European identity among ethnic majority and ethnic minority students: understanding the role of the school curriculum Open Access

Schools are an important setting wherein different identity dimensions are made available to youth. This paper argues that European identity can be a relevant unifying collective identity in ethnically diverse contexts. We study whether (or not) students who are confronted with a European and/or a multicultural dimension in their school curriculum, identify more strongly with a European identity. More specifically, we analyse whether the importance of these dimensions for European identification differs between a sample of ethnic majority and ethnic minority students. The results show that both the European and the multicultural curriculum predicted a stronger European identity for both groups. The effects of both dimensions on European identity were similar for ethnic majority students, while the multicultural dimension had significantly stronger effects on European identity for ethnic minority students. The paper thus highlights how curricula contribute to changing patterns of identification in a diversifying society.

Over the past decades, European societies have been characterised by increasing ethno-cultural diversification. This diversification has led to broad discussions on social cohesion and the meaning and position of national identity (Bauböck 2018). In response to these social changes, policy makers across European countries have turned to their public education system, stressing its role in fostering a shared national identity, mainly through its national curriculum (Doppen 2010). Yet, while national curricula traditionally focused on developing identification with ‘the nation’, most school curricula in Europe have also gradually incorporated teaching about cultural and national diversity and human rights, which can further promote identification with European and global identities (e.g. Ramirez and Moon 2013; Schissler and Soysal 2005). Students are taught not only to become good national citizens, but also to respect diversity and uphold ‘European’ values and ideals (Faas 2010). In this paper, we aim to deepen the understanding of the role of the curriculum in the construction of a European identity. More specifically, we study whether or not a European and/or a multicultural dimension in the school curriculum leads students to identify more with European identity.

Traditionally, state education plays a key role in national identity formation processes, and policy makers expect schools to socialise students into becoming ‘good’ citizens primarily able to navigate and identify with national societies (Green 2013; Youdell 2010). Moreover, in super-diverse contexts with highly diverse populations, dominant understandings of national identities often presuppose a specific ethnic national identity, which makes it difficult to include all citizens in such a shared national imagination (Alba and Foner 2015). Indeed, recent research shows that in a European context, national identities seem to be much less attractive to ethnic minority and migrant students than they are for native ethnic majority students (Fleischmann and Phalet 2018).

Interestingly, slightly different findings have been reported with regard to European identification: ethnic minority youth generally identify more strongly with European identity than with the national identity of the country in which they are living, and ethnic minority and ethnic majority students are quite alike in their strength of identification with Europe (Agirdag, Phalet et al.2016; Clycq et al.2021; Teney et al.2016). In this paper, we aim to explore these findings further and try to grasp some underlying mechanisms. More in particular we study the role of various types of the school curriculum in these processes. We focus on the curriculum as this is an important way of making knowledge, repertoires and identities available to students. Given that ethnic minority students seem more attracted to European identity than to national identity, we elaborate on findings from previous research suggesting that the imagination of European identity as a rather inclusive and multicultural identity might play an important role (Agirdag, Phalet et al.2016; Clycq et al.2021; Keating 2016). Therefore we study the impact of a multicultural and of a European dimension in the curriculum on the European identity of students.

We have carried out our study in urban schools in the highly diverse context of the city of Antwerp (Belgium). Such schools are particularly interesting settings for such a study, as identity and diversity-related issues are part of everyday school life, for school staff as well as students. At the same time, research shows that even in these schools there is considerable variation in the ways these topics are included (or not) in the curriculum and in the classroom practices (Agirdag Merry et al.2016). There are also practical reasons for focusing on urban schools, as we were able to recruit a sufficient number of pupils with and without a migration background from the same schools, enabling us to study whether they perceive the curriculum differently.

Against this background, we aim to investigate whether (or not) students who are confronted with a European and/or a multicultural dimension in their school curriculum (or not) identify more strongly with a European identity. Moreover, we analyse whether the importance of these dimensions for European identification differs between a sample of students with a Western European migration background, which we consider in this analysis as the ethnic majority, and a sample of students with a Turkish and Moroccan background, which we consider as the ethnic minority.¹ The following research questions will be addressed: (1) To what extent do Western European origin students and Turkish and Moroccan origin students identify with a European identity, (2) What is the effect of a European dimension, on the one hand, and a multicultural dimension in the curriculum on the other hand, on the European identification of (a) Western European students and (b) Turkish and Moroccan students?

