Minority bonding to political self-efficacy: a case study on bonding and bridging social capital and political self-efficacy among adolescents in Finland

Hannuksela, Venla; Tiihonen, Aino

doi:10.3389/fpos.2023.1270065

ORIGINAL RESEARCH article

Front. Polit. Sci., 08 January 2024

Sec. Political Participation

Volume 5 - 2023 | https://doi.org/10.3389/fpos.2023.1270065

This article is part of the Research Topic Power Relations in Society View all 6 articles

Minority bonding to political self-efficacy: a case study on bonding and bridging social capital and political self-efficacy among adolescents in Finland

$\r\nVenla Hannuksela$ Venla Hannuksela^*

Aino Tiihonen

Department of Political Science, Tampere University, Tampere, Finland

While social capital is claimed to predict, for example, a higher level of trust, better health, and higher voter turnout, its uneven distribution causes concern in society. In this study, we examine which aspects of social capital are connected to political self-efficacy, which is an important predictor of political participation and associated with for example political interest. According to our knowledge, there is a research gap both in terms of studies focusing on analyzing the effect of minority status on political self-efficacy via social capital and studies analyzing these patterns among adolescents living their formative years. In this article we test this mediating effect of social capital among majority and minority adolescents since minorities can be expected to have relatively more bonding social capital and less bridging social capital. In the Finnish context, this is particularly interesting since the Swedish-speaking minority is praised for its high level of social capital that is suggested to explain a lot of its success ranging from better health to a higher turnout compared with the majority. Analyzing a nationwide dataset of 15–16-year-old students (N = 5,189) completing their final year of comprehensive education in Finland in a multilevel structural equation model, we discover that belonging to the Swedish-speaking minority seems to strengthen bonding social capital (social networks, and community). The civic participation, community, and social networks aspects of social capital are positively connected to the level of political self-efficacy, while generalized trust has no connection to it. All these findings indicate that societies should take measures to strengthen bridging and bonding social capital among both adolescents and minorities.

1 Introduction

In multicultural democracies, all societal groups, including young people and minorities, are expected, and required to have enough political self-efficacy–individual trust in one's own abilities to understand politics and take part in political processes–to have control over political decision-making, and therefore participate in it (Craig and Maggiotto, 1982; Finkel, 1985). Yet there are various power relations even in democratic societies. Minorities are always in a vulnerable position due to being dependent on the good will of the majority according to the majority principle (Strubell and Boix-Fuster, 2011), and young people, whether part of the majority or a minority, can be sidelined. Due to this structural setting, it is essential to aim to strengthen both minorities' and youth's political self-efficacy by actively including them in political processes. Therefore, more knowledge is needed on how minority youth develop political self-efficacy. One route to political self-efficacy is through social capital–an interplay of trust and engagement in social networks (e.g., Tingaard and Svendsen, 2009)–that can be enhanced, for example, by creating more opportunities to engage in a constructive atmosphere.

Even though political engagement is largely built in adolescence during the so-called formative years, most research on the relationship between social capital and political self-efficacy in general or on minority groups in particular is about adults (see e.g., Lillbacka, 2006; Bäck and Kestilä, 2009; Abbott, 2010; Kim et al., 2020), and few consider many different aspects of social capital as potentially having different effects (e.g., Lillbacka, 2006; Coffé and Geys, 2007; Bäck and Kestilä, 2009). In the field of political science, there is a research gap in existing research. Specifically, there has been limited investigation into the influence of minority status on individuals' political self-efficacy when mediated through the concept of social capital. Political self-efficacy refers to one's belief in one's ability to effectively participate in political activities, while social capital encompasses the resources and connections individuals have within their social networks. Despite the importance of understanding how minority status may interact with social capital to shape political self-efficacy, this linkage remains largely unexplored in the current literature. Furthermore, we think that it is important to study this relationship among minorities since previous studies have shown them to have stronger social capital from their own homogeneous communities compared with majorities, and, correspondingly, less social capital that connects them to larger society (Scholten and Holzhacker, 2009; Frost and Meyer, 2012; Uekusa, 2020). Our other concern regards to the lack of studies analyzing these patterns among adolescents, who are living their formative years. Previous research has predominantly focused on adults, when analyzing these mechanisms.

In spite of the observed research gap in studies analyzing social capital's mediating effect on political self-efficacy, some literature gives support for the assumption that social capital and political self-efficacy might benefit from each other. Bandura's (1977, 1997) seminal theory on the concept of self-efficacy, is built on the idea of individuals' own perceptions of their abilities to accomplish tasks and generally succeed in life. In political science, political self-efficacy is considered as a form of self-efficacy, but not a synonym for it. Political self-efficacy does not measure concrete actions, skills, or activity, and thus it should be kept apart from concepts such as political knowledge or political participation (see, e.g., Reichert, 2016).

Even though self-efficacy is a clearly broader concept than political self-efficacy, Bandura's thoughts can be considered to offer valuable insights into on how self-efficacy can contribute to individuals' political engagement and participation. As to social capital and Bandura's concept of self-efficacy, social capital can contribute to the development of self-efficacy. In addition, both social capital and self-efficacy can impact on an individual's ability and motivation to engage in various activities, including political participation. Regarding Bandura's thinking (Reichert, 2016), support, resources, and connections within one's social network, which are aspects of social capital, can bolster an individual's belief in their capacity to influence and contribute to their community or political processes.

In our article, we fill this gap by studying how belonging to the Swedish-speaking minority in Finland is connected to different kinds of social capital, and through them, to political self-efficacy among adolescents. In other words, we test the mediating effect of social capital on political self-efficacy among majority and minority adolescents. We argue that the language minority offers a favorable perspective for studying social capital's effect on political self-efficacy for two reasons. First, the acknowledged higher level of social capital among the Swedish-speaking minority when compared with the Finnish-speaking majority has been given as an explanation for various things, for instance life expectancy, health, and turnout (Hyyppä and Mäki, 2001; Saarela and Finnäs, 2004; Paljärvi et al., 2009). Thus, we aim to discover whether it also explains the previously discovered differences in the levels of political self-efficacy among the Finnish- and Swedish-speaking adolescents in Finland. Second, from a normative point of departure, it is not enough for a healthy democracy that people in general have the necessary prerequisites to participate, but all societal groups must possess them. From this perspective, it is of particular importance that minorities tend to rely heavily on their own networks with relatively strong bonding social capital, and thus they often have limited networks beyond their homogeneous group, compared with non-minority groups (Scholten and Holzhacker, 2009). Previous literature suggests that strong bonding capital especially develops within socially vulnerable communities as a byproduct of social inequality and as a strategy to combat stigma (Frost and Meyer, 2012; Uekusa, 2020). Therefore, if certain dimensions of social capital benefit political self-efficacy more than others, it may either support or damage minorities' political self-efficacy compared to majorities, depending on which these particularly important dimensions are.

