Introduction

The development of civic engagement is an important part of identity formation in adolescence (Erikson 1985), involving an understanding of one’s role and connection to the broader society (Yates 1999) and is thus an integral component of citizenship. According to Erikson (1968), the achievement of community awareness is important for personal development during adolescence when youths develop their identity and start to question how they fit into the society that goes beyond their family and friends. Indeed, civic engagement refers to attitudes, behaviors, knowledge, and skills aimed to work for the common good. In the current work, we focused on the behavioral component of civic engagement, defining it as membership and participation in local organizations (Flanagan et al. 1998). Evidence shows that civic engagement positively relates to psychosocial adjustment (Schmidt et al. 2007). Adolescents have the ability to improve their schools, neighborhoods and the broader society by volunteering in community-based organizations and through engaging in informal prosocial activities. This involvement in community life can promote psychological, social and intellectual growth for young citizens (Fredricks and Eccles 2006; Johnson et al. 1998).

At the same time, adolescent civic engagement can provide services to the local community, thus promoting the effective functioning of society (Flanagan and Sherrod 1998). Since studies have shown that civic responsibility during adolescence predicts civic responsibility in adulthood (Youniss et al. 1997; Zaff et al. 2008), it is important to understand which factors predict the development of civic engagement at this developmental stage. Although historically civic competence has been included among the central tasks of adolescence (Havighurst 1972), the evidence on which factors can promote the development of civic engagement in adolescence is limited (Zaff et al. 2008).

Recent studies broadened the comprehension of the development of civic competence based on ecological systems theory (Bronfenbrenner 1979), showing the importance of life contexts for the socialization of adolescents to different goals and behaviors. More evidence on the positive association between community characteristics and civic participation has been gathered in adult populations than adolescents (e.g. Prezza et al. 2001). Research has shown, for instance, that a sense of community is a catalyst for participation (Cicognani et al. 2008). Moreover, in studies evaluating the relation between life contexts and civic engagement, the characteristics of these contexts are generally measured through individual perceptions, instead of aggregating them in order to obtain the average perception of individuals living in a particular context. Recent methodological developments in the study of life contexts show the importance of measuring them both at the individual and at the contextual or aggregate level (Shinn and Toohey 2003). Furthermore, international studies in this field are rare, but are helpful for understanding both the influence of different cultural contexts and universal characteristics that can foster civic engagement (Flanagan et al. 1998; Torney-Purta et al. 2008). The current research extends previous literature evaluating the role of socio-economic status, school and neighborhood contexts, conceptualized both at the individual and at the contextual level, in fostering civic engagement in adolescents from five countries.

Defining Civic Engagement

Broadly construed, civic engagement includes attitudes, behaviors, knowledge, and skills that benefit society and derive from the interest in improving the common good. Civic attitudes, variably called civic responsibility, civic mindedness, or civic identity (Youniss et al. 1997), indicate feelings of responsibility toward the communities in which an individual is embedded, and the idea that everyone has a central role in influencing the well-being of society. Civic behaviors include actions that, based on this belief system, aim to resolve community issues and improve the welfare of the society, such as volunteer work and supporting charities. Such behaviors are generally termed civic participation (e.g. Zaff et al. 2008), civic involvement or civic engagement, but the terms are often used interchangeably. We emphasize membership and participation in local institutions as defining elements of citizenship and civic commitment (Flanagan et al. 1998).

Defining and measuring civic engagement is particularly challenging during middle adolescence when political involvement is uncommon. At this developmental stage, the activities of political and voluntary organizations are both aimed at fulfilling the needs of the local community (Youniss et al. 1999). Adolescents’ involvement in volunteer and political service puts adolescents in contact with people in need, and allows youths to directly contribute to other people’s well-being. Other organizations that generally engage adolescents in community life include youth organizations (e.g. Scouts) and religious organizations. These kinds of organizations supplement the education of young people and help them create a value system that underlines their constructive role in improving society. Unlike political and voluntary associations, in these organizations contributing to the common good is not the only aim; activities of these organizations also include playing and having fun with other boys and girls. Finally, other clubs in which youth can participate in structured activities include sport and cultural organizations, which imply a minimal involvement in improving the community (aside from, e.g. organizing local events).

Thus, community organizations can be placed on a continuum based on their degree of civic purpose and community improvement activity. Based on the type of organizations in which youths are involved, it is possible to assume different amount of civic engagement: maximum when young people volunteer or participate in political activity in the community, minimum when they take part in sportive or cultural organizations. Like studies that suggested that extracurricular activities are an early manifestation of civic involvement (Sherrod et al. 2002; Yates and Youniss 1998), we define civic engagement in the present study as based on adolescents’ club memberships. We take into account both the number of organizations in which youth participate and the level of civic purpose that generally characterizes these organizations (Vieno et al. 2007a). By creating an index based on the amount of involvement in community life in which the different organizations tend to engage, the focus of the present study is on adolescents’ civic engagement: how much young people, depending on the kind of organization attended, participate and contribute to community life.