How and why do school curricula matter for students’ identification? And what do we understand by a European identity? There is extensive literature on the relationship between social and personal identity (e.g. Jenkins 2014), but we limit our discussion here to the significance of collective identity categories (such as ‘being European’) for the social identity of students. While identity generally refers to the way in which individuals describe or understand who they are in relation to other people, social identity is mostly conceived as the description or definition of the self in terms of social category memberships, such as gender, ethnicity, nationality or class (e.g. Tajfel and Turner 1979). When social category membership also implies a reference to a shared feeling of ‘one-ness’ or ‘we-ness’, this is usually conceived of as collective identity (Snow 2001, 2213). While there is no standard definition of collective identity, it always implies a shared awareness of being members of a collective or a ‘we’, that is anchored in real or imagined shared attributes and experiences (Snow 2001). Identification with a specific group forms the link between an individual's social identity and the broader, collective identity (with which he or she identifies).

Yet, the type of collective identities with which students identify is a highly contextual affair. In alignment with a social constructivist approach, we argue that collective identity categories such as ‘male’, ‘Belgian’ or ‘European’ are not stable and objectively given, but are constantly and collectively created, reconstituted, or combined. Whether or not students will feel themselves to be members of a ‘European we’, for instance, will partly depend on the vocabularies, stories and images that students associate with ‘Europe’. In this article, we hence explore how school curricula influence students’ European identification. In line with our social constructivist approach, we acknowledge that circumstances matter: school contexts and the curriculum made available to students can have an impact on their identification, as well as other factors such as socialisation strategies in the family (Reay et al.2011).

Due to continuing ethno-cultural diversification in most (European) societies, a new focus has emerged on the role of collective identities and shared feelings of belonging. In several countries, this has led to a stronger focus on ‘national identities’ (Clycq and Levrau 2017). Particularly in the last decade, loyalty to the ‘nation’ and the cohesion of society have become core objectives of government policies, and in particular of education policies (Alba and Foner 2014; Faas 2017; Verkuyten and Martinovic 2012; Colombo 2015). A shared national identity is promoted, yet often imagined in quite exclusive ways; in the UK, for instance, the focus on Fundamental British Values in the national curriculum is seen to prevent ‘radicalisation’ (Vincent 2019) . Such an approach seldom leaves room to include ethno-cultural diversity as part of this national identity, because it presupposes an assimilation of so-called problematic minorities into the national imagination. Several studies across Europe have indeed shown that ethnic minority youth identify much less with (sub)national identities than ethnic majority youth (Fleischmann and Phalet 2018; Jugert et al.2019). Since ethnic majorities identify quite strongly with the (sub)national identity, this reveals a gap, which leads us to ask what identities might be better suited to become a shared identity crossing ethnic, religious, and cultural lines in society.

Recent research has indicated that European identity is a salient identity category for individuals of various backgrounds, and can possibly function as such a shared collective identity (e.g. Agirdag et al. 2016; Clycq et al. 2021). This leads us to a further elaboration of what European identity might mean to young people. A European identity construct seems to be, in contrast to national identities, less ethno-culturally exclusive and more civically defined (Erisen 2017). Because one ‘only’ needs to be a legal citizen of a European country and there are no ethnic, religious, linguistic, or cultural traits that one is supposed to have or hold, it is, arguably, easier to identify oneself as a ‘European citizen’ than as a national citizen. Delanty (2005) states that ‘[to] be European is simply to recognize that one lives in a world that does not belong to a specific people’ (19), suggesting that by being European you cannot be excluded on the basis of your ethno-cultural heritage. Given the fact that Europe is by definition multilingual, multi-ethnic, multireligious, multinational, the elusiveness of European identity can be imagined in more inclusive terms (Alba and Foner 2015; Clycq 2021). Some political groups have tried to recast Europe in ethnic ways, i.e. a European identity based on white European ethnic roots and European Christian culture (e.g. Brubaker 2017). A person with a non-European ethnic identity – e.g. identifying as Moroccan, Nigerian, or Chinese – might find it difficult to identify with a European identity that is defined in such a way. Yet, we believe that, overall, the imagined European identity includes respect for cultural and religious diversity, and specifically in schools students will be taught to uphold the ‘European’ values of diversity and pluralism.