Despite the vast amount literature on the concept of social capital, in this article we rely on the seminal theory by Putnam (2000), who brought the democratic and civic perspective to social capital by introducing “bridging” and “bonding” social capital. In the introduction section, we first go through how minorities develop social capital, and then how social capital is connected to political self-efficacy. By analyzing a nationwide dataset of 15–16-year-old students (N = 5,189) completing their final year of comprehensive education in Finland in a multilevel structural equation model, we find out that belonging to the Swedish-speaking minority is connected to more bonding social capital (a sense of belonging to a community, and social networks), and through that to higher political self-efficacy, but it has no connection to bridging social capital (generalized trust, and civic participation). Moreover, our results show that generalized trust is not linked to political self-efficacy, unlike the other studied dimensions of social capital. In the light of political self-efficacy development, we therefore believe that societies should invest in both the bonding and the bridging social capital of youth and minorities.

2 Social capital, political self-efficacy, and minorities

2.1 Social capital and minorities

Social capital has been an enormously used concept across social sciences, especially in economics, sociology, and political science. For example, it has been used to explain health (Hyyppä and Mäki, 2001; Poortinga, 2012; van der Star and Bränström, 2015; Bamford et al., 2021; McAlpine et al., 2022), wellbeing (Heim et al., 2011; van der Star and Bränström, 2015), success in the labor market (Baalbergen and Jaspers, 2023), a higher level of entrepreneurship (Dana et al., 2018), political trust (Bäck and Kestilä, 2009) and political participation (Verba et al., 1995; Lee, 2022). While the concept of social capital has been defined in multiple ways (see e.g., Bäck, 2011), in political science, social capital is often understood as an interplay of trust and engagement in social networks (see, e.g., Tingaard and Svendsen, 2009).

Perhaps the most seminal theory of social capital in the twenty-first century was created by Robert Putnam, who regarded social capital to consist of relations between people, social networks, and norms of trust and mutuality (Putnam, 1993a,b, 2000). In the center of Putnam's work are the concepts of bridging and bonding social capital, which especially classify different types of social ties between individuals. Briefly, bridging social capital refers to connections between diverse people, whereas bonding social capital refers to connections between similar people (Putnam, 2000; see also Putnam and Campbell, 2010). Sometimes, in addition to bridging and bonding social capital, scholars have distinguished linking social capital. It is said to describe the ties between individuals inside institutional networks and hierarchic communities, such as workplaces (see, e.g., Poortinga, 2012). From a societal perspective, bridging social capital has often been considered more beneficial since it can create unity and cohesion by connecting different groups and people (Putnam, 2000, p. 22–23).

Because social capital is considered to be that important, it can be regarded as somewhat problematic that minorities in general have lower social capital than majorities (Heim et al., 2011; Bamford et al., 2021; McAlpine et al., 2022; Lenkewitz, 2023). This is often explained by their weaker socioeconomic status (SES), possible discrimination and/or their structurally weaker position compared with the majorities (Heim et al., 2011; McAlpine et al., 2022; Lenkewitz, 2023), although these factors do not necessarily explain the total difference (Lenkewitz, 2023). However, minorities can be equal with majorities in terms of their SES and law-based rights, as is the case for the Swedish-speaking minority in Finland (Hansén, 1987). They can also have higher social capital than majorities (Baalbergen and Jaspers, 2023), which also holds true for the Swedish-speaking Finns (e.g., Hyyppä and Mäki, 2001, 2003; Nyqvist et al., 2008; Paljärvi et al., 2009). Still, minorities are inherently in a different societal position than majorities because of being a minority: they are always an exception to the societal norm, which is why they tend to develop a minority identity and find it important to preserve their culture and defend their rights (Branscombe et al., 1999; Frost and Meyer, 2012). Even though national and group-based contexts vary extensively, minorities in democratic societies are in a more vulnerable societal position than majorities because the majority has some power over their rights based on the majority principle (Strubell and Boix-Fuster, 2011). In this study, we analyze a minority that is in a comparatively very good position but still has a clear minority position and experiences minority-specific circumstances and challenges (Liebkind et al., 2007; cf. Mişcoiu, 2006). By investigating such a privileged minority, we hope to be able to shed some light on how a minority can be strengthened or weakened (Liebkind et al., 2007).

Because of their vulnerable societal position, minorities tend to have a strong in-group identity, which predisposes them to have a stronger in-group bias, that is, stronger motivation to prefer interacting with members of their own minority group (Branscombe et al., 1999; Frost and Meyer, 2012; Uekusa, 2020). As a result, minorities tend to develop relatively strong bonding social capital (Scholten and Holzhacker, 2009; Uekusa, 2020), apart from they believe that socializing with the minority leads to negative consequences (Uekusa, 2020). This tendency to develop strong bonding social capital is underscored in minority-specific contexts, such as minority sports clubs, where minority members enjoy the chance to interact with each other (Theeboom et al., 2012). A high level of bonding social capital, in turn, is expected to reinforce exclusive identities and homogeneous groups, promote strong ingroup loyalty, and support specific reciprocity and mobilization of within-group solidarity (Lee, 2022). Although young people are studied considerably less than adults, these tendencies also apply to youth, who seem to prefer to rely on bonding social networks, even though they recognize the importance of bridging networks (Jørgensen, 2017).

Due to the tendency to bond within their own minority community, minorities often have limited networks beyond their homogeneous group compared with non-minority groups (Scholten and Holzhacker, 2009). This is amplified by the majority attitudes toward the minority: the more hostile the environment, the more minority members seem to identify with the minority and the less with the majority (Heim et al., 2011). Compared with homogeneous minority contexts that enhance strong ties between individuals, heterogeneous contexts and intergroup contacts seem to create weaker ties but contribute more to social skills, intercultural learning, volunteer work, and self-confidence (Theeboom et al., 2012). Thus, minorities can be assumed to have relatively high levels of bonding social capital, such as feelings of belonging to a community, but relatively low levels of bridging social capital, such as generalized trust, even though bonding and bridging capital do not need to be inversely correlated (Lee, 2022). Bridging social capital in general or generalized trust in particular can be further weakened by a hostile public opinion toward the minority (van der Star and Bränström, 2015), experiences of harassment (Heim et al., 2011; Lee, 2022) or discrimination (Bäck and Kestilä, 2009), or a threat from the majority (Lee, 2022). Such negative experiences, in turn, are common among people belonging to minorities (Heim et al., 2011; Lindell, 2020).

As for language minorities, language is a strong identity builder (see, e.g., Anderson, 1983/2006, pp. 73–77), but it also creates a barrier. In other words, language is a factor that makes it easier to bond within one's group because it is more pleasant and effortless to communicate in one's native language than a foreign one (Theeboom et al., 2012). It also highlights the importance of minority communities because possibilities to interact in one's mother tongue are more limited for minorities (Jørgensen, 2017). Hence, individuals belonging to language minorities are in different positions: the bilingual can interact with the majority without constraint/trouble, while the unilingual cannot, although they can both be expected to identify with the minority community (see also Lindell, 2020, p. 33).