The Role of Family Socio-Economic Status, School and Neighborhood in Civic Engagement

Since civic engagement is a multidimensional concept, including attitudes, behaviors and skills, studies evaluating its correlates have used multiple definitions; some studies, for example, focused on civic attitudes, while other studies considered behaviors aimed to contribute to community life, mainly by participating in local organizations. Despite the differences in the definition of civic engagement, research on the correlates of civic engagement are based on a similar assumption: during adolescence, socializing agents transmit to youths different goals and behaviors, instilling a sense of commitment to work for the common good, which in turn makes it more likely they will join a community organization where it is possible to take action (e.g. directly helping people in need, collecting money for a cause).

Although the evidence on the predictors of youth civic behavior is limited, previous research suggests the importance of some individual and contextual contributing factors. Research on the origins of citizenship has analyzed correlates of civic participation such as knowledge about government and political processes (Furnham and Stacey 1991). Recent psychological research showed the roles played by demographics: being female (Da Silva et al. 2004; Flanagan et al. 1998), native-born (Torney-Purta et al. 2007), and having high socio-economic status (Atkins and Hart 2003), are all associated with higher levels of civic knowledge and engagement. Socio-economic status, in particular, relates to parental knowledge of opportunities in the local community and easier access to these resources. In families with a higher socio-economic status, parents are more connected with social networks and institutions, thus knowing and valuing community organizations that involve adolescents, and assisting their children in accessing organized out-of-school activities (Coulton and Irwin 2009). Moreover, more affluent families can better afford the costs associated with such participation.

Theoretical models have underscored socializing agents that foster or deter the development of civic engagement. Ecological systems theory (Bronfenbrenner 1979) and the Social Development Model (Catalano and Hawkins 1996) posit that family, peers, school and community contexts socialize individuals to different goals and behaviors, instilling in some a moral commitment to contribute to the common, helping other people and participating in community organizations. The Social Development Model, a theory initially developed for understanding antisocial behavior, posits that involvement with prosocial school and community members increases the likelihood that the adolescent will adopt the beliefs and behaviors of the group (Catalano and Hawkins 1996). Similarly, political theorist Walzer’s (1989) work on citizenship argues that youth learn the meaning of citizenship through their own experiences of membership in local communities and institutions.

Schools contribute to students’ civic knowledge and foster engagement in community life. Teachers can help accomplish this by establishing a democratic climate for learning and social interaction (Flanagan et al. 2007; Vieno et al. 2005). An open school climate in which students take part in making rules and organize school events is positively associated with the development of democratic skills, such as perspective taking and trust in other people and institutions (Hahn 1998). Moreover, studies suggest that students’ perception of a democratic school climate is positively associated with knowledge about international affairs and ability to think and to act critically about civic issues (Newmann 1990), and commitment to vote (Campbell 2008). Thus, experiencing a democratic, participatory climate at school can foster adolescents’ commitment to certain goals and values related to the common good, which in turn increases the likelihood of joining a community organization.

In the youth development literature, there is increasing evidence that neighborhood social resources and adolescents’ ties to their community influence young people’s well-being (Leventhal and Brooks-Gunn 2000; Youngblade and Curry 2006). When youth feel that there are people in the local community they can turn to, they are less likely to experience emotional and behavioral problems and more likely to report feelings of competence (Pretty et al. 1996; Quane and Rankin 2006). Generally, these studies defined neighborhood social resources as the perceived level of connectedness among people in the community—that is, the extent to which they care about one another and their willingness to contribute to the common good. The perception of living in a neighborhood characterized by high levels of social connectedness is associated with individuals’ commitment to give back to their communities, working to make it a better place (Flanagan et al. 2007; Albanesi et al. 2007). This evidence underlines the reciprocal relationship between neighborhood social resources and civic engagement described in the literature on social capital, which refers to social networks characterized by norms of trust and reciprocity facilitating cooperative action among citizens and institutions (Putnam 1993). From a developmental perspective, living in a neighborhood with high social capital, in which relationships among people are characterized by trust and reciprocity, can nurture adolescents’ willingness to work on making the local community, and the larger society, a better place. There is a process of collective socialization in which helping behaviors and interest for the common good are learned from people who adolescents meet daily in the neighborhood (Jencks and Mayer 1990). Moreover, adolescents want to give back to communities that support them, by joining local organizations where they can volunteer helping people in need or organizing events in their community.

Thus, theoretical models and research evidence support the role of schools and neighborhoods as microcosms of public life, in which adolescents have opportunities to exercise rights and assume responsibilities as members of those local communities. A democratic school climate and high levels of neighborhood social capital can create norms and an infrastructure to support civic engagement (Putnam 2000).