The relative attractiveness of a European identity for ethnic minorities is also corroborated by research. Agirdag et al. (2016b) examined the European identity and (sub)national identities of pupils aged between 10 and 12 years old in their last years of primary school in Flanders. They found that pupils of Turkish origin identified more strongly with Europe than with Belgium, while ‘native’ Flemish students identified more strongly with Belgium. Likewise, Clycq et al. (2021) showed similar results when comparing ‘native’ Flemish students and Moroccan origin students aged between 16 and 19 years old in Flanders. Clycq et al. (2021) found that Moroccan students identify more strongly with a European identity than with a Flemish identity, yet, for native Flemish students this was the other way around. However, in both studies, the ‘identity gap’ between students with a Moroccan or Turkish background and native Flemish students was much smaller with respect to the European identity than for the Flemish identity (Agirdag et al. 2016b; Clycq et al.2021). These results suggest that a European identity might indeed become such a shared identity (Alba and Foner 2015). Thus, a first step in the analysis in the current paper is to identify the strength of the European identity of both subgroups, namely Western European students versus Turkish and Moroccan students, and to see whether or not they differ significantly in strength. This leads us to the first hypothesis which will allow us to further explain the (potentially similar) strength of European identity in the two groups, studying the impact of the school curriculum.

There are various ways in which schools can influence identity formation processes of students; through everyday practices and routines (e.g. students’ relationships with adults and with peers, and how the school seeks to guide and shape these), but also explicitly through lessons and the curriculum (Blaskó et al.2019). Recent research into (sub)national identifications of young people acknowledges that (sub)national and supranational identities are not solely constructed in school through cognitive elements and civic knowledge, but also through sociocultural interactions that take place within the school (Solano-Campos 2015). In this article, we are specifically interested in whether a European identity is associated with a curriculum with a multicultural dimension, which acknowledges and values diversity and differences between young people, or rather with cognitive knowledge about Europe achieved through a European dimension in the curriculum. Although there are different ways to measure curriculum content, we focus on the students’ perception and the extent that they think their school adopts a multicultural dimension and/or a European dimension within their curriculum. In this way, we aim to consider the ways students regard, understand, and interpret the curriculum as they experience it in their school. Multiple studies have shown that schools often endorse more than one cultural diversity approach and teachers often have quite some ‘freedom’ to articulate and present these topics in their teaching (Civitillo et al.2017; Plaut et al.2018). Moreover, it is highly possible that schools may communicate that they endorse a multicultural curriculum yet implement rather assimilationist policies that maintain inequalities by excluding certain identities or topics in their everyday practices (Dixon et al., 2007). As these different cultural diversity approaches may have different effects on multiple social outcomes of young people with minority and majority backgrounds (Celeste et al.2019), it is important to also study the perceptions of the students themselves and recognise their interpretations of the curriculum. In the next paragraph, we elaborate on how a multicultural dimension and a European dimension of the curriculum are related with one's identity construct.

Making ‘national’ identities cognitively available in school is traditionally done through explicit history, language and cultural education and more recently through citizenship education. The national curriculum focused on developing identification with ‘the nation’ by equipping young people with conventional knowledge, values, and attitudes to become active democratic citizens (Green 2013). This was quite a successful strategy, as for many young people, ‘national identity’ is one of their most important identities. Although all countries in Europe have considerable autonomy in the field of education, European policy makers believed that a European identity could also be cultivated through education. Schissler and Soysal (2005) explain that in the years following World War II, national histories were reframed within a broader European context, which altered educational curricula. The integration of teaching about Europe and providing opportunities for students and teachers to be in contact with each other across Europe to strengthen a European identity, through programmes such as Erasmus, has been advocated by several EU institutions, such as the Council of Europe and the European Commission (Savvides and Faas 2016).

