1 Introduction

Public opinion polls and political developments across advanced democracies suggest that public commitment to democracy might have weakened in recent years (e.g., Pew Research Center 2017; Plattner 2017). The other side of this coin is that support for authoritarian ideas and leaders is concomitantly on the rise in many places (Bermeo 2016; Wike et al. 2017; Levitsky and Ziblatt 2018; Waldner and Lust 2018). These large-scale trends—alternately labeled “democratic backsliding,” “democratic deconsolidation,” or “democratic recession”—call for renewed scholarly focus on citizen support for democracy (Diamond 2015; Norris and Inglehart 2019).Footnote 1 Given the demonstrated importance of mass support for democratic regime survival (Lipset 1959; Claassen 2020), continued research on this subject is essential. The purpose of our analysis is to unpack the individual-level causes of support for democracy.

Research on this subject is especially necessary because some evidence signals that young generations are among the least enthusiastic about democracy (Denemark et al. 2016; Foa and Mounk 2017; Norris 2017).Footnote 2 This comes as something of a surprise in light of theorizing that successive generations socialized into democratic norms produce a predictably steady supply of pro-democratic citizens (Mishler and Rose 2007; Pop-Eleches and Tucker 2017; Bacovsky and Fitzgerald 2023). If new citizen cohorts progressively come of age lacking positive views of democracy, we can expect the political implications over time to be dramatic and deleterious. We therefore approach our study with an eye on the development of democratic support among young people. Tracing responses to Swedish panel survey questions on views of democracy from year to year, we provide an uncommon insight into what shapes the democratic values of young people in a well-established democratic system. Our analysis is informed by existing theories of democratic support as well as new theoretical tools developed by scholarship on the political implications of belonging. This burgeoning belonging literature prompts us to ask whether and how feelings of belonging relate systematically to basic democratic values.

To summarize our findings: stronger and broader feelings of belonging promote principled support for democracy. Attitudinal indices of belonging to national and sub-national communities consistently correlate with democratic support. Feeling “at home” in Sweden, in one’s municipality, and in one’s neighborhood, predict stronger democratic values among respondents over time. One striking aspect of our findings is that the effect of belonging is significantly more powerful than factors theorized to matter for democratic values, such as participation in civil society and socio-economic status (SES). Moreover, we see that belonging is a significant predictor of democratic principles when we control for satisfaction with the way democracy works.

We also consider: what are the boundaries of belonging as it relates to support for democracy? Does any kind of belonging matter, or is the impactful type of belonging associated with social communities? In other words, does a blunt sense of belonging promote democratic values or does it depend on the reference group to which one feels they belong? Our results signal that specific kinds of belonging matter for democratic support, and this helps us to refine our account of the implications of belonging. Notably, we find that feeling at home in one’s family is not a significant predictor of support for democracy. Thus, what we identify in our analysis is a positive connection between perceived belonging in a place-based social community and pro-democratic orientations. As such, we provide a nuanced view of how democracy may be supported or undermined through social inclusion or exclusion.

Our keen interest in exploring support for democracy prompts us to ask some follow-up questions based on our findings. To learn who feels that they belong in the first place, we explore the correlates of belonging for our sample. We find that individuals of higher SES, those born in Sweden, and those who are most trustful of others feel the most at home in their country, municipality, and neighborhood, leaving the relatively poor, foreign-born, and distrustful to feel less at home.Footnote 3

2 Theoretical Framework

2.1 Existing Literature on Support for Democracy

One explanation for democratic support (or the lack thereof) among the mass public hinges on assessments of government performance. The logic is that people will be in favor of democracy over alternative regime types as long as they perceive democracy to be delivering for the citizens. Evidence of this evaluative mechanism comes in the form of matching objective conditions at the national level to aggregate levels of support for democracy. For instance, objective measures of quality of government are found to underpin aggregate-level support for democracy (Boräng et al. 2017). Other work tells a similar story, leveraging hierarchical models to nest individuals within their country contexts. Such work shows that macro-economic conditions and perceptions of government performance shape micro-level support for democracy (Magalhães 2014; Evans and Whitefield 1995).