Most studies that analyzed the role of school and neighborhood in promoting civic engagement measured these contexts at the individual level, through adolescents’ perceptions. The use of aggregate perceptions and other ecological predictors in multilevel models provides a more reliable estimate of school and community characteristics. Research on correlates of social competencies (Romano et al. 2005) underlines the importance of measuring life contexts at an aggregate level, through multilevel modelling analysis. Contextual levels of democratic school climate and neighborhood social capital, in particular, were found to influence adolescent general well-being (Fauth et al. 2007; Vieno et al. 2005), independent of individual-level effects.

The Importance of Cross-national Studies on Civic Engagement

Social and cultural norms towards civic engagement, as well as opportunities for youth to be involved in local organizations, may vary in different cultural contexts, but cross-national studies in this field are rare. Existing studies that compare adolescents’ civic participation across countries are usually based on the assumption that, although national differences may encourage or hinder civic participation (e.g. by providing different levels of opportunities to be involved in community organizations), it is important to detect correlates of civic engagement in more proximal contexts (e.g. school and neighborhood). Thus, the main assumption of these studies is that individual and mesosystem characteristics of family, school or neighborhood may be more closely associated with adolescent civic engagement than macro-level factors. In a study of adolescents in 27 countries (Torney-Purta et al. 2008), in spite of social and cultural differences across countries regarding youth civic engagement, experiences of democracy at school were associated with knowledge of human rights and political efficacy. After controlling for country level factors (freedom index and duration of democracy), the study showed that students who perceived an open classroom climate had a system of beliefs favoring social movement citizenship, positive immigrants’ rights attitudes, and higher levels of political efficacy. In another study comparing seven countries, Flanagan et al. (1998) obtained somewhat different results: adolescents’ sense of membership at school was positively related to students’ levels of civic commitment in most of the countries included in the study, while democratic school climate predicted adolescents’ civic commitment in only two of the countries included in the study. These results show how the effects of distal and proximal contexts may interact in creating a complex system of effects. For this reason, there is a need to more deeply explore correlates of civic engagement in a cross-national perspective and to identify both universal and country-specific factors related to civic behavior.

In the current study, five countries which differ in history and tradition regarding civic life were included: England, Belgium, Italy, Romania and Canada. The contextual background of participation in civic life varies greatly between Europe and North-America, and it is quite diverse even between European countries (GHK 2010). In England, for instance, there is a longstanding tradition of volunteering and social participation, which was nurtured by the government in the 1980s, with the introduction of a contract culture encouraging local organizations to provide services to the community. Volunteering, in particular, remains profoundly embedded within the UK’s social policies and it is considered a core element of strategies to increase collective citizenship among adults and young people. Belgium, compared to the rest of Europe, can be considered a typical European country in relation to civic attitudes and traditions, where people show average levels of trust and participation in local organizations (Newton 2007). A peculiarity of the Belgian context which deserves to be noted is related to civic education in schools, which is managed with a considerable degree of autonomy, with schools deciding how much effort to spend in civic classes and activities (Hooghe and Claes 2009). Italy also has a longstanding tradition of civic and voluntary organizations; the history of community organizations in Italy is mostly related to church-based institutions, which until the ninetieth century guided most of the charitable activities in the social, welfare and health domains. Although, particularly in the last decades, the number of secular civic organizations has grown, a large number of youth and adolescents are still involved in forms of community service related to religious organizations (Vieno et al. 2007a). Unlike England, Belgium and Italy, which have strong history and traditions of participation in community life (although with their own peculiarities), in Romania the voluntary sector is still emerging. It is worth noting that one of the main factors impacting participation in community organizations in post-communist countries is the heritage of communism. Although with some peculiar characteristics, post-communist societies show a somewhat negative attitude towards volunteering, which could partly derive from the communist era, when people were forced to participate to state organizations; as a consequence, many citizens tend to refuse to participate in any form of civic initiatives. Canada, finally, represents an example of country where norms of participation and civic commitment are strongly embedded in the everyday life, through charitable giving, volunteering and participating in non-profit organizations. As the Canada Survey of Giving, Volunteering and Participating (CSGVP) showed, a high percentage of Canadians are involved in local organizations: according to this survey, during 2004, 45% of the population aged 15 and over, volunteered in a community organization (http://www.givingandvolunteering.ca). The high levels of civic engagement among Canadians can partly derive from “service learning” programs, which involve mandatory service in community organizations and are part of the education in many high schools.

Along with the historical and cultural traditions in civic life, in order to compare the levels of opportunities to be involved in community organizations in these five countries, it is useful to consider the “clubs and associations” indicator included in the Indices of Social Development (ISD),Footnote 1 which includes data on membership of local groups, time spent socializing with relatives and in local clubs, attendance of community meetings and participation in non-profit associations.

The “clubs and associations” index is comprised between 0 (low levels of club and association participation) and 1 (high levels of club and association participation). In the 2005 Social Development Survey, Belgium’s and England’s score on this index was 0.55, while Italy’s score was 0.49, thus approximating the average score of the countries included in the study (M(SD) = 0.50(0.11)). In Canada, instead, the “clubs and associations” index was above the general mean (0.59), while Romania obtained a score below the mean (0.35).