However, up until now and despite these proclaimed policy aims, concrete EU actions to strengthen European identification have been rather limited and predominantly focused on facilitating cross-country mobility or providing general knowledge about Europe and the EU (Bergbauer 2018). Furthermore, there is little research on the relationship between learning about the EU and European identification. Verhaegen and Hooghe (2015) analysed the data of the largest international and only dedicated study of civic and citizenship education, the International Civic and Citizenship Study (ICCS) 2009, and showed that knowledge about the EU had significant but limited effects on youngsters’ European identity. They argue that personal benefits that come with being a European citizen are a more important determinant for European identity. Furthermore, Ziemes et al. (2019) found that opportunities for learning about Europe predicted European identity in all countries, while Jugert et al. (2019) found that opportunities for learning about Europe, together with social-economic status and migrant background, were negatively associated with European identity at the classroom level. They qualified this finding by stating that qualitative research is needed to examine how school curricula including a European dimension affect representations of identity. One of the few studies that analysed whether the integration of a European dimension into the school curriculum can transform ‘exclusive-nationalistic’ educational approaches and practices into more inclusive ones, is that of Philippou (2007). She showed that the incorporation of a European dimension – defined as an inclusive identity respecting diversity – could be helpful to let students redefine their identity. It should be noted that different teaching strategies on these topics can also stimulate different outcomes in students, but this was not the focus of our current study.

Thus, overall, research on the impact of a European dimension in the school curriculum on identity formation of students is scarce. The existing studies show only a modest impact, while the direction is not always clear. Therefore, based on existing research, we expect that obtaining more knowledge on Europe through the school curriculum will have limited effects on students’ European identity.

In response to increasing ethnic and cultural diversity, educational policies have addressed issues concerning citizenship and (sub)national identities in several ways. Three main approaches to addressing diversity can be identified (Guimond et al.2014; Haenni Hoti et al.2017; Schachner et al.2016). An assimilationist approach aims to reduce or eliminate diversity and expects ethnic minorities to adopt the dominant majority culture. A colour-blind approach ignores group differences and existing diversity (Guimond et al. 2014). A multicultural approach in education aims to include ethnic minority students by affirming their distinct ethno-cultural identities (Dovidio et al.2007).

Research conducted by Celeste et al. (2019) found no differences in feelings of school belonging and academic achievement between ethnic minority and majority students when diversity was acknowledged and valued (i.e. via a multiculturalist approach in the curriculum). On the contrary, when diversity was either ignored (in a colour-blind approach) or rejected (in an assimilationist approach), this was detrimental for the belonging and achievement of ethnic minority students. Likewise, Faas (2010) found that incorporating cultural pluralism into the curriculum only promoted social cohesion and collective identities of students in contexts where inclusive practices were integrated. In other words, to generate a positive effect on belonging and collective identification processes, school diversity policies need to incorporate diversity and allow students to (re)construct, define, and practice their identities that are recognised and valued at school. Moreover, promotion of ethnocentric views or solely and superficially celebrating diversity undermined social cohesion in schools and led to lower feelings of national identification.

These studies indicate that ethno-cultural approaches to national identity in education fail to create shared feelings of belonging, whereas definitions of a national identity that include peers who are culturally different from the majority group facilitate feelings of national belonging among minority students (Gharaei et al.2018). In the current study we elaborate on these findings and examine whether and how the presence of a multicultural dimension in the curriculum is associated with students’ European identity. Based upon the discussion above, we hypothesise that when students feel that their school positively supports discussions on diversity and incorporates these into the curriculum, this is positively associated with European identity as a collective identity construct.

Additionally, a key question in highly diverse contexts is whether there are differences between ethnic groups. Ethnic minority students are socialised in an education system that predominantly promotes the attitudes, beliefs, and traditions of the majority community (Brint 2017). Faas (2017) found that ethno-cultural understandings of (sub)national identities make it difficult for ethnic minority students to relate positively to that (sub)national identity. When Europe was framed in an ethno-exclusive way that only included white Christians, Turkish Muslims struggled to relate positively to Europe. Yet, Faas (2017) showed that when conceptions of Europe were affiliated with multiculturalism, Turkish Muslims found ways to relate to Europe in a positive way. Thus, multicultural teaching practices acknowledging and discussing cultural diversity have positive effects on the European identification of ethnic minority students, while they have no negative effects on the European identification of ethnic majority students. Therefore, in this study, we seek to explore whether we find differences in the effects of a multicultural dimension and/or a European dimension in the curriculum on the European identification of Western European students and Turkish and Moroccan students.