A concern regarding generalizability arises from the fact that most of the results on minority social capital concern adults (see however Jørgensen, 2017), which highlights the need for research on growing citizens during their formative years. While both adults and adolescents can have access to social capital, the nature and extent of their social capital differ due to their social networks, age, and life experiences (see, e.g., Stolle and Hooghe, 2004). Adults tend to have more extensive and more established social networks compared with adolescents, often due to them having had more time to develop and build their personal and professional networks. On the other hand, compared with adults, adolescents may prioritize their social lives, and their peers may form their most important networks. Whereas, occupation and professional networks play an important role in adults' social capital, school and extracurricular activities have a clear impact on adolescents' social capital (see, e.g., McDonald and Mair, 2010; Schwandel and Stout, 2012). In addition, compared with adults, adolescents are quite dependent on their parents for their social capital, and school is an important arena for them (Jørgensen, 2017; see also Putnam, 1993a,b; Stolle and Hooghe, 2004). Basically, adolescents regard their peers as an integral asset, highlighting the need to study peer social capital (Jørgensen, 2017). While we recognize that adolescents are influenced by many factors, not the least by the actors of political socialization, the effects of social capital among youth are understudied.

Moreover, most previous studies on minorities' social capital are studies on socioeconomically disadvantaged ethnic minorities (Branscombe et al., 1999; Scholten and Holzhacker, 2009; Heim et al., 2011; Uekusa, 2020), however, some studies are on sexual minorities, who are not necessarily economically disadvantaged (Frost and Meyer, 2012; van der Star and Bränström, 2015; Lee, 2022). Thus, there seems to be a research gap in the previous studies on socioeconomically non-disadvantaged ethnic minorities' bonding and bridging social capital. However, the fact that both socioeconomically equal non-ethnic minorities and socioeconomically disadvantaged ethnic minorities seem to function in the same way supports the assumption that these mechanisms are similar for the socioeconomically equal/advantaged Swedish-speaking minority in Finland.

Therefore, minorities can be expected to have relatively high levels of bonding social capital, such as feelings of belonging to a community, but relatively low levels of bridging social capital, such as generalized trust. As for the Swedish-speaking minority in Finland, previous results on adults suggest that the minority has a higher level of social capital than the majority, but results on the exact aspects of stronger social capital are mixed (Hyyppä and Mäki, 2001; Nyqvist et al., 2008; Paljärvi et al., 2009; Myllyniemi, 2014, p. 55). However, we expect that the bonding social capital of the Swedish-speaking Finns is strengthened by the fact that the Finnish school system, from daycare to university and teacher education, is almost completely separated based on language, and the minority community is strong and has plenty of resources with which to develop its culture (see, e.g., Hansén, 1987). On the other hand, the minority is often harassed and feels threatened (Lindell, 2020, p. 21), and the political discussions on weakening the strong position of the minority are constant (Lindell, 2020, p. 20), which can be expected to increase the threat that the minority feels and lower the bridging social capital. It is also worth noting that the Swedish-speaking minority in Finland is equal or more advantaged compared with the majority when it comes to economic capital, human capital, and cultural capital (see, e.g., Saarela, 2006), although there are significant internal differences (Liebkind et al., 2007). On the one hand, this makes it suitable for studying the effect of belonging to an ethnic minority without the interference of a disadvantaged societal status, yet on the other hand, it weakens the generalizability of the results to disadvantaged minorities. However, we find it important to analyze the potential effects of a minority status when the minority is well-integrated. Thus, we expect the following:

H1a: Adolescents belonging to the Swedish-speaking minority in Finland have a higher level of bonding social capital than the majority-belonging adolescents in Finland.

H1b: Adolescents belonging to the Swedish-speaking minority in Finland have a lower level of bridging social capital than the majority-belonging adolescents in Finland.

2.2 Social capital and political self-efficacy

In this article we are interested to test the mediating effect of social capital on political self-efficacy. While analyzing this effect, our focus is on social capital's two dimensions: bridging and bonding social capital. Overall, previous studies have indicated that individuals with higher levels of social capital are more likely to have the resources, connections, and information necessary to engage in political activities, which can increase their political self-efficacy (Browning et al., 2004; Anderson, 2010). Moreover, social capital may indirectly boost political self-efficacy by providing individuals with a sense of empowerment and support, allowing them to feel more capable and confident in their political participation (Levy, 2013, 2018; Quintelier, 2015; Du et al., 2022). Social capital may also expose individuals to diverse political perspectives and ideas, which can broaden their political knowledge and competencies, further increasing their self-efficacy in the political sphere. Therefore, social capital and political self-efficacy can also easily create self-reinforcing cycles, since social capital enhances political self-efficacy and participation through it and participation in turn increases social capital (Putnam, 2000; Levy, 2018). Bandura et al. (2001) also found that people who have confidence in their abilities (self-efficacy) and believe that positive outcomes will happen (outcome expectancy) are more likely to do something. Self-efficacy and outcome expectancy play a big role in how people think and act, so they can help explain why using information and having social connections affect political involvement (see also Kim et al., 2020). In other words, their study suggests that using information and having diverse social connections can influence how politically involved people are, because they affect how confident people feel and what they expect to happen (Bandura et al., 2001).

Even though social capital's role in political engagement has interested scholars for a long period of time, some scholars have questioned whether social capital as a whole, or rather some of its components lead to specific outcomes. For instance, Bäck and Kestilä (2009) found that generalized trust is associated with political trust, while they found no connection between civic participation and political trust. Hyyppä and Mäki (2001), in turn, discovered that religious activity, close friends, and generalized trust were better measures with which to explain self-rated health rather than, for example, hobby participation. When considering how both bridging and bonding social capital are connected to an individual's level of political self-efficacy, we must start with not only observing how bridging and bonding social capital are found to affect political self-efficacy, but also, how previous studies have operationalized the concepts.

Empirically, bridging social capital is typically measured with generalized trust (social trust) and the number of social networks. Oftentimes in social science studies, generalized trust has in fact been used as a proxy for social capital as a whole (Uslaner, 2006; Liu et al., 2018). Bonding social capital, on the other hand, is often measured with the number of close friends, the sense of community, and the similarity of values within the networks (Putnam, 2000; Coffé and Geys, 2007). A sense of community, or community connectedness, means the feeling of belonging to a group or groups, in which members matter to one another and to the group, and there is a shared faith that the members' needs will be met through their commitment together (Anderson, 2010).

According to McMillan and Chavis (1986), a sense of community is composed of four elements: (1) membership, (2) influence, (3) the integration and fulfillment of needs, and (4) a shared emotional connection. Thus, a sense of community requires not only formal membership or activity, which are often used as a measure in social capital research, but also emotional aspects: feeling like an important part of the community, feeling that the community fills some needs and shares some values of the individual, and identifying with the community through, for example, sharing a common history (Anderson, 2010). On the other hand, generalized trust is not required, only some trust toward the in-group. The sense of community is expected to enhance political self-efficacy by offering individuals the chance to feel meaningful and influential, providing a context with shared values, and helping to gain relevant skills and experiences (Anderson, 2010).

Even though the components of social capital have been measured separately, they are still usually expected to function in the same way, that is, to increase political self-efficacy (Du et al., 2022). However, during recent years, scholars have still increasingly approached the concept of social capital more critically by aiming to create a converging theory, which could comprehensively bring together all of its dimensions (e.g., Claridge, 2018; Gelderblom, 2018). Although the relationship between trust and engagement is far from clear (Lillbacka, 2006; van Ingen and Bekkers, 2015), these dimensions are usually found important when measuring social capital. After all, it should not matter whether engagement enhances trust or whether the more trusting get more engaged (van Ingen and Bekkers, 2015), as the result is anyway that the two dimensions support each other.