Overall, considering both the civic traditions and opportunities characterizing the countries, we can note that three of them (England, Belgium and Italy), despite their own particularities, tend to promote individuals’ civic engagement and participation in community life by providing an adequate quantity of opportunities to be involved in local organizations. In Canada, norms of participation and civic commitment are highly embedded in everyday life; this country is characterized by high levels of participation in local organizations, which is considered a critical aspect of civic life; at the same time, there are many opportunities to participate in community organizations, especially for young people. On the contrary, Romania is a country where both levels of participation and opportunities to be involved in local organizations tend to be lower.

Aims and Hypotheses

The principal aim of the current study was to understand the role of family affluence, school climate and neighborhood social capital in promoting civic engagement in a representative sample of adolescents from five countries (Belgium, Canada, Italy, Romania, England). We hypothesized that higher levels of family affluence, perceived democratic school climate and neighborhood social capital predict participation in community organizations.

The second goal of the study was to compare the relative contribution of these influences at the level of individual perceptions and the aggregate school or neighborhood level. Since present data come from nationally representative surveys, the school level was used as a proxy of area of residence for measuring neighborhood social capital at the aggregate level. Similarly, family affluence at the aggregate level was measured at the school level. We expected family affluence, democratic school climate and neighborhood social capital to show effects at both individual and aggregate levels.

Finally, the current study aimed to evaluate whether predictors of adolescent civic engagement are consistent or whether they vary across countries. Social and cultural norms towards civic engagement, as well as levels of participation and opportunities for youth to be involved in local organizations, vary across the five countries included in the study. Despite these differences, we expect that family, school and neighborhood contexts influence the levels of adolescents’ civic participation in the five countries. Moreover, we expect that contextual influences will be stronger in countries where social and cultural norms strongly encourage civic participation, where average levels of civic involvement are high, and there are many opportunities available to participate in community organizations (Canada, England, Italy); indeed, it is plausible that social and cultural norms at the country level impact neighborhood, school and family contexts, because distal factors (e.g. government decisions and social policies) may influence more proximal experiences (Torney-Purta et al. 2008).

Methods

Setting and Sampling

Data were collected in the 2005/2006 Health Behaviour in School-aged Children (HBSC) study, a standardized, cross-national survey carried out in collaboration with the Regional Office for Europe of the World Health Organization (Aarø et al. 1986). The research protocol involved repeated cross-sectional surveys of 11-, 13- and 15-year-old students in representative samples of schools in 42 European and North-American countries and regions, although only responses from 15-year-olds in the five countries that administered the HBSC Optional Questionnaire (Belgium, Canada, Italy, Romania, England) were used here. Indeed, “civic engagement” is part of an HBSC optional package (Currie et al. 2002), so that researchers in each country can decide whether to include or not this measure: in the 2005/2006 Survey, five countries with different cultural and political characteristics chose to include this measure, so that data from these countries were considered suitable for the aims of the current study. The students completed a standardized self-administered questionnaire during a school lesson, following the instruction from a trained adult. Parental permission was obtained before the administration.

Participants

Each national sample comprised of students in the relevant age groups from a random sample of schools, but the research protocol established that some of the variables included in the present analyses (neighborhood social capital, civic engagement) can be reliably measured only on the 15 year-olds. Thus, the study included 8,077 secondary school students from five culturally and economically diverse countries (see Table 1).

Table 1 Descriptive statistics for civic engagement, family affluence, democratic school climate and neighborhood social capital by gender
Full size table

The current analyses excluded 660 students (8.2%) with missing information on one or more of the variables of interest (family affluence, democratic school climate, neighborhood social capital, civic engagement). We compared the sub-sample excluded from the analysis and the final sample in terms of gender distribution and, if known, family affluence (FAS) levels. The excluded sub-sample does not differ from the final sample in terms of gender distribution (χ² (2) = 1.59, n.s.), but it differs significantly in terms of FAS levels, which were lower among excluded students (F (1, 7,763) = 17.71, p < .001).

The sample was obtained through a multistage cluster sampling procedure, in which first the schools were randomly selected; then, in selected schools, one class for each age group was sampled randomly. Due to differences in the school-systems across countries, however, some national adaptations were made. Due to the multistage sampling, non-response may occur at the school, class, and student level. The rate of schools agreeing to participate was generally high, with the majority of the countries above 80% (Currie et al. 2008).

Measures

The self-report questionnaire devised by the HBSC international group assessed health behaviors of adolescents. For the present study we focused on 15-year-olds’ reports of civic engagement, family affluence, gender, democratic school climate, and neighborhood social capital.