The current study is conducted in the context of the Flemish education system, which proves to be an interesting setting, as we argue below. Flanders is the Dutch-speaking semi-autonomous region of Belgium and is fully autonomous in terms of education. In recent years, issues of (sub)national identities and belonging have been prominently discussed in politics and the media. In Flanders, the (sub)national identity is predominantly defined by the ethnic majority culture (Gharaei et al. 2018). Students with a migrant background often identify much less with Flemish subnational or Belgian national identity (Agirdag, Phalet, et al.2016). Moreover, in Belgium, public education is a regional (subnational) competence. Thus, the education policies implemented in Flanders focus on the further development of Flemishness and Flemish identity. This is done through quite strict language policies (often prohibiting the use of the home language in the school context), making the school environment largely monolingual Dutch (with the exception of valued languages such as French, English, and German) (Pulinx et al.2017). However, in the past two years the discussion on the position of Flemish identity has broadened and the new Flemish government – driven by its new conservative right-wing Minister of Education – has also formulated the policy goal of including a ‘Flemish Canon’ – presenting the ‘core’ of Flemish identity, history and culture – in the regular curriculum of public education. This makes studying issues of collective identity a salient topic in Flanders, but also relevant, as similar policy initiatives can be found across European education systems. Moreover, the Flemish education system is characterised by some of the largest performance inequalities related to students’ migration background and socio-economic background (Danhier et al.2014). In the Flemish education system, much more than in most systems, these background variables negatively with educational performance and feelings of school belonging (Van Praag et al. 2019).

These recurring patterns make it relevant to study issues of collective identity, in particular in relation to the position of adolescents from diverse ethnic background in public education. Barrett (2000) indicated that identity becomes more salient in childhood and develops during adolescence as individuals seek to explore and find their place in society and in interpersonal contexts. In this study, we focused on students in their final years of secondary education, as they can be expected to have developed their identities further and to have learned to cultivate attitudes to social and political issues, having also navigated the Flemish education system for nearly twelve years (Reijerse et al.2013).

In this study, we used data from 636 students (35.8% boys, 63.6% girls²) in the final two years of their secondary education. These students are between 16 and 19 years old. The data originated from survey data from 1039 students at 17 secondary schools in Antwerp, Belgium and were collected in 2017 for a research project³ on collective identity formation in education, conducted by the authors (see also Clycq et al.2021). The sample used for this study was limited to students of Western European origin (n =  288) and students of Turkish and Moroccan origins (n =  348). Focusing on students with a Turkish and Moroccan background, and bringing them together in one category of ethnic minority students, is interesting for several reasons. They are not only the largest non-EU groups in the Flemish education system (which is the context of our study), but they also experience some of the largest inequalities in education, and often have similar identification processes related to national and European identity (Agirdag et al. 2016b; Clycq et al.2021). Moreover, despite their long presence in Flanders (Belgium) since the start of labour migration from Morocco and Turkey in the 1960s, they are still considered as ‘different’ from Flemish mainstream society and their identification processes are often questioned (Clycq and Levrau 2017). From a more statistical perspective, it was important to construct these two categories, as it enabled us to compile two groups of sufficient size to allow comparisons. Students were labelled as ‘Western European’ when they were born in Belgium, the Netherlands, Germany, France or the UK, and when both their parents and grandmothers were also born in one of these countries. Students were categorised as ‘Moroccan/ Turkish’ if either they were born in Turkey or Morocco, one or both parents were born in Turkey or Morocco or if both grandmothers were born in Turkey or Morocco. These groups were formed in line with research done by Agirdag et al. (2011). We are aware that this categorisation can be problematic, because hypothetically these students could identify as Belgian and not as Moroccan or Turkish. However, our preliminary analysis indicated that the bivariate correlation between the ethnic identification of the students and our categorisation was extremely high, yet, the variable asking for ethnic identification had missing variables. Therefore we chose to categorise on the basis of the birth country of the parents and/or the grandparents. Moreover, this categorisation based on the country of birth is in line with other research done in the same field (e.g. Agirdag et al. 2016b, Jugert et al. 2019). Furthermore, research done by Agirdag et al. (2016b) and Clycq et al. (2021) has shown that Moroccan and Turkish students construct European identities in similar ways. Preliminary ANOVA- analyses confirmed that no differences in the main variables were reported within the subgroups of students with Western European migration backgrounds, or between Turkish and Moroccan migration backgrounds.