Bridging social capital is often considered societally more beneficial for political self-efficacy, because it connects people (see, e.g., Putnam, 2000; Wilhelmsson, 2015). This has been used to motivate assimilationist integration policies, because they contribute to bridging social capital instead of bonding, although they can otherwise be criticized. Empirically, there seems to be fewer results. A study from England on social capital's influence on building community resilience and health (n = 17,572) has shown that the indicators of the different types of social capital are only weakly interrelated. Furthermore, the results suggest that the various types capture different aspects of the social environment and that, in particular, bonding and bridging social cohesion, civic participation, heterogeneous socio-economic relationships, and political efficacy and trust appeared important for community health after controlling for neighborhood deprivation (Poortinga, 2012). In line with the expectations, most indicators of bonding, bridging, and linking social capital have been significantly associated with neighborhood deprivation and self-reported health (Poortinga, 2012). There are also studies that indicate that ethnically diverse sports clubs are more likely to boost adolescents' self-confidence compared with sports clubs with uniform ethnicity among their members (Theeboom et al., 2012). On the other hand, it has been found that participating in minority-specific civil society associations, in addition to participating in more general organizations, contributes to political participation (Wilhelmsson, 2015; Lee, 2022).

After all, since bonding and bridging social capital do not need to be inversely correlated (Lee, 2022), we can assume that both can contribute to political self-efficacy. On the one hand, bridging social capital can connect adolescents to the wider societal context, giving them confidence in their possibilities to understand it and operate as parts of it. On the other hand, bonding social capital can offer them opportunities to feel heard, connected, trusted, understood, and supported, enhancing their self-efficacy also in the political domain. Thus, we expect the following:

H2a: The Swedish-speaking minority youth is likely to exhibit a higher level of bonding social capital, which, in turn, is anticipated to boost their political self-efficacy.

H2b: The Swedish-speaking minority youth is likely to exhibit smaller exposure to diverse political perspectives, leading to a lower level of bridging social capital, which is anticipated to hold back their political self-efficacy.

To conclude the theoretical discussion and literature review, our analysis will be guided by the following four hypotheses. The first two hypotheses focus on the relationship between minority status and social capital, suggesting that the Swedish-speaking Finnish adolescents have higher bonding social capital (H1a) and lower bridging social capital (H1b) than the majority-belonging adolescents. In practice, we operationalize bonding social capital as social networks at school and among peers, and as feelings of community and togetherness. Bridging social capital, in turn, is operationalized as generalized trust, that is, as a trust in people in general, and as participation in different civil society organizations, which are typically used measures for (bridging) social capital (see, e.g., Putnam, 2000; Newton, 2001; Uslaner, 2002; Nyqvist et al., 2008; Paljärvi et al., 2009). The second pair of hypotheses more generally regards the mediating effect of social capital on political self-efficacy via minority status. Based on theory, we expect both bonding and bridging social capital to be connected to political self-efficacy. Thus, we hypothesize that bonding social capital contributes to political self-efficacy among minority adolescents (H2a) but that bridging social capital holds it back (H2b). The hypotheses are illustrated in Figure 1.

Figure 1

Figure 1. Individual level dimensions of social capital predicting the level of political self-efficacy among majority and minority adolescents.

3 The study context: social capital among Finnish-speaking majority and Swedish-speaking minority youth in Finland

The Swedish-speaking Finns are a small language minority accounting for about 5.2% of the Finnish population. Yet, as a national minority that once formed the elite of society (Väistö, 2017, p. 49–50), they have a strong law-guaranteed status: Finnish and Swedish are stated as the two equal national languages in the Finnish constitution. Other laws strengthen this by giving the Swedish-speaking Finns, for example, equal rights to go to school in Swedish, get official services in Swedish, and have national broadcasting services in Swedish. However, over 50% of the Swedish-speaking feel that they do not get services in Swedish, almost a third has been harassed for using Swedish, and over 20% feel that they are treated as second-order citizens because of their native language (Lindell, 2020, p. 21). Correspondingly, only a third of the Swedish-speaking Finns trust unknown Finns (Lindell, 2020, p. 38) or feel togetherness with them, while over 80% feel togetherness with other Swedish-speaking Finns (Lindell, 2020, p. 32). From a comparative point of view, the Swedish-speaking Finns have been considered one of the most privileged minorities of the world–yet it can be argued that they are not equal with the majority in practice (Liebkind et al., 2007). Thus, we believe that we can say something about minorities in general by studying this specific minority, and eventually tell something about how and why a minority can be strengthened or weakened (cf. Liebkind et al., 2007).

Furthermore, the inward-looking nature of the Swedish-speaking community can be expected to be strengthened by the fact that the Swedish-speaking Finns are a language minority, for whom their own language is central (Lindell, 2020, p. 100). Even though most people in Finland are or have been required to study the other native language (Väistö, 2017, p. 13), language is often a barrier for the Finnish-speaking to participate in Swedish-speaking activities (Lindell, 2020, p. 24). On the other hand, the minority often prefers Swedish-speaking activities, although about 50% of the Swedish-speaking are bilingual (Lindell, 2020, p. 24).

The Swedish-speaking are often claimed to have a higher level of social capital than the Finnish-speaking (e.g., Hyyppä and Mäki, 2001, 2003; Nyqvist et al., 2008; Paljärvi et al., 2009). That higher level of social capital, in turn, is suggested to explain, for instance, the Swedish-speaking Finns' better health (Hyyppä and Mäki, 2001, 2003; Suvisaari et al., 2014), higher life expectancy (Hyyppä and Mäki, 2001), and their higher civic participation (Nyqvist et al., 2008) when compared with the majority. However, these suggestions are also sometimes contested (see, e.g., Hyyppä and Mäki, 2003; Nyqvist et al., 2008). While it is often not specified which aspects of social capital are higher among the Swedish-speaking, some results point to the Swedish-speaking having a higher level of civic participation (Nyqvist et al., 2008; Paljärvi et al., 2009; Myllyniemi, 2014, p. 55), tighter social networks (Hyyppä and Mäki, 2003; Nyqvist et al., 2008; Paljärvi et al., 2009), more active hobby or cultural participation (Hyyppä and Mäki, 2001, 2003; Paljärvi et al., 2009), more active religious participation (Paljärvi et al., 2009), and a higher level of generalized trust (Hyyppä and Mäki, 2001; Nyqvist et al., 2008) than the Finnish-speaking. On the other hand, some studies find no differences in associational participation (Hyyppä and Mäki, 2001), religious involvement (Hyyppä and Mäki, 2001, 2003), the number of close friends (Hyyppä and Mäki, 2001), or generalized trust (Paljärvi et al., 2009). Yet others find that the Finnish-speaking have a higher level of associational involvement (Hyyppä and Mäki, 2003). Swedish-speaking Finnish youth's social capital has rarely been studied, and when they have been included, the sample size for the minority has been low (Myllyniemi, 2014), which highlights the importance of this study.