Civic Engagement

This was operationalized as the number of different kinds of community organizations in which adolescents are involved weighted by the general civic purpose of each type of organization. The items asked participants in which of the following organizations they participate: sports clubs, voluntary service, political organizations, cultural associations, church or religious groups, and youth clubs (by indicating “yes” or “no”). We attributed different scores based on the contribution that these organizations generally give to the local community: voluntary and political organizations scored 3 (high civic purpose and frequent involvement in community improvement), youth and religious groups scored 2 (moderate civic purpose and occasional involvement in community improvement), and sport and cultural associations scored 1 (low civic purpose and infrequent involvement in community improvement). Then, responses were added up, obtaining a single measure of civic engagement ranging from 0 to 12. For instance, a young person participating in a sportive club (score 1), a voluntary organization (score 3) and a religious group (score 2) would score 6. Although the response categories do not indicate someone belongs to more than one organization in the same category, this operationalization of the variable allowed us to take into account both the number of different kinds of organizations in which adolescents participate and the level of civic purpose or community improvement that characterize their activities (because a higher score means participation in many organizations and high amount of civic engagement).

Demographics

Students reported their gender and family socio-economic status, which was measured by the Family Affluence Scale (FAS), a four-item measure developed and validated in the HBSC study (Boyce et al. 2006), which included four indicators of family affluence: family car ownership, unshared rooms, number of computers at home, and times spent on holiday in the last 12 months. Responses were summed and the total scores (ranging from 0 to 9) were divided into three groups, adopting the cut points recommended by previous research (Boyce et al. 2006): students scoring between 0 and 2 were placed into the low affluence category, those with scores between 3 and 5 were grouped into the moderate affluence group, and those between 6 and 9 in the high affluence category.

Democratic School Climate

Individual perception of democratic school climate was assessed using the following five items (Vieno et al. 2005): (1) “In our school students take part in making rules”; (2) “The students get involved in organizing school events”; (3) “The rules in this school are fair”; (4) “I am encouraged to express my own views in my classes by my teachers”; (5) “Our teachers treat us fairly.” Responses were rated on a 5-point scale ranging from 1 = strongly disagree to 5 = strongly agree. Alpha reliability for the five-item scale was 0.63; items were averaged for the measure of student perception of democratic school climate, so that higher scores indicated higher perceived democratic school climate.

Neighborhood Social Capital

Similar to previous work (Boyce et al. 2008), neighborhood social capital was measured with a 5-item scale developed in the HBSC study, including :“people say ‘hello’ and often stop to talk to each other in the street,” “it is safe for younger children to play outside during the day,” “you can trust people around here,” “there are good places to spend your free time,” and “I could ask for help or a favor from neighbors”. Responses ranged from 1 = strongly disagree to 5 = strongly agree. All items were extensively piloted in the HBSC survey, except for “you can trust people around here,” which was included based on Kawachi et al. (1999). The Cronbach’s alpha for the scale was 0.72. A single measure of neighborhood social capital was created by averaging participants’ responses to the five items.

Analytic Approach

Data at the school level were obtained by aggregating FAS, democratic school climate and neighborhood social capital, as reported by student respondents. Since these kinds of data are inherently clustered, with adolescents having been sampled within schools, we used the multilevel regression technique of hierarchical linear modeling (HLM; Raudenbush and Bryk, 2002). A brief description of the statistical models used is provided in the analysis section.

The dependent variable for this study (civic engagement) assumed the shape of a Poisson distribution. Therefore, Poisson regression was used to model the rate of civic engagement as a function of individual and aggregate variables. The link function is the logit link. A model with the total sample was first run, and then we evaluated independent models for each country.

The within-school (level 1) model estimates the influence of family affluence, perceived democratic school climate and neighborhood social capital on civic engagement for student i in school j, controlling for gender. Family affluence, perceived democratic climate and social capital were centered around the school mean, entailing that the estimate of school-mean measures are unadjusted for between school variation in these variables; this way it is possible to examine the between-school influence of the aggregates of these variables at level 2 (Raudenbush and Bryk 2002). The individual-level model includes three predictors and one demographic control variable:

$$ \eta_{ij} = \gamma_{0j} + \gamma_{1j} \left( {\text{Female}} \right) + \gamma_{2j} \left( {\rm{FAS}} \right) + \gamma_{3j} \left( {{\text{Dem}} . {\text{ School}}} \right) + \gamma_{4j} \left( {\text{Social Capital}} \right) $$

We considered the contextual effects on adolescent levels of civic engagement as a function of family affluence, school climate and neighborhood social capital average levels. We explored possible effects on the adjusted school log-odds of civic engagement, γ0j.

$$ \gamma_{0j} = \beta_{00} + \beta_{01} \left( {\text{MEANFAS}} \right) + \beta_{02} \left( {{\text{MEANDEM}} . {\text{ SCHOOL}}} \right) + \beta_{03} \left( {\text{MEANSOCIALCAPITAL}} \right).$$

Results

Preliminary Analyses

Descriptive statistics for the variables on each level are shown in Table 1. On average, the reported levels of civic engagement were modest (total sample mean = 1.62; 34.8% of participants not involved in any kind of organizations). In particular, in Belgium, Italy and England the levels of adolescents’ involvement in community organizations were quite similar (respectively 1.52, 1.60 and 1.69); Romanian adolescents were the least involved in civic organizations (1.27); Canadian youth had the highest levels of civic engagement (1.87), but that was still modest in a 0–12 scale.