The data were collected in February and March 2017. The participating schools varied with regard to neighbourhood (Antwerp, Berchem, Merksem, Wilrijk, Borgerhout, Deurne, and Linkeroever), ethnic composition (heterogeneous and homogeneous), educational track (academic, vocational, and technical track⁴) and educational network (public Flemish Community networks, denominational (mainly catholic) network and Municipal and Provincial network). Consent to administer the questionnaires was obtained from the school administrators. Students filled out a questionnaire on ethnic, religious, cultural, and (supra)national identity formation in the classroom, with one or two researchers present. The questionnaire was administered in Dutch and took approximately 40 min to fill out. The questionnaires were anonymous, and the students were informed that their answers would be treated confidentially. Student names were not recorded, and respondents were given the opportunity to stop participating any time they wanted.

European Identification was assessed with three items that aimed to determine an individual's feeling of belonging in Europe, based on Keating (2016). We adjusted this scale in such a way that the items measuring feeling part of the European Union were not included, as this was not the focus of our study.⁵ These items were ‘I am proud to live in Europe’, ‘I feel part of Europe’, and ‘Indicate how strongly you identify with being European’. For the first two items, response options ranged from 1 ‘strongly disagree’ to 5 ‘strongly agree’, and for the third item from 1 ‘not at all’ to 5 ‘very strongly’. The level of internal consistency, as determined by a Cronbach's alpha, for the total sample was α = .714, for the sample that consisted of the students with a Western European background it was α = .718, and for the sample of students with Turkish and Moroccan backgrounds it was α = .747. The responses to these items were averaged.

The Multicultural Dimension in the Curriculum was captured with ten items, with answer categories ranging from 1 ‘strongly disagree’ to 5 ‘strongly agree’. These items were inspired by the School Diversity Inventory developed by Gottfredson et al. (2011) and adapted to fit our broader study. Examples of items were ‘I feel that there is enough attention paid to themes such as cultural diversity in courses in school’, ‘I feel that we can talk about different cultural backgrounds in school’, and ‘My school offers students from different ethnic or religious groups opportunities to understand each other’. The Cronbach's alpha for the total sample was α = .735, for the sample that consisted of the students with a Western European background it was α = .723, and for the sample of students with Turkish and Moroccan backgrounds it was α = .735. The responses to these ten items were averaged. See appendix 1 for the exact wording of all ten items.

European Dimension in the Curriculum. To measure whether Europe is part of the curriculum in schools, eight items were used, on a scale from 1 ‘strongly disagree’ to 5 ‘strongly agree’. The scale ‘European Dimension in the curriculum’ is inspired by items used in the ICCS measuring ‘Students’ reports on opportunities for learning about Europe at school’ (see also Keating 2016) and adapted to fit our broader study. This scale included items such as ‘My school gives me opportunities to learn about European history’, ‘My school gives me opportunities to learn about solidarity between people in Europe’, and ‘My school offers enough opportunities (e.g. travel, projects) to get in touch with other Europeans’. The level of internal consistency, as determined by a Cronbach's alpha, for the total sample was α = .813, for the sample that consisted of the students with a Western European background it was α = .800, and for the sample of students with Turkish and Moroccan backgrounds it was α = .827. The responses to these eight items were averaged. See appendix 1 for the exact wording of all eight items.

Controls. Gender was used as a dichotomous variable (female = 0, male = 1). In research done by Agirdag et al. (2016b), it was found that ‘native’ Belgian girls identified less with Europe than ‘native’ Belgian boys. Moreover, Verhaegen et al. (2013) showed similar results and found that girls reported lower levels of perception of opportunities to learn about Europe in school. Research done by Agirdag et al. (2016b) and Jugert et al. (2019) shows significant effects of parental SES on the European identification of students. Therefore, we included students’ parental educational level as a control variable. This variable was measured by taking the highest educational level of both parents as an indicator. Four categories were created: ‘lower education’, ‘middle education’, ‘higher education’ and ‘other’. The reference category was ‘lower education’. This variable is related to the socioeconomic status of the family in which the adolescent grows up. The educational track of the students was divided into three groups: ‘Academic’ (ASO), ‘vocational’ (BSO), and ‘technical’ (TSO) education. ‘ASO’ was selected as the reference category. We take this control variable into account since a higher percentage of Turkish and Moroccan youth are attending vocational education. Moreover, the curriculum of the different educational tracks differs; it is therefore important to take this control variable into account. Lastly, at the school level, we measured the ethnic composition of the school by the proportion of students in the school who speak a different language at home, using data gathered from the Flemish Educational Department. The ethnic composition of the schools is relevant, since the ethnic diversity of the school may influence different educational outcomes, such as perceived belonging and academic achievement (e.g. Agirdag et al. 2016a; Celeste et al.2019). The ethnic composition of the total student population of the participating schools ranged from 7.21 percent to 81.43 percent, meaning that in the school with the lowest ethnic diversity, 7.21 percent of the students speak a different language at home in addition to Dutch, and in the school with the highest ethnic diversity, 81.43 percent of students speak a different language at home, in addition to Dutch.