4 Materials and methods

4.1 Sample and procedure

The data of the study was collected in the spring of 2021. Thus, the research strategy was adapted to the COVID-19 situation. The data was collected in classrooms across Finland, without any of the research team being physically present. Each research situation began with a professionally produced video that introduced the study and explained how to participate in the study. The video was designed to inspire pupils to respond, replace the presence of a researcher, and standardize the research situation. After the video, the students responded individually using electronic devices, while the teacher was present in the class to monitor the situation and provide practical assistance if needed. The teacher could call the research team at any time, in case of any issues. Most of the students answered during a lesson in class (94.2% of the weighted sample, 90.8% of the unweighted sample), and some during remote learning arrangements (5.8% of the weighted sample, 9.2% of the unweighted sample). In such remote learning arrangements, the whole class answered at the same time, while the teacher supervised the situation and advised the pupils, if needed, to secure equality between the research situations. There were no significant differences between data collected in class and data collected in remote learning arrangements. The sample consists of 5,189 15–16-year-old youth in total, 4,133 studying in Finnish-speaking schools, and 1,087 studying in Swedish-speaking schools.¹

The sample is built as a strategic cluster sample. When recruiting the sample, geographic diversity and representation, population density, language diversity, the average education level of the municipality, the schools' size and the school's type were considered. We contacted 109 schools in total, but some of them declined to participate, predominantly due to the extra burden related to the COVID-19 situation in their municipality, or participation in other research activities. When a school declined, we recruited another school from a similar neighborhood, if possible. In most schools, all classes participated in the study, as requested. Even though the sampling of the municipalities and schools was not random, the data can be regarded as a good mini-representation of that cohort in Finland because collecting the data during lessons enabled us to reach out to all pupils regardless of their background. Yet, it was voluntary to participate, and thus, we cannot exclude potential bias resulting from, for example, pupils who are less interested in politics not responding. Based on the teachers' notes, the individual-level response rate was about 80%.

The data was collected in 379 classes in 80 schools, 17 of which had Swedish as the language of teaching, in 38 municipalities or local strategic partnerships. Because of some schools' reluctance to participate, and unequally divided effects of COVID-19, there were some geographic discrepancies in the sample (see Appendix A). Due to the strategic overrepresentation of the Swedish-speaking and the geographical discrepancies, geographical and linguistic weights are applied (see Appendix A for further details).

4.2 Measures

All items used for each measure are specified in Appendix D.

4.2.1 The dependent variable

4.2.1.1 Political self-efficacy

Political self-efficacy is measured using a six-item scale, based on the ICCS 2009 scale, with McDonald's omega of 0.85 (Cronbach's alpha 0.84). The classical measure of political self-efficacy has remained relatively similar since the inception of the concept [e.g., European Social Surveys (ESS) and the International Social Survey Programme (ISSP)]. The respondents responded to statements using a five-point Likert scale, ranging from totally disagree to totally agree and coded so that higher values imply a higher level of political efficacy. These statements are: 1. Sometimes politics seems so complicated that I do not quite understand what is going on (reverse-coded); 2. I trust in my own abilities to participate in politics; 3. I understand the most important political issues in Finland; 4. I know more about politics than most of my age; 5. When political issues or problems are discussed, I usually have something to say; 6. I have political opinions that are worth hearing. We have run a confirmatory factor analysis (see Appendix C) to control the fit of the scale to the data at hand.

4.2.2 The independent variables

4.2.2.1 Social capital

Social capital is measured using four scales, all of which are recoded so that 0 means the lowest level of social capital and 1 means the highest level of social capital. The scales are theory-based and formed using exploratory factor analysis (EFA), after which they have been controlled using confirmatory factor analysis, which yields satisfactory fit statistics (the analyses are presented in detail in Appendix B).² After having found a satisfactory factor solution, the scales to be used in the analyses were calculated as sum variables.³ These four scales are Generalized trust, Civic participation, Social networks, and Community. The two first are used to measure bridging social capital, while the two last are used as proxies for bonding social capital. Generalized trust is measured with the standard question (“Would you say that most people can be trusted or that you can't be too careful in dealing with people?”). Previous studies focusing on measuring social capital have considered this question as the most suitable standard question for generalized trust and hence bonding social capital as well (e.g., Putnam, 2000; Newton, 2001; Uslaner, 2002). Civic participation consists of five items measuring whether one has actively or less actively (response alternatives: never; yes, but not very actively; yes, and I have participated intensively) taken part in different civil society organizations with a societal or political goal (political youth organizations, environmental organizations and groupings, associations for animal rights, human rights associations, or other voluntary associations), and it has a McDonald's omega of 0.77 (Cronbach's alpha 0.71). Social networks scale consists of four items that measure whether one feels lonely (constantly/quite often/sometimes/rarely/never), whether one feels like an outsider in one's school (range: totally agree…totally disagree), whether one is afraid of being bullied (range: totally agree…totally disagree), and how many close friends one has (none/one/two/several). The scale has a McDonald's omega of 0.74 (Cronbach's alpha 0.74). Community consists of four items measuring whether one feels belonging to a community that is important to oneself, important for other people, trusted, and able to get help when one needs it (each ranging from agree…totally disagree). It has a McDonald's omega of 0.79 (Cronbach's alpha 0.78).

In this study, we thus use four measures to encompass different aspects of social capital: generalized trust and civic participation for bridging social capital, and social networks and community for bonding social capital. Although different, these aspects are also connected. The McDonald's omega for all the social capital variables–one variable for generalized trust, five variables for civic participation, four variables for social networks, and four variables for community–is 0.71. However, if the aspects are used as scales, the McDonald's omega drops to 0.56, which is significantly lower than the McDonald's omegas for each scale separately. The correlations between individual items between the different aspects are usually around 0.2 or lower, while correlations between the items within an aspect are mainly around 0.4 or higher, with civic participation measures being the least correlated with other measures.

4.2.2.2 Minority status

Minority status is operationalized as the school language of the respondent (i.e., Swedish = 1; Finnish = 0) to measure the socialization environment of the respondent. When the weights are applied, 99.6% of the respondents who speak Finnish but not Swedish at home study in Finnish-speaking schools, while 84.3% of the Swedish-speaking who do not speak Finnish study in Swedish-speaking schools. The weighted sample's bilingual adolescents who speak both Finnish and Swedish at home study slightly more often in Finnish-speaking schools (52.6%) than in Swedish-speaking schools (47.4%). The share of immigrant-origin pupils is almost the same in both groups, both measured as other home languages than Finnish or Swedish and measured as the country of birth of the respondents and their parents.

4.2.3 The control variables

4.2.3.1 Family SES

Socioeconomic status (SES) refers to education, profession, and economic situation, which all are highly correlated. The educational level of the respondents themselves is the same–each respondent is in Year 9 of comprehensive education–so that cannot be used as a test variable. For youth, family background is also assumed to impact the most, which is why we measure Family SES. The questionnaire includes questions on, both parents' education, employment and profession, and the family's economic situation. The first two concepts unreliable measures since the young were clearly unsure about the educational level of their parents and because many either did not know what their parents did for living or answered vaguely. Previous studies have also shown that children, particularly children from low-SES families, do not know the educational level of their parents and tend to overestimate it (Engzell and Jonsson, 2015). Therefore, the family's economic status, measured as youth's self-assessment of the economic situation of their family, is used as a proxy for the Family SES of the respondents. In our dataset the family's economic status has been measured with a question: “How is the economic situation of your family?” with a five point response scale: 1 = very bad, 2 = quite bad, 3 = neither good nor bad, 4 = quite good, 5 = very good. The variable's scale has not been recoded in the analyses. The variable has been coded in compliance with its original scale in the analysis.

