A sense of belonging is one of an individual's basic psychological needs, capturing experiences of acceptance, being connected with others, being supported and being safe (Maslow, 1943; Ryan and Deci, 2000; Wenger, 1998). In an educational context, sense of belonging is linked to students' identity development and engagement with study communities (Korhonen et al., 2019; Allen et al., 2018). Research has differentiated academic and social aspects related to sense of belonging to an educational institution (Ahn and Davis, 2020; Korhonen et al., 2019).

While inequalities in academic achievement have attracted wide scholarly attention, students' perceptions of schools or other educational institutions and their place in them, such as sense of belonging have been studied far less. Yet, in addition to educational achievement, students' sense of identification with the context in which learning takes place is essential to their overall success in various stages of their educational journey. Increasingly, the literature has demonstrated that students who feel a sense of belonging in an educational institution, such as school or university, are generally doing well academically, socially and emotionally (Ahn and Davis, 2020; Graham et al., 2022; Korhonen et al., 2019; Korpershoek et al., 2020).

It has been found that students' sense of belonging to an educational institution is associated with study progress and success, achievement, overall educational attainment and wellbeing related to studies (Finn and Zimmer, 2012; Graham et al., 2022; Pedler et al., 2022; Ulmanen et al., 2016; Korpershoek et al., 2020). The interrelation between scholastic achievement and students' wellbeing is reciprocal. In other words, stronger achievement boosts students' sense of belonging and likewise, the feeling of being in the right place in education also affects educational outcomes. Current research has found that having a perception of alienation in education may lead to challenges in academic success and wellbeing, and that this vicious circle may continue into subsequent stages of a student's educational pathway (Salmela-Aro et al., 2018). In contrast, positive experiences in education lead to stronger academic achievement and social acceptance and as a result, an increased sense of belonging in education (e.g., Ahn and Davis, 2020; Wentzel, 1998). An important finding from the life course perspective is that those students with a weak sense of belonging in education are the ones who are disinterested in pursuing further education, an association that exists even after controlling for other demographic factors (OECD, 2019, 2017). For these reasons, it is important that educational research focuses not only on academic achievement, but also on the sense of belonging in education.

Previous research has highlighted that the sense of belonging is a complex, fluid and dynamic construct (Graham et al., 2022; Dias-Broens et al., 2024; Allen et al., 2018). Sense of belonging is constructed through various experiences in education and multiple interactions with other students and teachers over a long period of time, so it is regulated by various factors (Korhonen et al., 2019; Allen et al., 2018). The role of students' close, friendly relationships with teachers has been highlighted by Chiu et al. (2016) and Ulmanen et al. (2016). In relatively recent studies, sense of belonging in education has been considered mainly from socio-ecological perspectives, i.e., emphasizing social processes and relations, close relationships and connectedness with immediate surroundings, rather than educational policies or structures of educational systems (Aldridge and McChesney, 2018; Allen et al., 2018; Brown and Shay, 2021; European Commission, 2021; Glazzard, 2019). Factors critical to students' sense of belonging include caring relationships with teachers and functioning relationships with peers, as well as pedagogies that enable active student participation, engagement in meaningful learning activities and collaborative learning with peers (Atkinson et al., 2019; European Commission, 2021; van Gijn-Grosvenor and Huisman, 2020; Weare and Nind, 2011). Both the school and classroom climates have been found to be important. Furthermore, students' surroundings (i.e., the geographical, environmental and cultural contexts of the higher education institution and their living spaces in the local area) and personal spaces (i.e., life satisfaction, attitudes to life and personal interests) are found to be important factors influencing students' sense of belonging in education (Ahn and Davis, 2020; Allen et al., 2021). Additionally, previous studies have noticed variation in the sense of belonging in education between students with different socio-demographic characteristics (OECD, 2019, 2017; Allen et al., 2021), although the research is far from being systematic or conclusive and there are contradictory findings (see Ahn and Davis, 2023; Korpershoek et al., 2020). Considerable cross-national differences have been detected in sense of belonging (OECD, 2017).