Yet these analyses obscure key dimensions of support for democracy that are distinct from assessments of how well democratic governments are doing their jobs. Although the more transactional version of democratic support may wax and wane over time in response to (perceived) government performance, a more stable set of values underpin “diffuse” or “principled” support for democracy, also termed a “general attachment” to democracy (see Easton 1975; Bratton and Mattes 2001; Klingemann 2018). It is this intrinsic support for democracy—independent of more flexible evaluations of government effectiveness—that interests us here. This is because it serves as a crucial reservoir of support for democratic regimes even during difficult eras. Without it, the fate of democratic institutions depends on societal conditions such as economic strain or inequality (Pennings 2017; Andersen 2012). For democracy to withstand hard times, democratic values among the population are essential (Inglehart 2003).

Existing studies tell us a great deal about the sources of this principled support for democracy. Social capital theories posit that membership in various kinds of organizations and groups enhances pro-democratic values (Putnam 1993; Meyer et al. 2008). Alternatively, resource-oriented accounts theorize that those who are educated, skilled, and otherwise elevated in SES are also more appreciative of democracy’s basic tenets (Lipset 1959; Cho 2014) and are better able to engage effectively in democratic life (Verba et al. 1995; Rosenstone and Hansen 1993).

Still other works point to certain attitudes that underpin individuals’ intrinsic support for democracy. Trust in others (also known as generalized or social trust) is consistently identified as a central orientation characterizing those who feel very supportive of democracy (Dowley and Silver 2002; Cho 2014). Existing evidence demonstrates that internal efficacy (the notion that one has the ability to participate in politics and have an influence) is important for the development of democratic norms and values. For instance, Finkel (1985, 1987) connects internal efficacy to positive system orientations and broad political support in democracies.

Finally, studies of support for far-right extremist parties, which challenge key democratic principles such as tolerance and equality, approach the matter from a different angle. Perhaps most notably, they find that individuals who are intolerant of immigrants and immigration are overwhelmingly represented among voters of radical right parties (Rydgren 2008), which threaten key dimensions of democratic governance. Societal-level research offers complementary observations. For example, Freeman (1997) observes that immigration has led to declining public faith in and support for their governments since the 1960s,Footnote 4 making citizens more receptive to the messaging of far-right extremists (see also Shehaj et al. 2019).

The cumulative knowledge represented above relates to citizens across age groups. What do we know about the ways in which young people develop their orientations to democracy? Although studies of young citizens parallel those of their elders in society relating to the main sets of theories outlined above (see Šerek and Lomičová 2020), scholarship on adolescents also draws attention to the ways in which democratic support is shaped by experiences that are intimate, routine, and close to home. For instance, research shows that family lives, emotions, and experiences play a significant role in shaping young citizens’ orientations to democracy (Miklikowska and Hurme 2011). Similarly, personal experiences in defined social contexts such as the school shape young adolescents’ understanding of key democratic principles and promote support for equal rights (Torney-Purta et al. 2008). The climate of social environments such as classrooms and interpersonal networks shape democratic values among young people as well (Reichert et al. 2018).

This collection of studies focusing on adolescents and young adults can help to inform our general theorizing and research on support for democracy. They signal that we can enhance our understanding of how support for democracy develops and lasts through exploration of people’s feelings and foundational orientations to the world around them. Existing theories of social capital, SES, and political attitudes do not effectively capture this socio-psychological dimension of life that animates basic democratic inclinations. Our focus on feelings of belonging to place-based social communities, therefore, stands to enrich our understanding of how pro-democratic orientations develop.

2.2 Belonging and Political Behavior

Feeling like one belongs is a foundational concept for human motivation and behavior. Research in psychology and sociology provide the theoretical foundations for its study. A long-established fact of human psychology is that belonging is a central aspect of an individual’s sense of self (Maslow 1943; Baumeister and Leary 1995; Osterman 2000; Ryan and Deci 2000). As sociologist Yuval-Davis explains, “Belonging is about emotional attachment, about feeling ‘at home’ and … about feeling ‘safe’” (Yuval-Davis 2006, p. 197). Objects of belonging evoke emotional attachments, feelings of closeness, and relate to specific social and cultural spaces in society. Duyvendak (2011) refers to “home” as a “fixed place where people belong, a place which is theirs” (Duyvendak 2011, p. 2). He evokes concepts of nation, community, and the household in his treatment of the subject.

Both Yuval-Davis and Duyvendak identify the close connection between the power relations associated with belonging and politics. Further connecting the sociology of belonging to the political realm, Kuurne and Gómez (2019) write, “the question of belonging has been viewed as twofold, first concerning how persons ‘feel at home’ (or not) in a certain setting, and second, how the politics of belonging produce patterns of inclusion and exclusion” (Kuurne and Gómez 2019, p. 214). With regard to politics, the very notion of citizenship, itself, can be conceptualized as a form of belonging in the psychological sense (Shotter 1993).