4.2.3.2 Family discussion

Family or parents is considered an important political socialization agent for youth. For instance, family can give examples of participatory norms and politically efficacious attitudes (Almond and Verba, 1963; Levy, 2013) and it can involve children in political discussions (Sherrod et al., 2010). Thus, political socialization from parents is controlled for with a question about the frequency of political discussion with parents. Family discussion has been measured with a question: “How often do you do the following things outside school hours? – Discuss political or societal matters with your parents. There are four different response categories: daily or almost daily, weekly (at least once a week), monthly (at least once a month) and more rarely or never. The variable has been coded in compliance with its original scale in the analyze.

4.2.3.3 Living environment

Living environment controls whether the respondent lives in a center of a large town (over 100,000 inhabitants), in the center of a small town (< 100,000 inhabitants), in a suburb, in a municipal center or other population center in a rural area, or in a sparsely populated rural area.

4.2.3.4 Living in Finland

Living in Finland controls for the time the respondent has lived in Finland, thus accounting for differences that derive from different possibilities to have gained knowledge about the Finnish society and social capital in Finland. It is asked with a multiple-choice question with five response alternatives. Living in Finland has been measured with a question: How long have you lived in Finland? There are five different response categories: 1 = My whole life; 2 = Over 10 years but not always; 3 = 5–10 years; 4 = 1–4 years; 5 = Less than a year. Living environment -variable controls for the time a respondent has lived in Finland. There has been given five response categories: 1 = My whole life; 2 = Over 10 years but not always; 3 = 5–10 years; 4 = 1–4 years; 5 = Less than a year.

4.2.3.5 Living in the current municipality

Living in the current municipality controls for the time the respondent has lived in the municipality in which they currently live. In literature, internal migration has been reversely connected to economic and social capital, and the Swedish-speaking population has been found to stay in their birth region significantly more often than Finnish-speakers (Saarela, 2006). Living in the current municipality is asked with a multiple-choice question with five response alternatives. Living in the current municipality -variable controls for the time respondent has lived in their current home municipality. It has been measured with question: “How long have you lived in the municipality where you currently live?” and there are five response categories: 1 = My whole life; 2 = Over 10 years but not always; 3 = 5–10 years; 4 = 1–4 years; 5 = Less than a year.

4.3 Analytical strategy

Our data have a hierarchical structure with individuals (level 1) nested within classes (level 2) nested within schools (level 3) nested within municipalities (level 4). Therefore, we estimate multilevel regression models within a structural equation modeling framework in Stata 17.0 to minimize the risk of underestimating the standard errors of regression coefficients caused by individuals within a class, a school, or a municipality to be more like one another than individuals from different classes, schools, or municipalities (Barr et al., 2013; Tarka, 2018).

First, we test how the variance is divided between the levels by running empty models (see Appendix D). Even though there is slightly more variance on class level than on school-level when using weighted data, we find school-level more important to control in the main analysis, because our minority status variable is measured on the school level. A valid concern is whether our multilevel models may suffer from intercepts not being able to vary by class or municipality (see also Barr et al., 2013). However, to fit such a model is computationally intensive and leads to an unreasonable computation time. Hence, we are unable to run three-level models, but we run two-level models where intercepts can vary by class as a robustness check. This approach is also theoretically motivated, since previous literature has found Finnish schools to be relatively homogeneous though some sociodemographic differentiation between and within lower secondary schools in urban areas has been discovered (Kosunen et al., 2016).

5 Results

5.1 Descriptive statistics

In Table 1, descriptives on the social capital and political self-efficacy scales are presented for the whole group and for the Finnish-speaking and the Swedish-speaking separately. According to these analyses, both political self-efficacy, social networks, and community are statistically significantly higher among the Swedish-speaking than among the Finnish-speaking, while there is no difference between the groups in terms of generalized trust and civic participation. Furthermore, social networks and a sense of community are quite strong among all, and also generalized trust is slightly above the arithmetic mean of the scale, while political self-efficacy falls under the arithmetic mean of the scale, and civic participation is very rare.

Table 1

Table 1. Political self-efficacy and social capital by minority status (mean, standard deviation, skewness, kurtosis, independent samples t-test).

Table 2, in turn, presents the descriptive statistics of the background variables for the Finnish-speaking and the Swedish-speaking. Swedish-speaking adolescents appear to discuss politics slightly more often than Finnish-speaking adolescents, and they more often find their family's economic situation to be very good but less often to be quite good. The Finnish-speaking have slightly more often lived in Finland their whole lives, but they have moved from one municipality to another slightly more often than the Swedish-speaking. The Swedish-speaking live in sparsely populated rural areas over twice as often as the Finnish-speaking, and the Finnish-speaking live in suburbs over twice as often as the Swedish-speaking.

Table 2

Table 2. Categorical variables by minority status, crosstabulation (weighted sample).

5.2 The structural equation: analyzing the effect of minority status on political self-efficacy via social capital

The results (see Figure 2) indicate that Swedish-speaking minority adolescents have a higher level of bonding social capital–both social networks and sense of community–than Finnish-speaking majority adolescents, even after controlling for socioeconomic and sociodemographic background. These results suggest that the Swedish-speaking Finn community is indeed relatively rich in bonding social capital, at least among adolescents. The strength of these associations is quite low, only 0.02, but it is statistically significant. Thus, H1a, which hypothesizes that the minority has a higher level of bonding social capital, gets some support. On the other hand, the Swedish-speaking minority adolescents do not have a lower bridging social capital than the majority adolescents. In fact, there is no statistically significant connection between minority status and either of the measures for bridging social capital (generalized trust and associational engagement). Hence, H1b is not supported.

Figure 2

Figure 2. The impact of language group, bridging social capital, and bonding social capital on political self-efficacy. Multilevel generalized structural equation model (standard errors in parentheses). *Significant on 0.05-level, **significant on 0.01-level, and ***significant on 0.001-level.

The results also show that both community, social networks, and civic participation are positively associated with political self-efficacy. The two measures for bonding social capital have estimates under 0.2 (0.17 for social networks and 0.16 for community). Therefore, H2a gets some support: bonding social capital seems to contribute to political self-efficacy. In practice, a young person with a very high sense of community or very good social networks is hence expected to have a 0.16 or 0.17 units higher level of political self-efficacy on a scale from 1 to 5 than a young person with a very low sense of community or very weak social networks. On the other hand, active civic participation clearly reaches the highest estimate (0.50), and passive civic participation, too, gets a relatively high estimate (0.22), which implies that a very engaged young person is expected to have a 0.50 or 0.22 units higher level of political self-efficacy than a person who is not at all engaged, everything else held constant. However, generalized trust, our other measure for bridging social capital, is not connected to political self-efficacy, contrary to H2b. Thus, H2b gets partial support.

Consequently, belonging to the Swedish-speaking minority is positively connected to bonding social capital, which, in turn, is positively connected to political self-efficacy. Therefore, through bonding social capital, belonging to a minority can partly explain the higher political self-efficacy of the Swedish-speaking adolescents when compared with the Finnish-speaking majority. However, it is worth noting that the minority has higher political self-efficacy than the majority even after controlling for social capital and all the control variables which suggests that there are also other explanations for the higher political self-efficacy among the minority adolescents.

