models, non-immigrant girls, who are generally the  most politically active, were treated as the reference category. All models controlled for peer effects, political discussion at home, parents who are active in organizations, SES and grade. The results suggest a significant and positive relationship between organizational activity in socially oriented and leisure organizations and political participation. The relationship was strongest for socially oriented organizations. Young people who were active in these organizations were more likely to participate in political activities than those who were not. The organizational 'effect' did, however, vary among the four groups of youth. The interaction terms between gender/immigrant background and organizational activity showed the relative effect for each group (compared to non-immigrant girls). The interaction terms suggest that organizational activities in sports organizations were more strongly related to political participation among students of immigrant background than those of non-immigrant background. Thus, there was an organizational 'effect' of being active in sports organizations on political participation, but only among students of immigrant background. Among boys in general, but especially among those of immigrant background, we see the same relationship for activities in leisure organizations. Finally, the table reveals an organizational 'effect' among immigrant youth and non-immigrant boys in religious/ethnic organizations.
The control variables predicted political participation largely as one would expect, with positive effects of peers and family socialization (political discussion at home and parents who are active in organizations). Interestingly, SES did not predict political participation, which was partly due to multicollinearity: excluding peer effects and family socialization from the models, SES positively predicted political participation.
In sum, the regression analyses suggest that on a general level there was a positive relationship between organizational activities in socially oriented and leisure organizations and political participation. Among students of immigrant background, we see an additional effect of some forms of organizational activities, most notably participation in sports and religious/ethnic organizations.
The main results from the regression models are illustrated in Table 5 as predicted values. The entries in the tables show the average number of political activities for each group (non-immigrant/ immigrant boys/girls and country of origin) as contingent upon whether or not they had participated in different organizations, holding all control variables at mean value. Table 5 shows that, among youth who did not participate in any organization, non-immigrant girls (0.95 political activities) were most likely to participate in political activities, while immigrant boys (0.64) were least likely to participate in political activities. Among students who were active in organizations, the average number of political activities increased more for immigrant students than for non-immigrant students. For instance, nonimmigrant youth who were active in sports organizations were no more likely to participate in politics Entries are predicted probabilities from regression models (see Table 4), with all control variables set to mean value.
than non-immigrant students who were not active in any organization. Among immigrant youth who were active in sports organizations, however, the average number of political activities increased from 0.64 to 1.71 (immigrant boys) and from 0.79 to 1.14 (immigrant girls). A similar effect, albeit stronger for immigrant girls than minority boys, was evident among students who were active in religious/ ethnic organizations. Finally, it is worth emphasizing the strong 'organizational effect' of participation in socially oriented organizations. Across the various groups, students who were active in these organizations participated, on average, in more than two political activities.
Discussion: a stepping stone for political integration In this paper, we examined the link between voluntary engagement and political participation among 16-19-year-old boys and girls of immigrant and non-immigrant backgrounds in Oslo, Norway. Citizens of immigrant descent constitute a group of growing political significance in most western democracies, but they are generally less involved in associational and political life. Lacking a long history of network and family resources makes it particularly relevant to study the potentially reinforcing relationship between associational involvement and political participation among youth of immigrant descent. Overall, the results suggest that young people who are involved in voluntary organizations are more likely to participate in political activities. However, this relationship varies across types of organizations. In line with the first hypothesis, the relationship is stronger among youth involved in instrumental organizations (i.e. socially oriented) compared to those involved in expressive organizations (i.e. leisure, sports or religious/ethnic). Instrumental organizations are concerned with issues such as animal and human rights and the environment and climate, which is outward-oriented and -in the tradition to the 'Scandinavian model' -closely linked to politics (Enjolras and Strømsnes, 2018). Previous research has shown that such organizations produce social trust and political commitment among their members (Vogel et al., 2003;Wollebaek and Selle, 2002), and they do so to a higher degree than expressive organizations, which are mainly 'inward-oriented' in their production of goods or activities for their members (Wollebaek and Selle, 2002). In line with Van Ingen and Van der Meer (2016), it is also reasonable to assume that instrumental organizations are more likely than expressive organizations to attract politically interested youth.
Furthermore, in line with the second hypothesis, the results suggest that the relationship between associational involvement and political participation is often stronger among immigrant youth than among their non-immigrant peers and that the relationship often holds even for students who are involved in expressive organizations. Even though the relationship between associational involvement and political participation was weaker and more complex across gender and immigrant background than what was found for instrumental organizations, the results suggest that a larger share of politically engaged minorities are also active in expressive organizations. Most notably, the results suggest that active participation in a sports organization, which is overall the most popular activity among youth in Norway, is positively related to political participation among immigrant youth and not among nonimmigrant youth. Previous research from Norway has demonstrated that participation in sports organizations among youth in general has a weaker effect on social capital than in voluntary organizations in general (Seippel, 2006). It is, thus, noteworthy that this relationship appears to be different for immigrant youth. We found a similar relationship among boys who were active in leisure organizations.