Despite these observations, the nexus between feelings of belonging and democratic support has not been sufficiently explored in empirical research. In the absence of dedicated literature on this relationship, we assemble a broad set of publications from a range of disciplines. Taken as a whole, we see that feeling part of a particular place-based community can have potent effects on citizens’ political orientations and political behavior.Footnote 5 An intriguing aspect of existing findings about belonging is that it seems to be linked to divergent outcomes in relation to democratic support.

On the one hand, some research on the political behavioral implications of belonging point toward negative effects on support for democracy. For instance, Cramer (2016) explores the aspect of belonging known as place attachment to illuminate how “rural consciousness” can animate support for right-wing populist politics in the USA. In a similar vein, Fitzgerald (2018) finds that a form of belonging rooted in place attachment drives support for radical-right parties in Europe. These findings align with research demonstrating the ways in which feelings of national belonging and identity motivate support for nondemocratic ideas and programs, notably on the far right (Mummendey et al. 2001; Billiet and De Witte 1995; Skenderovic 2009). Nationalism’s tendency to draw striking contrast between those who belong and those who do not is well known to promote anti-democratic extremism (Ignatief 2001; Brubaker and Laitin 1998; Weinberg and Assoudeh 2018).

On the other hand, research from a range of fields suggests that feelings of belonging promote and reinforce democratic values. Though few have connected belonging directly to support for democracy, taking stock of published work reveals that 1) belonging has significant, positive psychological effects, and 2) healthy psychological traits are important prerequisites for holding democratic values.

Maslow (1943) asserts that the need to belong is foundational for people; indeed, he argues that it is the third most basic need following physiological and safety requirements. Hogg and Abrams (1993) and Hogg (2009) theorize that a sense of belonging staves off the psychological strains of uncertainty. Further psychological research demonstrates that feelings of belonging promote self-esteem (Baumeister and Leary 1995; Begen and Turner-Cobb 2015), which in turn promotes support for democratic values and diminishes authoritarian inclinations (Shaffer and Hastings 2004). Similarly, perceived social connectedness enhances feelings of social solidarity, reduces anxiety, and engenders positive self-image (Lee and Robbins 1998), factors that in turn promote pro-democratic orientations (Lane 1962; Sullivan et al. 1981; Marchlewska et al. 2019).

Experts in natural resource psychology find that a sense of belonging to a particular social space engenders goodwill toward others and promotes civic action (Payton et al. 2005). Community psychologists report that feelings of belonging in a community promote democratic values (Anderson 2010) and that feeling part of a collectivity encourages civic participation (Lewicka 2005). These are important findings, as well, given the centrality of political engagement for democratic support (Sullivan and Transue 1999).

These observations align with a classic literature in political psychology that identifies psychological security as a requisite for democratic support (Lipset 1959; Campbell et al. 1980). Sniderman (1975) demonstrates that tolerance for diverse viewpoints and openness to compromise—key facets of democratic life—require psychological security. They also complement some research on the predictors of radical right support such as Sachweh (2020), who finds that subjective social marginalization predicts support for the extreme right AfD in Germany. Our chief expectation, therefore, is that feelings of belonging are on balance positive predictors of support for democracy.

3 Data

Our analysis uses data collected in Örebro, Sweden, a medium-sized city of 140,000 inhabitants at the time of the study (2010–2015). The city is close to the national average on factors such as population density, income level, ethnic diversity, and unemployment (Statistics Sweden 2010). Sweden is a useful case to analyze mainly because it provides analytical leverage: Sweden consistently tops lists of democratic quality (Freedom House 2015) and public support for democracy (Norris 1999). As such, we can to some extent set aside theories linking mass support for democracy with (objective) assessments about how well democratic governments are working.

The Swedish Political Socialization Panel (SPSP) survey investigates young people’s political development over time as part of the Youth and Society (YeS) study at Örebro University (Amnå et al. 2009). Researchers conducted annual assessments in ten middle schools and three high schools in the city.Footnote 6 The SPSP is the product of scholars representing the disciplines of political science, developmental psychology, and communications. Thousands of young people took part in the survey, and most did so repeatedly over time. Beyond its breadth in terms of participants and the nature of its valuable panel, the SPSP stands out for its theoretically derived and methodologically sophisticated item batteries that capture key concepts for behavioral analysis. The data therefore allow researchers to trace the development of political orientations over time during adolescence and young adulthood: crucial years during which patterns of citizenship and views about politics are formed (Sears and Brown 2013).