As robustness checks, the model was rerun including different measures for civic participation. In the first of these checks, the continuous measures of active and passive participation in civic organizations were replaced by dummies since most adolescents had not participated in any civic organizations or, if they had, they had only participated in one type of civic organizations. This model changed no other results to a significant degree, but it made passive participation insignificant. This suggests that exploratory, less active participation may be connected to political self-efficacy when the young person does it a lot, while occasional passive participation in something may be insignificant. In another model, active participation in religious congregations or other religious groups and active participation in sports clubs or other hobby groups were included as dummies. A third model even included passive participation in sports clubs or other hobby groups were included as dummies. These two models showed that neither religious nor sport activities were connected to political self-efficacy among the sample of Finnish adolescents. The Swedish-speaking adolescents had more often actively participated in sports activities than the majority, and correspondingly, had slightly less often passively participated in sports activities than the majority, which contradicts H1b. When it comes to religious participation, no statistically significant differences were found. Robustness checks were also run by replacing the school level in the analysis by the class level. These analyses did not change any of the main findings.

As to the control variables (see Table 3), political discussion with family is a robust predictor of political self-efficacy, and socioeconomic status has also a weaker connection, but other control variables were statistically non-significant. However, family SES is significantly associated with bonding social capital (social networks and community), which suggests that it may have a mediating effect to political self-efficacy through bonding social capital. Otherwise, those who have recently immigrated to Finland seem to be more engaged in active civic participation, but immigrants who have spent over 5 years in Finland have a lower sense of community than Finland-born. These results contradict the estimates of the Swedish-speaking minority, but the sample sizes and heterogeneity of these immigrant groups prevent further examination. Internal migration, in turn, has no connection to social capital. Those living in suburbs and sparsely populated rural areas have engaged in active civic participation slightly less, and those living in suburbs and municipal centers have engaged in passive civic participation slightly less than those living in the centers of large cities, perhaps reflecting longer geographic distances to possibilities to engage. The variances and covariances of the main model are presented in Table 4.

Table 3

Table 3. Structural equation modeling, control variables.

Table 4

Table 4. Variances and covariances of the main analysis.

6 Discussion and conclusions

In just multicultural democracies, all societal groups, including young people and minorities, are expected, and required to possess enough political self-efficacy to be able to participate in political decision-making (Craig and Maggiotto, 1982; Finkel, 1985). By analyzing a nationally representative dataset from Finland (N = 5,189) with an oversample of Swedish-speaking Finnish minority adolescents (n = 1,089), we examine how different aspects of social capital may help to enhance political self-efficacy among adolescents in general and among a socioeconomically non-disadvantaged minority in particular. By better understanding political self-efficacy development among these societally vulnerable groups, we may be able to contribute to more societal equality. While previous studies have examined social capital and sometimes its different aspects among minority-belonging adults (e.g., Branscombe et al., 1999; Scholten and Holzhacker, 2009; Frost and Meyer, 2012; Uekusa, 2020), minority-belonging youth (see however Jørgensen, 2017) and socioeconomically and societally well-off minorities have rarely been studied. Furthermore, this study takes a step further by analyzing the possible effects of minority status on political engagement through social capital, which has rarely been done (see however Wilhelmsson, 2015; Lee, 2022). Our study indicates that minority adolescents' political self-efficacy may be boosted through their higher bonding social capital. Our results show that both civic participation, community, and social networks are positively connected to the level of political self-efficacy, while generalized trust has no connection. Therefore, we think it would be beneficial to consider and measure the different aspects of social capital separately in future studies the different aspects of social capital.

We find our results interesting especially regarding the Swedish-speaking minority adolescents, whose political self-efficacy seems to be boosted by their higher bonding social capital, which is probably partly generated by their strong minority community and identity (see Branscombe et al., 1999; Frost and Meyer, 2012; Lindell, 2020, p. 32; Uekusa, 2020). However, their bridging social capital is not significantly lower than that of the majority adolescents, probably at least to a part because the minority is in a good societal position. Considering previous research on disadvantaged minorities who may even suffer from more deliberation (Gherghina et al., 2021), we assume that a good societal position and a high degree of integration protects a minority from having low bridging social capital compared with the majority and thus prevents the political marginalization of the minority. Therefore, from the point of view of equal participation, creating societal conditions that are as equal as possible seems important. Moreover, the political self-efficacy among minority adolescents seems to be stronger than among majority adolescents despite controlling for social capital. This suggests that there are likely to be other underlying mechanisms that explain the differences in political self-efficacy levels among Finnish-speaking and Swedish-speaking adolescents, such as differences in their socialization.

Our results support several notions. Interestingly, our results indicate that generalized trust is not connected to political self-efficacy, but instead all measures for bonding social capital (social networks and sense of community) are. This finding is somewhat contradictory to Putnam's (1993a,b, 2000) findings. However, we think that there might be some reasons that could explain this somewhat surprising finding. First of all, this supports the notion that context always matters. The effect of generalized trust on political self-efficacy may vary based on the social context and political culture in which the adolescents live. In practice, this means that in some contexts and cultures trust in strangers may have a more significant role in enhancing political self-efficacy than in other contexts. For instance, in countries with strong democratic traditions, well-functioning democratic institutions, and high levels of political trust, adolescents may feel that their trust in strangers is less relevant to their political self-efficacy. Finland is a good example of this kind of country context. Second, age and developmental factors in terms of emotional and cognitive skills might have an impact. There is likely to be a lot of variation in both the emotional and cognitive skills of adolescents at the age of 15. This can directly affect, for instance, their critical thinking skills, identity formation, social awareness and their peers' influence on them, which can all affect their level of generalized trust and, furthermore, the extent to which it has an influence on their political self-efficacy. In summary, the adolescents' trust formation might be an ongoing process at the age of 15 when they are becoming more independent thinkers.

In addition, we also think that this highlights the importance of considering social capital as a result of its dimensions and not as a unidimensional concept. This result is not unique in suggesting that different aspects of social capital may be related to different outcomes. For example, Bäck and Kestilä (2009) find that generalized trust is related to political trust while civic participation is not, and Hyyppä and Mäki (2001) discovered that religious activity, close friends, and generalized trust were better measures for explaining self-rated health than for example hobby participation. Given that this study has solely focused on adolescents, some of its results might also be explained by the differences in social capital formation between adolescents and adults. For instance, social networks can play a particularly important role for youth, when compared with adults. On the one hand, other students at school are important in making adolescents enjoy being at school, which helps them to learn (Jørgensen, 2017). On the other hand, friends can help adolescents to learn, discuss society with them, and so on (see Jørgensen, 2017).

While we do not find Swedish-speaking minority adolescents to be held back by lower bridging social capital because their level of bridging social capital is not statistically different from the majority, our results support previous results on bonding social capital among minorities (Scholten and Holzhacker, 2009; Uekusa, 2020). Swedish-speaking Finnish adolescents, who predominantly study in their minority language schools, seem to have stronger communities and social networks. This is in line with previous results on a strong minority identity among Swedish-speaking Finnish adults (Lindell, 2020, p. 32), and their higher level of social capital particularly of the bonding sort (e.g., Hyyppä and Mäki, 2001, 2003; Nyqvist et al., 2008; Paljärvi et al., 2009). Despite our results, we cannot rule out the possibility that, in other contexts, higher thresholds for the minority to engage in active civic participation may hold the minority back, given that civic participation is a rather strong predictor of political self-efficacy among all the measured social capital variables.