Overall, a magnitude of variables has been found to influence sense of belonging, and all of these factors, associations and differences are influenced by macro-level educational systems and educational policies, but research has yet to explore how such macro-level variables are associated with students' sense of belonging in education.

Educational inequalities refer to systematic variation in several aspects of educational attainment between individuals, based on a range of socio-demographic characteristics such as gender, ethnicity, age, disability and socioeconomic background (Codiroli Mcmaster and Cook, 2019; Gross et al., 2016; Nichols and Stahl, 2019). The educational opportunities of different social groups have been shown to be hampered by these educational inequalities (Codiroli Mcmaster and Cook, 2019; Gross et al., 2016; McCall, 2005). Educational inequalities have multiple consequences with regard to access to education, learning processes, educational outcomes and level of education completed (Gross et al., 2016; Jacobs, 1996). Previous research has shown that gender, migration status and family's socioeconomic status are the most prominent dimensions of educational inequalities at various educational levels (Bodovski et al., 2020; Marks, 2015); in the present study, we will also focus on these dimensions. While the majority of research has examined the effects of these dimensions of inequality separately (Zimmermann and Seiler, 2019), it has been found that they intersect, intertwine and cumulate. Such intersectionality leads to an intensification of educational disadvantages in certain groups (Codiroli Mcmaster and Cook, 2019; Gross et al., 2016; McCall, 2005). Therefore, it is not sufficient to focus only on individual dimensions of inequality when it is clear that socio-demographic characteristics operate simultaneously and jointly, accumulating over the course of a life. Consequently, inequalities in education should be viewed from an intersectional perspective, which requires broadening the view from individual dimensions of inequality to a complex and intertwined set of different characteristics and circumstances (Allen et al., 2021; Codiroli Mcmaster and Cook, 2019). Such extensive intersectional research combining several dimensions is relatively scarce (Graham et al., 2022; Strand, 2014). In the present study, we address the challenge of intersectional inequalities by creating intersectional groups that combine three major dimensions of inequality (see full details in Data and method section).

Answering complex questions such as variation in inequalities in sense of belonging in education requires an intersectional approach, recognizing that individuals can belong to multiple social groups and contexts simultaneously (Ahn and Davis, 2020; Graham et al., 2022; Rainey et al., 2018) and the intensity of their belonging to these various groups and contexts can vary (see Allen et al., 2021; Mullen and Yifang, 2024). However, previous research has mostly focused on single dimensions of inequality and individual-level sources of inequality.

Among individual-level sources, gender has been found to be one of the major determinants of sense of belonging in education. Research on primary school students is scarce, but findings indicate that girls feel a stronger sense of belonging in education (Sayer et al., 2013). More research exists on secondary school students, and this shows wide, cross-national variation; namely, that different genders have a stronger sense of belonging in different countries (OECD, 2017). In many European countries, boys tend to report a stronger sense of belonging than girls, for instance in Denmark, Finland, Ireland and Norway (OECD, 2017). In contrast, girls had a stronger sense of belonging in a smaller number of countries, for instance in Austria, the Netherlands, Spain and Turkey (Aliyev and Tunc, 2015; OECD, 2017). In tertiary education, the trend is that females experience a weaker sense of belonging than males (Rainey et al., 2018; Museus and Maramba, 2010). In intersection with ethnicity or race, the effect intensifies: females of color are even more likely to have a weak sense of belonging (Rainey et al., 2018).