Five cohorts of respondents participated in multiple years from 2010 to 2015. We use data from cohorts 1 and 2 in our analysis because these are the (only) individuals who answered survey questions in consecutive years.Footnote 7 This allows us to trace development of key political orientations on an annual basis. For the sample we use here, the youngest are 13 years of age at the start, and the oldest are 19 at the end of our panel. A range of questions were asked of all respondents in each wave; other questions were asked of only some cohorts and in only certain years. The assembled data for our purposes are 5915 person-year observations from 2696 individuals.

The dependent variable in our study is principle-driven support for democracy. This is based on five questions with a single opening prompt. It reads: “There are different conceptions of what society should be like. What do you think yourself?

  • Women should have the same rights as men.

  • All citizens should be free to choose their leaders.

  • There are better ways to govern a country than through democracy.

  • The most knowledgeable in society should take over power if the government is incompetent.

  • Under some circumstances, a nondemocratic government is to be preferred.”Footnote 8

The response options to these items are: agree absolutely, agree, disagree, disagree absolutely. We re-code some of these items so that they all run in a pro-democratic direction and combine them into a support for democracy index with an alpha of 0.6. These five items also load onto a single factor with an eigenvalue of 1.1.Footnote 9 This index has a total of 20 possible values and is coded—as are all of our variables—to run from 0 at the minimum value to 1 at the maximum value. Please see Table 4 in the Appendix for descriptive information on each of the variables we use in our models.

The main independent variable of interest is belonging in a particular geographically oriented community. This is based on three questions with a single root prompt: “How well do the following statements describe you? I feel at home in …

  • Sweden as a country.

  • the municipality where I live.

  • the neighborhood where I live.

  • my family or with relatives.”

The response options for these prompts are: doesn’t apply at all, doesn’t apply so well, applies quite well, and applies perfectly.Footnote 10 We combine the first three of these items (Sweden, municipality, and neighborhood) into an additive Belonging index, which has an alpha of 0.76 and that loads on a single factor with an eigenvalue of 1.46. Feeling “at home” in a specific place measures a sense of belonging and place-based sense of community (see Yuval-Davis 2006; Anderson 2010; Duyvendak 2011).Footnote 11 We use the fourth item, on feeling at home in the family, separately to better distinguish between the social, public dimensions of belonging and the private, personal dimensions with respect to their implications for democratic values.Footnote 12

We also include a measure of organizational memberships to set these kinds of ties and experiences apart from the psychological dimension of belonging represented by the belonging index. Respondents are asked: “Are you a member of an association/associations?” This is followed by a list of options from which to choose: sports association, religious association, outdoor recreational association (field biologists, scouts, or the like), immigrants’ association, political association, association for peace or human rights (the Peace Movement, Amnesty International, or the like), cultural association (theater group, role play, dance, music group, art, playing instruments, or the like), hobby association (photography, carpentry, sewing, stamps, chess, or the like), environmental association, or other (which?). We combine binary responses on each of these into a count variable from zero memberships to five or more. Associational memberships are a common measure of social capital (Putnam 1993).

Socio-economic status (SES) is measured in the SPSP in a somewhat unconventional way given that many of the respondents are quite young. The survey asks three questions:

  • “If you want things that cost a lot of money (e.g., a computer, skateboard, cell phone), can your parents afford to buy them?” Response options are: absolutely, not, probably not, yes maybe, yes probably, yes absolutely.

  • “What are your family finances like?” Response options are: my parents always complain that they don’t have enough money, it happens often that my parents complain that they are short of money, my parents seldom complain about being short of money, my parents never complain about being short of money.

  • “If you compare with others in your class, do you have more or less money to buy things?” Response options are: I have much less money than others in my class, I have a bit less money than others in my class, I have the same amount of money as others in my class, I have a bit more money than others in my class, I have a lot more money than others in my class.

These items correlate with an alpha of 0.75. The load onto a single factor with an eigenvalue of 1.35. We add them together to create our SES index, and as with all other variables we code the index to run from a value of 0 at the minimum to 1 at the maximum. Previous studies show that SES (both in absolute and relative terms) can predict support for democracy (Lipset 1959; Ceka and Magalhães 2020).