First, we examined the extent to which students identified as European, by comparing the mean level of European identification of the two groups, using an independent sample t-test. We also used an independent sample t-test to examine whether the groups differ in their perception of the extent to which a European dimension and/or a multicultural dimension had been included in the curriculum. Second, we studied the effects of the main variables (the European and the Multicultural dimension in the curriculum), as well as the control variables gender, educational track, educational level of the parents and the ethnic composition of the school, on European Identity for both groups, and estimated a random intercept model to explore the individual-level variables. We carried out each analysis separately for ‘Western European’ students and ‘Turkish and Moroccan’ students, since we hypothesised that the association between a European dimension and a multicultural dimension in the curriculum might have different effects on the extent of European identification of both groups.

Students with missing values for background characteristics included as control variables in the empirical analyses were excluded from the models. This explains the lower n for the unconditional model and the main model compared with the N that is given in the descriptives. The metric predictors are grand mean centred. To compare the impact of the two main variables on the dependent variable, standardised effects are reported in the results section. To make interpretation possible, unstandardised effects are reported in the tables. There was no evidence of multicollinearity, as assessed by tolerance values greater than 0.1.

Since the data have a nested structure of students clustered within schools, multilevel regression analysis is appropriate (Hox 2010). This allows us to model the variation between students and the variances between schools. In this study, we distinguished between two levels; (1) the student level, and (2) the school level (using RStudio, two level procedure). An unconditional model was determined, including none of the independent variables. The unconditional model for both groups revealed that most variance was situated at the individual level, see Table 1. For the subset including Western European students, no significant variance was found at school level (p > .05). This suggests that schools were not a contributing factor to the variance in European identification for students of Western European descent. Nevertheless, justifying the need for a multi-level analysis, Table 1 indicates that a significant variance of 4% was found between schools for the subset that consists of Turkish and Moroccan students (p < .05).

Using independent sample t-tests, we examined the mean differences between students with Western European origins and Turkish and Moroccan students. A Welch t-test was run to determine if there was a statistically significant difference in European identity between the two groups due to the assumption of homogeneity of variances being violated, as assessed by Levene's test for equality of variances (p = 0.001). No significant difference (M = 0.10) was found, t(632.359) = 1.731, p = 0.084. The Cohen's d (0.14) indicates that this is a very small effect. This result is in line with our first hypothesis (H1).

Furthermore, no significant difference was found between the groups concerning their perception of whether the school included a European dimension in their curriculum, (M = 0.006), t(630) = −0.120, p = 0.904. However, an independent t-test shows a significant difference between the two groups in their perception of the extent of a multicultural dimension in the curriculum in schools, (M = 0.18), t(631) = 3.72, p < 0.001. However, the effect size is rather small (Cohen's d = 0.30). This result suggests that Western European young people have views on whether their school includes and accepts students from different religious, cultural and ethnic backgrounds which differ from the views of their peers of Turkish and Moroccan origin.

Lastly, multiple correlations were assessed to determine whether there were associations between the main variables and to exclude issues of multicollinearity, as well as possible interaction effects between the main variables (Table 2). Results can be found in the table in appendix 2.

The main models for both subsets are significantly better than the null model (⁠$χ2<.05$ for −2LL tests, for null models see Table 1), which indicate that the models with the main variables and the control variables for both subsets are more consistent with reality. Table 3 shows the results of the complete models, including the independent variables and control variables. The results suggest that students’ perception of a European dimension in the curriculum as well as their perception of a multicultural dimension in the curriculum have significant and rather similar effects on the European identification of Western European young people, respectively (⁠$β=0.092$⁠, p < .05 and $β=0.130$⁠, p < .01). These results are in line with our H2, and H3. Moreover, we find that a higher level of education of the parents is positively related with stronger European identification. The other student-level control variables as well as ethnic composition at school-level are not significantly related with higher levels of European identification of Western European youth.