A second important socio-demographic determinant of sense of belonging in education is migration status. Sense of belonging is shaped by students' ethnic and migrant background across the educational pathway (Graham et al., 2022). In many European countries, e.g., Austria, Portugal and Switzerland, native students report a stronger sense of belonging in secondary education than students with migrant backgrounds. However, in just as many other countries, e.g., Denmark, Germany and Norway, there are no differences between these groups; and in some, e.g., Australia, the situation is even reversed (OECD, 2017). In tertiary education, migrant students experience a weaker sense of belonging than native students (Stebleton et al., 2010). A rare finding on the intersection between socioeconomic and migrant status and gender has indicated that migrant boys with low socioeconomic status tend to be disengaged and burned out in education, especially in secondary education (Salmela-Aro et al., 2018). It is noteworthy, that a concept of migrant is ambiguous, and findings vary depending on the status (refugee or asylum seeker) or the generation (newly arrived or children of migrated parents) of the migrant. The influence of migrant status on students' sense of belonging in education may also be related to the fact that students with migrant backgrounds are often a minority in schools or classes. In contrast, a larger representation of students with migrant backgrounds mitigates the effect of migration status on sense of belonging: students with migrant backgrounds do not report a weaker sense of belonging when their ethnic group is relatively well represented in their class or school (Graham et al., 2022; Tellhed et al., 2017). In general, underrepresented groups (e.g., non-white, non-heterosexual, non-Christian, disabled) report a weaker sense of belonging than students representing the majority (Fan et al., 2021; Kleemola et al., 2025).

A third important socio-demographic determinant of sense of belonging is socioeconomic status. In particular, students with combined high socioeconomic status and high achievement feel a stronger sense of belonging in education than disadvantaged students (OECD, 2017). However, there is also contrary evidence on socioeconomic background having no association with the sense of belonging (Korpershoek et al., 2020). It appears that achievement mediates the effect of socioeconomic status on the sense of belonging, at least in primary and secondary education (Cueto et al., 2010). In a similar vein, in tertiary education, there is strong evidence of students feeling a stronger sense of belonging if they have a higher socioeconomic status, compared with students with low socioeconomic status (Duran et al., 2020; Jury et al., 2019; Ostrove and Long, 2007; Stebleton et al., 2014; Trawalter et al., 2021). A weak sense of belonging is particularly evident in students who are the first in their family to attend tertiary education (Bowman, 2010; Rubin, 2012; Stebleton et al., 2014). Thus, parents' educational level as an indicator of socioeconomic status is important in terms of the association between it and the students' sense of belonging. However, in tertiary education, the selectiveness of student admissions often decreases the diversity of the student population in favor of students from high socioeconomic backgrounds (Kleemola et al., 2022). Such homogenisation likely weakens the sense of belonging in diverse groups (see Fan et al., 2021). Not only the socioeconomic status of students' families is important for their sense of belonging; a similar interaction has been found between students' sense of belonging and the average socioeconomic status of their school's student population. Students in schools with a low average socioeconomic status have a weaker sense of belonging in comparison to their peers in more advantaged schools (OECD, 2019). However, cross-national differences have been detected in this association, with the largest gaps in Europe seen in Bulgaria, Hungary and Luxembourg (OECD, 2019). Consequently, individual-level determinants of inequality exist in their contexts and interact with contextual factors (Mullen and Yifang, 2024).

Taken together, the growing body of literature demonstrates that sense of belonging is a complex, dynamic and multi-dimensional construct that changes across time and contexts (Ahn and Davis, 2020; Graham et al., 2022). On micro-level, there exists a complex interplay between individual and contextual aspects and courses of development of sense of belonging in various time points. However, the findings concerning associations between students' sense of belonging in education and gender, migration status and socioeconomic status seem to vary depending on the educational and cultural context. The reasons behind this variation have not been pinpointed. Associations between individual-level characteristics and sense of belonging have been studied extensively and large cross-national differences have been detected (see OECD, 2019; OECD, 2017; Mullen and Yifang, 2024), but little research exists on contextual macro-level factors that might explain the differences and inequalities in students' sense of belonging. However, macro-level determinants should be considered in research, as they have been found to influence the sense of belonging in students (Chiu et al., 2016). In studies focusing on achievement—strongly correlated with sense of belonging—it has been found that early tracking in the educational system increases inequalities in achievement (Contini and Cugnata, 2020). In contrast, increased female access to education has been found to boost female achievement in education (Buchmann and DiPrete, 2006). Existing studies on the influence of different policies on students' sense of belonging in education indicate that associations may not be straightforward and intuitive. Contradictory findings show that increasing inclusive policies may both support and decrease the sense of belonging in intended groups (Ham et al., 2017; Högberg and Lindgren, 2023). Counterintuitive is the finding that gender equality in society may actually increase wellbeing gaps to the benefit of males (Guo et al., 2022). As research is so scarce, it is still unclear whether the cross-national variation in students' sense of belonging in schools is somehow related to the institutional characteristics of different countries' education systems; for instance, with regard to the organization of the education system, female access to education and the way in which countries treat their ethnic minority population.