We also include a measure of the more evaluative dimension of support for democracy as a control to better test theories about diffuse or principled support for democracy. The SPSP asks, “On the whole, how satisfied are you with the way democracy works in Sweden?” Response options are very satisfied, quite satisfied, not especially satisfied, and not satisfied at all. The SPSP survey designers followed Norris (1999) and Linde and Ekman (2003) in creating this measure of instrumental support for democracy as it operates in one’s own country. We call this variable system satisfaction.

We include additional opinion items found to predict levels of support for democracy in our study. We measure trust in others using two survey items with the opening prompt: “If you think about people in general, how far do you agree with the following statements?

  • Most people can be trusted.

  • Most people are fair and don’t try to exploit you.”

Response options for both are: doesn’t apply at all, doesn’t apply so well, kind of applies, applies quite well, and applies perfectly. We combine these items into a summary trust index, which has an alpha of 0.85. These two items load onto a single factor with an eigenvalue of 1.29. See Flanagan et al. (2010) on measuring trust.

Efficacy is measured with a set of items that begin with: “If I really tried I could …

  • Write a letter to a newspaper or web editor about problems in my hometown or society in general.

  • Actively contribute to the work of organizations trying to solve problems in society.

  • Be a leader of a group working with societal issues.

  • Help to organize a political protest.

  • Be an active member of a political organization.

  • Discuss politics with people more experienced than me.

  • Take part in a demonstration in my hometown.

  • Take on responsibility in a political party.

  • Convince others to sign petitions concerned with political or societal issues.

  • Take part in a demonstration in my school even if my friends or parents are against it.

  • Paint graffiti or write political messages in public places to demonstrate my view.”

Response options are: I could definitely not manage that, I could probably not manage that, I could probably manage that, and I could definitely manage that. These 11 items correlate with an alpha of 0.93; the load onto a single factor with an eigenvalue of 6.25. We add the values of these constitutive variables together to create our efficacy index, coding it to run from 0 at the minimum to 1 at the maximum value.

The SPSP measures anti-immigrant sentiment with eight questions. The battery begins: “What are your views on people who have moved here from other countries?

  • Our culture gets enriched when people from other countries move here.

  • It happens only too often that immigrants have customs and traditions that do not fit into Swedish society.

  • In the future, Sweden will be a country where there are exciting encounters between people from different parts of the world.

  • Immigrants often come here just to take advantage of the welfare in Sweden.

  • That people move to Sweden is good for the Swedish economy.

  • Immigrants often take jobs away from people who are born in Sweden.

  • We should welcome people who have fled from problems in their own countries.

  • Immigrants should have the same rights as people born in Sweden.”

The response options are: does not apply at all, does not apply so well, applies quite well, or applies very well. The resultant index (which reverse-codes the positively worded items above), Anti-immigrant, has an alpha of 0.77. Factor analysis loads all eight items onto a factor with an eigenvalue of 2.62. The resultant index is a summary variable of the values of the variable itemized above.

An age variable that ranges from 14 to 18 years old (approximately)Footnote 13 and a dichotomous variable denoting whether the respondent states that she is femaleFootnote 14 are included as covariates. All of our variables are coded such that the lowest value is 0 and the highest value is 1 for ease of statistical interpretation. Models are time-series, fixed-effects regression models with robust standard errors to make the most of the panel nature of the data. Coefficients are interpreted as within-individual differences (between-individual difference R2s are also presented). Models specify robust standard errors and include a dummy variable for each wave to account for time-specific factors associated with the broader socio-political and survey contexts.

4 Results

How does the place-based belonging index relate to the support for democracy index among young Swedes? Table 1 presents the results of three fixed-effects regression models. Model A is a simple bivariate regression that focuses on the relationship between belonging support for democracy. It shows that belonging is a positive and significant predictor of support for democracy. Shifting belonging from its minimum to its maximum value is associated with a 7 percentage point increase in support for democracy over the previous year. Models B and C introduce covariates to establish the effect of belonging more fully. We see that the belonging index is robust to these controls. Yet these independent variables are also illustrative in their own right. Membership in a number of different organizations is positively associated with support for democracy, but it is not statistically significant, challenging social capital theories. This tells us that belonging more effectively accounts for variation in democratic support than does a well-recognized predictor of democratic values. Socio-economic status (SES) is not related to democratic support per these models, in contrast with resource-based theories of democratic values. Respondents in their later teens are no more supportive of democracy than younger respondents, though recall that the age range for these participants is from 14 to 18.Footnote 15 We are therefore unable to contribute to debates over the relative positivity levels toward democracy between younger and older citizens. Individuals who self-identify as women and those born in Sweden are not more supportive of democracy than others.