Despite the uniqueness and applicability of both the data and research models, we are aware of some limitations this study holds. Because the data was not specifically designed for examining different dimensions of social capital, we could not exclusively divide bonding and bridging social capital. For example, civic participation was used as a measure for bridging social capital, suggesting that it creates weak ties between different people. Yet, civic participation can also create bonding social capital, since organizations also can bring like-minded individuals together and help people form close relationships (see Poortinga, 2012). Thus, further research is needed to deeply understand the mechanisms between minority status, social capital, and political self-efficacy. The mechanisms should also be tested for more disadvantaged minorities as well and for different aspects of political engagement, to be able to develop societies to become more inclusive and democratic.

There are also limitations regarding the causal relations, which we have not been able to examine with our data. For instance, it is possible that political self-efficacy encourages adolescents to get involved in civic participation, rather than civic participation strengthening political self-efficacy. Given that neither sports or hobby group activities nor religious group activities were connected to political self-efficacy, we are reluctant to discard this possibility. Yet it seems unlikely that political self-efficacy contributes to social networks (mainly school-related social networks in our data) or feelings of community, because school, friend groups, and communities inside and outside school are mainly accessed regardless of the political engagement of the adolescents. Hence, these connections should be tested with longitudinal data.

Certain limitations in the context of minorities need to be considered as well. First, it is important to note that minorities, by their nature, constitute a distinct and often marginalized group within society. Despite huge variance in national and group-based contexts, there are power relations between minorities and majorities. In democratic societies, minorities are in a more vulnerable societal position than majorities because the majority has some power over their rights based on the majority principle (Strubell and Boix-Fuster, 2011). This distinctiveness necessitates a deeper understanding and increased knowledge about various minority groups, in different societal positions. Additionally, it is worth highlighting that examining the social capital, or the resources and networks available to individuals or groups in society, can be done without solely focusing on socioeconomic status (SES). This allows for a comprehensive analysis of the social dynamics prevalent within minorities and between minorities and majorities. In this study, a socioeconomically and societally strong minority was in focus, which weakens the generalizability of the results to more disadvantaged minorities but makes it a good case with which to examine social capital without involvement from SES.

Considering these limitations, we suggest that states should avoid emphasizing assimilationist approaches to minority integration policies. Such policies can inadvertently undermine the political self-efficacy of minority groups, since they weaken the ties within minorities. According to our results, it is crucial to recognize that both bridging and bonding social capital are vital for minority communities' political engagement. In particular, bonding social capital, which encompasses connections within a specific community, plays a significant role in fostering higher levels of political self-efficacy (PSE) among both the Swedish-speaking minority of this study and other minorities in earlier studies (see, e.g., Wilhelmsson, 2015; Lee, 2022). At the same time, it is essential to avoid creating or maintaining disadvantages for minorities in terms of bridging social capital, that is, connections across different communities, since these connections connect them to the wider society and its decision-making processes. In addition, it is vital to acknowledge the unique challenges faced by minorities, such as language barriers, socioeconomic disadvantages, discrimination, and feelings of exclusion, and the need for a more nuanced understanding of their experiences. By considering the limitations mentioned and rejecting assimilationist policies, states can better support minority communities and foster their political engagement in general and political self-efficacy in particular.

Regarding youth politics, we propose the allocation of resources toward initiatives aimed at enhancing the social capital of adolescents during their crucial formative years. In our results, we see that social capital is a building block of political self-efficacy, but at the same time, social capital is largely correlated with other resources, such as the socioeconomic status of the family. One potential approach is fostering positive relationships among students within schools, thereby promoting a sense of camaraderie and preventing incidents of bullying. It is also crucial to ensure that community-based hobbies and recreational activities are accessible to all young individuals, irrespective of their family's socioeconomic background and their ethnic group. By implementing such measures, we can contribute to the overall wellbeing, development and political engagement of our youth.

All in all, our findings show that the Swedish-speaking minority exerts a stronger presence on political self-efficacy compared with other groups, even after controlling for social capital. This suggests the existence of other underlying mechanisms that contribute to this phenomenon. An important discovery from our study is that bonding social capital is associated with political self-efficacy among young individuals, contrary to concerns raised by scholars like Putnam. However, this outcome can be logically explained by the fact that close-knit networks not only provide the necessary support for adolescents to feel influential but also serve as potential platforms for motivation and participation. In other contexts, higher thresholds for the minority to engage in active civic participation may yet hold the minority back, given that civic participation is the strongest predictor of political self-efficacy among all the measured social capital variables.

Data availability statement

The datasets presented in this article are not readily available because the dataset is currently under an embargo, but will be published in the Finnish Social Science Data Archive at a later stage. Requests to access the datasets should be directed to Elina Kestilä-Kekkonen, ZWxpbmEua2VzdGlsYS1rZWtrb25lbiYjeDAwMDQwO3R1bmkuZmk=.

Ethics statement

Ethical approval was not required for the studies involving humans because an ethical approval was not required according to the instructions of the guidelines of the Finnish National Board on Research Integrity TENK. The participants of the study were 15 or 16 years old, and the study had over 5,000 participants, which makes them unidentifiable in the final results. The recommendations of the Finnish National Board on Research Integrity TENK were meticolously followed thoughout the research. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants' legal guardians/next of kin in accordance with the national legislation and institutional requirements because in Finland, 15-year-olds are considered old enough to give their own consent to participate in research. According to the national guidelines, their parents or guardians were informed about the study, and given a chance to forbid their child to participate. Written informed consent was not obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article because the manuscript does not contain any potentially identifiable data on individuals.

Author contributions

VH: Conceptualization, Data curation, Formal analysis, Methodology, Visualization, Writing—original draft, Writing—review & editing. AT: Conceptualization, Writing—original draft, Writing—review & editing.

Funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. The authors have received funding from the research project Inkludering av svenskspråkiga elever i studie om medborgarfostran i skolan (Inclusion of Swedish-speaking pupils in a study of citizenship education at school), funded by The Swedish Cultural Foundation in Finland (grant number 156 786), and from the research project Koulutuksen merkitys kansalaispätevyyden kehittymisessä (Education, Political Efficacy, and Informed Citizenship—EPIC), funded by the Academy of Finland (grant number 323608).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpos.2023.1270065/full#supplementary-material

Footnotes

1. ^The sample consists of students during the last year of their common education, i.e., most of them were born in 2005. Some had started their schooling a year earlier or a year later or doubled a year, and were thus born in 2004 or 2006 (239 students in total).

2. ^Apart from chi² tests, yet as they are dependent on the sample size, and other fit statistics indicate acceptable or close fit, we deem the model acceptable.

3. ^Ideally, we could use latent variables to account for measurement error, but as this is computationally intensive and leads to an unreasonable computation time, we choose to rely on sum variables, whose reliability we have tested to the best of our ability.

References

Abbott, S. (2010). Social capital and health: the role of participation. Soc. Theor. Health 8, 51–65. doi: 10.1057/sth.2009.19