An additional challenge in the existing research on sense of belonging is, as stated previously, the lack of attention given to the intersectionality of inequalities. Thus, missing from the current literature is the use of an intersectional approach to give systematic attention to students' sense of belonging in different stages of the educational pathway.

The aim of the present study was to explore inequalities in students' sense of belonging in education across educational levels and contexts using an intersectional approach to inequalities. Example countries representing different educational contexts were used to demonstrate variation in the inequalities. Combining large datasets allowed us to investigate this complex interaction across all educational levels, from primary to secondary and tertiary education.

Our research questions were:

Based on previous research, we expected to find that male gender (Museus and Maramba, 2010; OECD, 2017; Rainey et al., 2018), non-migrant status (OECD, 2017; Stebleton et al., 2010) and high socioeconomic status (Duran et al., 2020; Jury et al., 2019; OECD, 2017; Ostrove and Long, 2007; Stebleton et al., 2014; Trawalter et al., 2021) were associated with a stronger sense of belonging in education across educational levels. However, as intersectional research on these associations is almost non-existent, we expected that findings may also be surprising, especially with regard to the intersections of these dimensions.

As there did not exist previous empirical research on the association between macro-level indicators of educational contexts and sense of belonging, we could not base our hypotheses on other studies. We therefore pursue an explorative strategy, investigating in detail example countries that represent different educational contexts. We chose these example countries by using macro-level correlates of the research on general educational inequalities. Accordingly, we expected that countries with favorable policies such as low tracking and high inclusiveness toward migrants would show as smaller inequalities in the sense of belonging. We considered it reasonable to expect to see less inequalities in sense of belonging where women are better represented in higher levels of education.

Secondary data from large, cross-national assessments were used in the study. Countries in the dataset covered European, North American and Oceanian countries (see Table 1). For the primary level, datasets from Trends in International Mathematics and Science Study (TIMSS) and the Progress in International Reading Literacy Study (PIRLS) were used (IEA., 2019, 2015; IEA, 2006, 2001; Kelly et al., 2020; Kennedy et al., 2007; Mullis et al., 2016, 2006, 2002). The datasets from TIMSS 2015 and 2019 and PIRLS 2001 and 2006 were used. For the secondary level, datasets from the Programme for International Assessment (PISA; OECD, 2000, 2003, 2006, 2009, 2012, 2015, 2017, 2018) were used. The datasets from 2000, 2003, 2012, 2015, and 2018 were used. For tertiary level education, Eurostudent was used (Cuppen et al., 2021). The dataset from Eurostudent VII (2019) was used. Detailed information about each dataset is available through above references. Some countries were dropped from the final compiled dataset due to lack of relevant information, namely socio-demographic information or the item used for measuring sense of belonging in education. Altogether 39 countries (as well as some of regions in UK) were included in the final dataset (see Table 1). Only seven of these countries had a dataset that covered all educational levels.

Macro-level indicators were used to reflect educational contexts and to pick example countries that represent different contexts. For this, data from UNESCO, the Migrant Integration and Policy Index (MIPEX) and the Educational System Database were used (see detailed description below).

Ordinary least squares (OLS) regression analysis was conducted for each country in each dataset with sense of belonging as a dependent variable and intersectional groups as independent variables, the reference group being “native males with high socioeconomic status” (MHN). Appropriate weighting variables were used in each dataset. Control variables for the language spoken at home, the degree of urbanization and survey years (see above) were used in the analyses when applicable. The resulting b coefficients reflected gaps in the sense of belonging between the reference group and the other seven intersectional groups. Average gaps were used to represent the entire dataset, and country-specific gaps to represent example countries.

Gaps in the sense of belonging between intersectional groups and the reference group are presented in Figure 1.