Table 1 Belonging Index and Support for Democracy (fixed-effect regression models)

Model C includes system satisfaction, which captures the evaluative dimension of support for democracy as it functions in Sweden (also termed “specific” support) (Easton 1975). Shifting system satisfaction from its minimum value to its maximum value is associated with a 6% increase in support for democracy. We would expect a strong correlation here, given that satisfaction with democracy and support for democracy are quite close conceptually. What one might not expect is that the coefficient for system satisfaction is similar in size to that of the coefficient of the belonging index. Furthermore, the fact that the belonging index maintains significance with system satisfaction in the model speaks to its value for explaining support for democracy in principle (representing the concept of diffuse support). The belonging index is also robust to inclusion of attitudes often found to promote democratic values, especially trust and efficacy, and to attitudes found to undermine democratic values, especially anti-immigrant attitudes. Anti-immigrant attitudes, we note, correlate significantly and negatively with support for democracy. It is the single most impactful variable in our models, signaling tension between intolerance for immigrants and democratic values.

Over time, the place-based belonging index exerts an independent effect on support for democracy. This is a particularly stringent test of the role belonging plays in support for democracy over time.Footnote 16 In substantive terms, belonging outperforms nearly every other variable in the model. The only exceptions are attitudinal variables: system satisfaction and immigration attitudes. This underscores the importance of feelings of belonging for democratic values and ultimately democratic stability.

These results underscore the role of belonging in the development of democratic values, but they also obscure potentially telling insights by using the collapsed belonging index. We next disaggregate this index to evaluate whether all forms of belonging in the index—in Sweden, in the municipality, in the neighborhood—shape democratic support. Table 2 presents the full models for each of these index items. We see that they are all predictive of democratic support—regardless of the many control variables in these models. We also see that perceptions of belonging in the municipality are slightly more impactful than those of the other two measures. In bivariate versions of these models (see Table 5 in the Appendix), this finding about the municipality is corroborated. More broadly, we learn from these models that feeling at home in a range of collective places (from the more defined neighborhood to the expansive nation) promotes democratic orientations and diminishes openness to authoritarianism.

Table 2 Belonging disaggregated and support for democracy (fixed-effects regression models)

To this point we have focused on the concept of belonging as a feeling of closeness to and safety in a range of specific place-based communities. And yet it is possible that belonging is a state of mind and that any kind of perceived belonging might be associated with support for democracy. If this were the case, then our narrative of the importance of belonging to a broad, social collective that is bounded in spatial terms would be imprecise. To gain better purchase on how, exactly, feeling at home in the country, municipality and neighborhood shapes democratic values, we leverage the fourth item in the belonging battery: feeling at home in one’s family.

If there is something specifically community-based about the role of our belonging index’s impact on democratic support, then feelings of belonging in the family should not have this same impact. A societally related mechanism would not be captured by feeling at home with relatives. Alternatively, if in our above models we are picking up on a more generic feeling of belonging that spans the public and private realms, then we might expect feeling at home in the family to have the same effects. We test this by running two additional models.

The first model in Table 3 evaluates the role that feeling that one belongs in the family plays in shaping support for democracy in parallel fashion to the models in Table 2. It shows that family belonging is not a statistically significant predictor of principled support for democracy. The second model includes our standard belonging index (in Sweden, municipality and neighborhood) as a control. The two coefficients tell a nuanced story about belonging. The forms of belonging that are most relevant for democratic values are social and denote particular geographically defined communities.

Table 3 Belonging in family and support for democracy (fixed-effects regression models)

As with most studies, ours answers some questions and raises others. With new light shed on belonging as a source of pro-democratic support over time, more information on who these high- and low-belongers are is called for. We approach this step in our inquiry in simple, descriptive terms, estimating pairwise correlations of a range of variables from our study with our three-item belonging index (in Sweden, in the municipality, in the neighborhood). We present the resultant patterns as Fig. 1 in the Appendix. Who feels that they belong? Individuals who are members of a range of organizations and who are highly trusting are most likely to feel that they belong. Similarly, those at relatively high levels of SES and those born in Sweden feel more “at home” in the geo-social communities we study.