These findings suggest that the 'Scandinavian model' of voluntary organizations may still play an important role by functioning as a mediating structure, especially between politically marginalized groups, such as immigrant youth, and political institutions. This is also the case for involvement in religious/ethnic organizations, indicating that young people involved in such organizations, especially immigrant girls, were more likely to participate in political activities. This finding somewhat challenges the social capital tradition of Putnam (2000). In such organizations, building relationships among members of networks based on religious, ethnic or cultural similarities seems to operate in terms of 'bonding' rather than 'bridging', as they do not necessarily strengthen relationships with individuals and networks that are different from themselves (Putnam, 2000). Similar findings have been observed in Norway, where religious and ethnic organizations, to a limited degree, have formal structures or institutional relations with the wider community or civil society in general (Ødegård et al., 2014). Qvist (2018) shows that immigrants' lack of informal networks seems to be a significant factor in explaining the gap in secular voluntarism. In his study of young people's activism, Sakellariou (2018) challenged the traditional understanding of religious organizations as solely bonding-oriented. He found that religion takes up a secondary role in young people's motivations and that togetherness and social activism may become more important than religious affiliation in sustaining participation. Sakellariou further argues that both bonding and bridging social capital are cultivated within these organizations, serving as a reasonable interpretation of our data. We found the relationship between religious/ethnic associations and political participation to be significant but weak, appearing strongest among immigrant girls.
Previous studies have repeatedly found that women are more likely than men to be involved in charity and religious organizations and activities and that young women are more volunteer-oriented but less attentive to political affairs than their male peers (see Galligan, 2015). By combining gender and immigrant background, our findings provide a more complex pattern. In general, young people of immigrant background are significantly less likely to participate in voluntary organizations compared to their non-immigrant counterparts, and girls appear to be the least active group. However, the immigrant and gender gaps are mainly visible in sports associations, where especially immigrant girls are underrepresented. In a previous study concerning sports participation among immigrant and nonimmigrant youth, Strandbu et al. (2017) found that much of the gap between youth of immigrant and non-immigrant backgrounds is due to differences in socio-economic resources and that religious denomination also explains, to some extent, the immigrant-non-immigrant gap among girls. This underlines the need for research to look at various factors so as to attain a better understanding of the gender differences in participation patterns, especially among immigrant groups.
Even though we found that youth of immigrant background are, in general, somewhat less involved in political activities, the general pattern masks important gender differences. Immigrant girls are significantly more involved in political activities than non-immigrant boys. We also identified an important difference in types of activities in which immigrant youth are more involved than non-immigrant youth in institutional political activities (participation in the youth wing of a political party, other political organizations and youth councils) but less involved in informal activities (signature campaigns, political events and boycotting). Previous research from Norway found that structural and cultural conditions influence informal political activities more than traditional and institutional activities (Ødegård & Berglund 2008). We also found the predictive power of family cultural resources to be stronger for informal activities than for institutional activities. In general, immigrant youth have fewer educational, cultural and economic family resources (Andersen and Bakken, 2015), and this seems to partly explain their lower levels of participation in informal activities.
In summary, the results suggest that associational involvement seems to be a more important 'stepping-stone' for political mobilization among immigrant youth than among non-immigrant youth. Based on the data, we cannot draw strong causal inferences on whether voluntary organizations work as 'pools' or 'schools' of democracy. However, combined with previous research suggesting that young people of immigrant background have less access to social network resources than their peers with Norwegian-born and -raised parents, these findings underline the potential significance of associational involvement for political socialization among citizens who are less integrated in the political system. For these citizens, voluntary organizations can contribute towards strengthening their social and institutional networks, which serve as a substitute for limited political socialization in the family (cf. Humphries et al., 2013).
Even though our findings might seem promising in terms of the political integration of immigrant youth, this study calls for further research on the relationship between various forms of associational involvement and political participation, which might be useful in understanding how to reduce inequality in processes of political socialization for youth of immigrant background. Consequently, Young in Oslo 2015 is limited as it insufficiently explains why youth of immigrant background seem to benefit more than non-immigrant youth in their political mobilization by attending expressive organizations. The categories of associations presented in the paper are broad and include a number of organizations with variations in scope, such as scouts, marching bands and youth clubs. Even though these are crossethnic organizations and are usually structured in line with the Scandinavian voluntary sector model, the categories might mask important internal variations between organizations, which are important in understanding the specific mechanisms driving civic engagement. In addition, by limiting associational involvement to respondents' reports on whether they attend a number of different organizations, we do not know whether they are active members or passive supporters, which might influence their civic engagement. Moreover, we do not know how long they have been involved in an organization and whether they attend special kinds of activities (intra-organizational or extra-organizational activities), which might have an impact on political mobilization. We also acknowledge that 'youth of immigrant background' is an oversimplified category, which could mask important differences in civic and political involvement among immigrants. Length of residence in the host country might be of relevance, even among young people, as well as differences among ethnic groups. Indeed, the data suggest significant variation across different immigrant groups, but the sample was too small to draw strong concluding inferences. Thus, further research is needed to attain additional knowledge of motivational factors relating to the transnational dimension of political engagement and political socialization among migrant families. Finally, the data used in this study was limited to the capital of Norway, the municipality with the longest history of immigration and the highest share of immigrant residents today. Would the patterns of political mobilization among immigrant youth vary in a different context, for instance, in a city with a shorter history of immigration? Further research is needed to improve our knowledge of such contextual factors.

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

ORCID iD
Guro Ødegård https://orcid.org/0000-0001-6418-9613 Note 1. Initially, we constructed two dependent variables -'long-term political participation' and 'short-term political participation', distinguishing between, on one hand, participation in the youth wing of a political party, participation in another political organization and involvement in a youth council and, on the other hand, taking part in signature campaigns, attending political events (public meeting, demonstrations, etc.) and boycotting certain products. Distinguishing between these two variables in regression models did not, however, yield substantially different results (which may be retrieved upon request). Moreover, for the sake of parsimony, we stuck to the single index of political participation in this paper.