In primary education, all females, regardless of their migration or socioeconomic status, scored higher in their sense of belonging than the reference group of native males with high socioeconomic backgrounds (MHN). However, the difference was significant only in the case of native females with high socioeconomic backgrounds (FHN), who had by far the strongest sense of belonging of all the groups. Migrant females with high and low socioeconomic backgrounds (FHNN and FLNN) score slightly higher than the reference group in their sense of belonging, but the difference was not significant. In contrast, migrant males with high and low socioeconomic status (MHNN and MLNN) experienced the weakest sense of belonging in school, with a significant difference between them and the reference group. Native males with low socioeconomic status (MLN) were in between the reference group and the migrant males in their sense of belonging. In summary, there was a pattern of females having a stronger sense of belonging than males. Furthermore, having a migrant background was a disadvantage within the groups of females and males. Within the groups of natives, higher socioeconomic status indicated a stronger sense of belonging in education.

This pattern underwent clear changes in secondary and tertiary education. One pattern that remained constant is the top position of native females with high socioeconomic status (FHN) that extended through all stages of the educational pathway. However, in contrast to primary education, high socioeconomic status was the dividing factor in secondary and tertiary education. In secondary education, the group of native females with high socioeconomic status (FHN) was the only intersectional group that has significantly stronger sense of belonging than the reference group (MHN). The gaps between the groups were not very large overall in secondary education. Migrant females with high socioeconomic status (FHNN) were very close to the reference group, as were migrant males with high socioeconomic status (MHNN), but the difference was significant only for the latter group. Considerably larger was the gap between the reference group and migrant and native males with low socioeconomic status (MLNN and MLN) and migrant and native females with low socioeconomic status (FLNN and FLN). However, for female migrants with low socioeconomic status the difference was not significant. In tertiary education, the gaps between the intersectional groups grew notably larger compared with secondary education, particularly between the groups with low socioeconomic status. However, none of the differences were statistically significant. All intersectional groups with high socioeconomic status (MHNN, FHN and FHNN) were above the reference group (MHN) in their sense of belonging. Migrant and native males with low socioeconomic status (MLNN and MLN) were quite far below the reference group in their sense of belonging, but even lower were the respective groups for females (FLNN and FLN). Migrant females with low socioeconomic status (FLNN) had by far the weakest sense of belonging in tertiary education.

Next, we examined the gaps in sense of belonging in our example countries in detail (see rationale in choosing the countries in chapter 2.2.4). The country-specific gaps are presented in Table 3.

In Finland, the gaps in the sense of belonging were mostly small and/or not significant across the intersectional groups and across educational path. An exception to this was the native females (FHN and FLN) in tertiary education, with a stronger sense of belonging than the reference group (MHN). Compared with the average gaps in entire data, a stronger sense of belonging in female groups in primary education could not be detected in Finland, in fact, the direction of the gaps in female groups was negative, although findings were not significant. Likewise, neither in secondary nor in tertiary education, could a pattern with socioeconomic status be detected that appeared as dividing factor in the entire data. In contrast, the stronger sense of belonging in tertiary education in native females that was evident in the Finnish data, was not detected in the entire data.

In Hungary, too, the gaps in the sense of belonging were mostly small and/or not significant across the intersectional groups and across educational path. An interesting exception to this is the migrant females with low socioeconomic status (FLNN) in tertiary education with considerably stronger sense of belonging that the reference group. Compared with the average gaps in entire data, the gaps in the sense of belonging in Hungary in primary and secondary education were roughly aligned with the sizes and directions of the gaps in the entire data, although they are smaller and more often not significant. However, interestingly in tertiary education, most intersectional groups reported stronger sense of belonging compared with the reference group although the gaps were in large part not significant.

In Ireland, more significant gaps in the sense of belonging could be detected compared with the first two example countries. This was especially true in the primary education, whereas the gaps faded in secondary and tertiary education. Compared with the average gaps in entire data, the direction of the gaps was aligned. However, in the primary education, the gaps appear larger, with male groups reporting considerably weaker and female groups stronger sense of belonging than the reference group.