In contrast, our oldest respondents feel like they belong less than our youngest respondents. Because the age range is approximately 14–18 years in the sample, this implies a loss of belonging as an individual enters the young adult years. Identifying as female, efficacy, and anti-immigrant views are not associated with belonging in pairwise fashion. If we reverse our thinking and consider the opposite side of belonging, this table tells a story of the relatively poor, the foreign-born, and the untrusting as feeling that they do not belong in social places. Importantly, we make no causal claims with these correlations, but in descriptive terms they can provide focus for how to think about belonging and perhaps how to go about encouraging it. With respect to these observations about feelings of belonging (or lack thereof) among less trusting and foreign-born respondents, we note the importance of patterns of residential diversity versus segregation between nationals and immigrants for understanding social integration more broadly (see Jünger and Schaeffer 2023, this issue).

5 Discussion

In this paper, we explore the link between feelings of belonging and principled support for democracy. The present analysis allows us to focus on a particularly important sub-population for present and future trends: adolescents. Furthermore, our focus on Sweden simplifies the analysis to a single socio-political context, in effect controlling for national-level factors such as government performance and other society-wide factors. By utilizing a study from a single city in Sweden, we narrow the number of potential confounding variables even further. Leveraging panel data provides granular and dynamic insights into the ways that support for democracy develops over time.

The main take-home message from our study is that belonging—in the form of feeling “at home”—is associated with greater support for democracy. Moreover, we supply evidence that belonging shapes orientations toward democracy over time. Feeling at home in one’s country, in one’s municipality, and in one’s neighborhood independently and collectively promote democratic values. We gain confidence in our interpretation that our findings shed light on the implications of belonging in geographically defined social communities—independent of more diffuse feelings of belonging—because feeling at home in the family yields no significant effects on democratic support.

Our study relates in key ways to this special issue’s broad themes on societal integration set forth in the introduction (Grunow et al. 2023, this issue). First, the outcome of interest, principled support for democracy, connects to the consensus-based dimension of integration. Societal agreement on foundational political orientations is essential for developing common purpose among democratic citizens. It is especially pressing in these times of democratic strain. Second, our focus on belonging is essential for understanding what it means for society when individuals feel “at home” in different geo-social spheres. In this case it means stronger support for democracy, which should help to buttress the essential institutions and norms in times of political distress. Third, our findings relate to young people, whose socialization into democratic norms and inclusion in social fabrics are crucial for a well-functioning social order. The patterns we identify for the young Swedes in our study can serve as a roadmap for promoting societal integration in a range of ways.

Generally speaking, the pro-democratic effects of this narrative of belonging augur well for democratic politics in the future. We make this claim based in part on research that points to aggregate-level strengthening over time of feelings of place-based belonging, attachment, and identity (see Fitzgerald 2018, Chap. 1 for a discussion). If indeed, the widespread trend that scholars assert is happening, then as feelings of belonging rise so will the deep appreciation for democracy among the mass public. To this rosy scenario we can add our evidence that young people exhibit this pattern: for young people in a relatively highly functioning democratic context, belonging promotes support for democracy. If support for democracy among new generations is of concern, then here we are able to provide insight into how promotion of democracy works for tomorrow’s leaders, voters, and activists.

Although these implications may seem quite positive on balance, the converse consideration of our findings merits careful attention: not feeling “at home” is associated with weaker commitments to democratic principles. Our study shows that those who are relatively poor, those who were not born in Sweden, and those to do not feel trusting toward others are also the ones who feel least “at home” in society. We also see that as young people age, they feel less and less “at home” in their country, municipality, and neighborhood. Keeping an eye on effects over time within generations is just as important as examining effects between generations when studying societal trends that matter for democratic governance.

Indeed, the ways in which citizens relate to and feel about their geo-social surroundings merit significantly more scholarly attention to better appreciate their complexity. This study takes a step toward enrichment of our collective understanding of support for democracy. Belonging entails a feeling of closeness and being at home in a particular social space. It represents a form of social inclusion that shapes how people see themselves in relation to the world around them. If feelings of belonging can be encouraged in a way that does not denigrate a “them” but instead enhances the positive, psychologically beneficial aspect of feeling part of a place-based “us,” then its implications for an inclusive, equitable democracy can be realized.