Within- and Between-School Analysis

The within- and between-school HLM models predicting civic engagement for the total sample and separately for each of five countries are shown in Table 2. The within-school model includes the three predictors and the demographic control variable (gender). In the total sample model, each individual-level predictor showed a significant effect on students’ civic engagement: students who belong to more affluent families (OR = 1.18), and who perceive a more democratic climate in school (OR = 1.13), and social capital in the neighborhood (OR = 1.10) reported more involvement in community civic organizations. No gender differences were found.

Table 2 Multilevel logit regression estimates for “civic engagement” for total sample and by countries
Full size table

In the independent models run in each country, results generally followed the above pattern with slight variations. Belonging to more affluent families is associated with higher levels of civic engagement in all but one country in the study (Italy). Similarly, student perception of democratic climate at school was related to an increased likelihood of being involved in community organizations in three out of five countries (Canada, Romania, England). Individual perception of social capital in the neighborhood was associated with higher levels of civic engagement in Canada, Italy and Romania. Gender differences were found only in Canada, where levels of civic engagement were higher in girls, and in Italy, where boys were more involved than girls.

In order to evaluate the impact of between-school variation in family affluence, perception of democratic school climate and neighborhood social capital on students’ civic engagement, we included the school-level mean of family affluence, democratic climate and neighborhood social capital as predictors at level 2. In the model including the total sample, school-levels of family affluence (OR = 1.69) and democratic climate (OR = 1.35) had a positive overall effect on adjusted school mean civic engagement: being part of a school in which students, on average, come from more affluent families and perceive higher levels of democratic school climate was associated with a higher involvement in community organizations. Neighborhood social capital, measured at the aggregate level, showed no effects on adolescents’ civic engagement.

Considering the single models evaluated in each country, results showed that family affluence at the aggregate level had a positive effect on adjusted school mean civic engagement in all but one country (Romania) included in the study. Similarly, school levels of perceived democratic school climate were positively associated with a higher participation in community organizations in three out of five countries: Belgium, Canada and England. Neighborhood social capital was the only variable that, at the aggregate level, had no effect on civic engagement in any of the five countries.

Discussion

The study explored the role of family affluence, democratic school climate and neighborhood social capital in predicting civic engagement during adolescence in a representative sample of students in Belgium, Canada, Italy, Romania and England. Results showed that, in the total sample, family affluence, a democratic school climate in which students can express their point of view, and perceiving that one lives in a neighborhood where people trust and help each other, are associated with more involvement in civic organizations; however, results also showed some cross-country differences in correlates of civic engagement.

For the first aim of the current work, we analyzed data from the total sample. At the individual level, each predictor (other than the control variable gender) affected civic engagement. In line with previous research (Atkins and Hart 2003), family affluence seems to promote involvement in community improvement. In families with more economic resources, in fact, adolescents can have easier access to organized activities because they can afford the costs of participation (e.g. membership dues, transportation). Moreover, family affluence is usually related to social and educational status, increasing the likelihood that affluent parents know the local organizations and value adolescents’ participation in them (Coulton and Irwin 2009). Economic security can also improve the quality of parent–child relations (Leventhal and Brooks-Gunn 2000), creating a family environment in which it is easier for parents to encourage youth to participate in structured activities. All these characteristics can result in a wide range of opportunities for adolescent involvement, and in higher levels of participation.

Regarding the role of schools, we found that a democratic climate in the classroom and elsewhere in school related to civic engagement. Students who feel free to express their point of view and participate in organizing school events tend to be more involved in organizations outside of school. As demonstrated by previous studies (Flanagan et al. 2007; Vieno et al. 2005) schools can influence nurturing students’ engagement in community life. In particular, beside the central role of school civic education (Syvertsen et al. 2009), teachers can transmit civic values by establishing an open climate for learning. In a learning environment that allows students to influence school life, students have the opportunity to exercise rights and assume responsibilities as active members of the institution, thus improving their democratic skills (Hahn 1998; Newmann 1990), civic knowledge (Torney-Purta et al. 2008) and commitment to the common good. A school climate that encourages students’ democratic participation can foster civic values, knowledge and skills, that are the basis for a higher level of involvement in community life.

Neighborhood contexts might also influence civic engagement, but these findings suggest that it was individual perceptions of neighborhood social capital—rather than the aggregate, school-level social capital—that related to civic participation. One explanation is that just feeling that one lives in a neighborhood with high levels of social connectedness, where people trust and are willing to help each other, may enable adolescents to get involved in community organizations. Alternatively, participation in such organizations may positively influence young people’s perception of social capital, independent of the actual level of neighborhood trust and cohesion. Most likely, civic engagement and perceptions of social capital are mutually reinforcing. Interacting with people in community organizations that are based on norms of trust and reciprocity, adolescents can learn civic values and develop an interest to improve the local community (“collective socialization”, Jencks and Mayer 1990). At the same time, the relations developed in the neighborhood can be an important source of support for adolescents, who may develop an emotional bond, and a commitment to give back, to their local community, working to make it a better place (Albanesi et al. 2007; Brown et al. 2003; Flanagan et al. 2007).

Since previous research on the role of family, school and neighborhood factors in promoting civic engagement measured these correlates at the individual level, we also evaluated these correlates at the aggregate level, using the school as level of aggregation. The analysis using the total sample (across all five countries) shows a positive association between both family affluence and school climate measured at the aggregate level and civic engagement. The link to family affluence is stronger when it is measured at the school level: adolescents attending a school where the average family affluence is high are more civically engaged than students in low affluence schools. Thus, the level of affluence of the family can influence adolescents not only through processes that occur inside the family, but also when a concentration of advantage/disadvantage is created. Schoolmates are central in the social networks of adolescents (Vieno et al. 2005), and they can impact students’ motivation to join an organization or participate in structured activities. Adolescents can learn about existing opportunities for involvement from their schoolmates; they can also decide to get involved in activities simply because friends are already participating or are deciding together to join a particular organization (Loder and Hirsch 2003). Families, schools and other institutions in more affluent areas may also have the resources to create more opportunities for youth to participate in civic activities (Leventhal and Brooks-Gunn 2000; Leventhal et al. 2009).

The results also show that aggregate perceptions of democratic school climate relate to adolescents’ civic engagement, and this link is stronger than with individual perceptions of school climate. This result reinforces the idea that a school can be conceptualized as a community (Allodi 2002; Vieno et al. 2007b); when students in a school share the perception that they can have a voice in their school and that their opinion is valued, the overall climate will promote participation in school life and may spill over to extracurricular life outside of school. The school is a microcosm of public life where students learn democratic principles and values and engage in civic life.

Unlike studies that define social capital as a characteristic of the context (Kawachi et al. 1999; Putnam 2000), neighborhood social capital, measured at the aggregate level, is not associated with adolescent civic engagement. According to Putnam (2000) and others, social capital operates not just at the level of individual perception but at the community level as well, as neighborhood social capital is thought to promote physical and psychological well-being for all people in the community. The results of our study, instead, support Bordieu’s (1985) definition of social capital as a product of social relationships (among individuals), and a psycho-behavioral conceptualization of social capital (Perkins et al. 2002). We cannot draw definitive conclusions about this dichotomy, because of two ecological measurement issues. First, neighborhood social capital was measured by perceptions of informal social cohesion among ambiguously aged “people” and “neighbors.” Those subjective individual perceptions motivate the youth who holds them toward organized civic engagement, but may not reflect other measures of neighborhood-level social capital, such as organized citizen participation, social networks, or how adults may perceive the same community. Second, we were constrained to using the school as a proxy for neighborhood. Although most students from the same school live in the areas closest to that school, that can still represent several different neighborhoods that we treated as one. Thus it is possible that school-level neighborhood social capital was unrelated to civic engagement because the level of aggregation was imprecise.

The last aim of the study was to evaluate whether predictors of adolescent civic engagement were consistent across countries. Results within countries show some variation. In Canada, all predictors showed significant effects, with the exception of social capital at the aggregate level. This is similar to the results for the total sample and may be due in part to Canada having the greatest level and variance of civic engagement among the five countries. Canada was the only country in our sample with more civic engagement among girls, which contrasts with Flanagan et al. (1998), who consistently found girls more likely to engage in volunteer work in Eastern European countries, Australia and the United States. Similar to the total sample, family affluence and democratic school climate had an even stronger effect in Canada at the aggregate than at the individual level. In schools where students come from more affluent families, adolescents are 2.64 times more likely to be civically engaged. The stronger effect for democratic school climate at the aggregate level suggests the possibility that individual schools or school districts in Canada may engage in service-learning projects that encourage students to get and stay involved in civic organizations outside of school. Indeed, the work of Pancer and Pratt (1999) shows the importance of mandatory community service at school for adolescents’ attitudes and intentions to volunteer: although many youths may begin to participate in community service because they are required to, a positive experience in the organizations increases adolescents’ participation in community life (Taylor and Pancer 2007).

In England, the results are similar to Canada except there was no gender difference and neighborhood social capital had no effect at either the individual or aggregate level. Family affluence and the individual perception of democratic school climate were positively associated with civic engagement, but the association is stronger when these predictors are measured at the school level. According to our results, in England the role of school in promoting civic engagement is central and—as in Canada—suggests that where students agree that their school is open and democratic, it may foster greater citizenship behavior outside of school.

The most notable result in Belgium regards the association between family affluence at the school level and civic engagement. Beyond a moderate effect of school-level democratic school climate and a modest individual-level effect of family affluence, school-wide family affluence was the strongest predictor of civic engagement. The Belgian sample was not more affluent than the other countries (with the exception of Romania), nor did it have greater inequality (variance) of affluence. It appears that in Canada and Belgium, neighborhoods or schools with more affluent families do an especially good job of providing opportunities for all youth in the community to participate in community organizations. While in Canada this result could be mostly derived from the availability, at school, of service learning programs, in Belgium, where civic education is managed with a high degree of autonomy, schools attended by students coming from more affluent families may invest more in civic classes and activities (Hooghe and Claes 2009). Moreover, it may be more common in those two countries for affluent families to subsidize such participation for all local youth through charitable donations to organizations or other means of holding down the direct costs of participation.

By contrast, in Romania, family affluence predicts civic engagement at the individual but not the aggregate level, suggesting no such community-level subsidy for youth participation. This may be partly explained by the greater variability on individual family affluence in Romania compared to other countries. Perceived democratic school climate and neighborhood social capital are also associated with higher levels of civic engagement among Romanian adolescents. But all these effects only occur at the individual level. Perhaps in Romania, decisions on civic engagement are more personalized and not something that is widely encouraged or organized at the community level. This result could be partly related to the tendency of post-communist societies to have relatively negative attitudes towards community participation; as a consequence, in Romania it may not be common to attend a school or live in a neighborhood where social norms encourage civic participation in adolescence.

Finally, Italy was the only country, either in our sample or the seven others studied by Flanagan et al. (1998), where boys were more engaged in civic organizations than were girls. This result can be related to the inclusion of all kinds of participation in the definition of civic engagement; Italian boys are generally more involved than girls in sport and political organizations. Notably, Italy was the only country in the study where the affluence of the family appeared to have no effect on civic participation. A possible explanation could be related to the high involvement of Italian adolescents in religious organizations that tend to gather people from different economic backgrounds. Family affluence had a significant effect only at the school level, due to a concentration of advantage/disadvantage that can influence adolescents’ opportunities to join an organization. Furthermore, similar to Canada and Romania, the perception of living in a neighborhood with high levels of social capital was associated with higher levels of civic engagement.

The results confirm that in most countries, family affluence, democratic school climate and neighborhood social capital are important correlates of civic engagement in adolescence. With some exceptions (mainly aggregate-level neighborhood social capital), they have a positive effect on adolescent civic engagement. Indeed, in each one of the five countries, at least two out of three contextual correlates considered (family affluence, democratic school climate, neighborhood social capital) were positively associated to adolescents’ civic engagement (either at the individual or at the aggregate level). In general, the results show the importance of family, school and neighborhood contexts for the development of civic engagement in adolescence. However, variations between countries suggest how cultural or political aspects characterizing these countries may influence the correlates of civic engagement.

Limitations, Strengths, and Conclusions

The study has some limitations to acknowledge. First, the cross-sectional nature of the data did not allow conclusions about the direction of the effects. It is possible that adolescents who are more involved in community organizations have a different perception of school climate and neighborhood social capital, being more aware of school and neighborhood opportunities and evaluating them in a more positive way. The direction of the relationship between these variables and levels of civic engagement can only be evaluated using longitudinal studies.

Another limitation was the use of self-report, rather than independent, measures of school and neighborhood characteristics. Using the same source and method for all data collection risks reporting biases because the outcome can affect the evaluation of school and neighborhood contexts (Diez-Roux 2007). For example, students who participate more in civic organizations may be more likely than those who participate less to report social resources in their school and neighborhood, irrespective of the actual condition of these contexts.

Third, the sample was slightly more affluent than average due to missing data. This bias was mitigated by including FAS in the models at both levels thus controlling for its influence on civic engagement, although that did not completely control for sampling bias as the reduction in FAS variance means it might have had less influence on the models. Family and neighborhood affluence may play an even greater role in youth civic engagement than we found.

Beyond these limitations, a strength of the study was a large international sample, representing five European and North-American countries. Cross-national comparisons underscored that some characteristics of adolescents’ life contexts are more influential in some countries than in others. At the same time, the results showed that, although cultural contexts and educational systems may be different, family affluence, democratic school climate and neighborhood social capital are associated with adolescent civic engagement in different countries.

The current study demonstrated the need to study contextual correlates of civic engagement not only in different parts of the world but in greater qualitative depth in order to gain a better understanding of the psychological processes through which social contexts influence adolescent development. Indeed, qualitative methods (e.g. focus groups, structured and unstructured interviews) permit collection of the narratives of adolescents, thus obtaining a detailed description of their lived experience in the local community and their levels of civic engagement. This way, they are free to narrate their experience within the neighborhood, elucidating, for example, the processes through which normative systems in the local community can influence individual behavior. Understanding the mechanisms responsible for the association between neighborhood characteristics and adolescent civic commitment is critical in order to develop promotion programs based on empirical evidence (Durlak et al. 2007). Interventions that increase democratic school climate and at least adolescents’ perceptions of neighborhood social capital may promote young people’s civic engagement in different countries; and the actual level of neighborhood social capital tomorrow depends on young people’s civic engagement today.