For the subset of Turkish and Moroccan students, the perception of a European dimension to the curriculum also has a significant effect on the strength of their European identification (⁠$β=0.115$⁠, p < .05). Furthermore, a multicultural curriculum also has a significant effect on the European identification of Turkish and Moroccan students (⁠$β=0.202$⁠, p < .001). When comparing the two main variables, the standardised effect sizes suggest that a multicultural curriculum has stronger effects on the European identification of Turkish and Moroccan students than a European dimension in the curriculum. When looking at our control variables, our results show that Turkish and Moroccan boys identify significantly less as European than their female peers, when the other variables remain constant. Parental educational levels do not have a significant impact; however, students attending vocational education (‘BSO’) showed significantly higher levels of European identification than students attending academic education (‘ASO’). After controlling for gender, educational track and the educational level of the parents and the addition of the main variables to the model, schools explained 3.0% of the variance in European identification (p <.05). This suggests that the differences in the extent of European identification of Moroccan and Turkish students can mainly be explained by their individual-level characteristics. Lastly, the ethnic composition of the school appears insignificant and does not lead to stronger European identification for Moroccan and Turkish students. This means that the presence of more ethnic diversity, as measured through the use of languages other than Dutch at home, does not affect the strength of European identity.

The aim of the current study was to deepen our understanding of the role of the school curriculum in European identity formation. We first examined how students’ perceptions of a European dimension and a multicultural dimension in the curriculum predict their European identification. For this purpose, this study differentiated between students with a Western European background and students with Turkish and Moroccan origins. We found that both Western European students and Turkish and Moroccan students identify relatively strongly with Europe. In line with our expectations, we found no significant difference between the two groups in European identification, which underlines the importance of European identity as a collective identity. This is an interesting finding, because it suggests that a European identity can be(come) an identity which students with and without non-Western migration backgrounds can relate to; this contrasts with (sub)national identities, which are generally very attractive for ‘natives’ but much less so to individuals of ethnic minority or immigrant origin (Alba and Foner 2015). Results also suggest that the perceptions of the integration of a European dimension into the curriculum were the same in the two groups; Western European students, however, reported greater perception of a multicultural curriculum. This is a remarkable finding, given that a multicultural curriculum is often developed to explicitly include the experiences and background of minoritised groups. This could indicate that students with Turkish and Moroccan backgrounds feel that the multicultural dimension of the curriculum is limited, while Western European students feel that the school is already making considerable efforts to become more multicultural.

Secondly, we investigated how a European dimension in the curriculum and a multicultural curriculum were associated with a European identity. Both our main variables – a European and a multicultural dimension in the curriculum – predicted a stronger European identity for both groups, which was in line with our second and third hypotheses. The effect sizes of the two variables were rather similar for the students with Western European origins, and the effects of the two dimensions of the curriculum contributed equally to their European identification. Yet, for the Turkish and Moroccan students, a multicultural dimension in the curriculum had a stronger impact on their European identification than a European dimension in the curriculum, as assessed by the standardised regression coefficients. These results are the first exploration, to assess whether multicultural teaching practices acknowledging and discussing cultural diversity have positive effects on the European identification of ethnic minority students. Based on our results, we can conclude that a multicultural dimension in the curriculum is a stronger predictor for the European identification of Turkish and Moroccan students than a European dimension in the curriculum, whereas a multicultural dimension in the curriculum is equally important as a European dimension in the curriculum for the European identification of Western European students.

Lastly, we found limited effects of the control variables on students’ European identification. Contrary to existing literature (Agirdag et al. 2016b; Roose 2013), we only found two significant associations for the subset that consisted of students with a Turkish and a Moroccan background: namely being a male negatively predicted European identification, whereas attending the BSO level positively predicted European identification. The latter might be explained by the presence of more ethno-cultural diversity in the BSO (vocational) track which traditionally has a disproportionate number of students from ethnic minorities and/or lower SES backgrounds, compared with other educational tracks.

In this study, our main focus was on the school curriculum; this does not allow us to draw conclusions on the effects of existing social relationships and interactions in school on European identification. From previous research, we know that the ethnic composition of the school is significantly related to (sub)national identifications of young people (Agirdag et al., 2011). Moreover, positive intergroup contact between ethnic minority and majority peers plays a significant role in the development of a sense of national belonging (Fleischmann and Phalet 2018; Leszczensky et al.2016). All participating students are attending schools in a highly diverse city. It is possible that these cities could be ‘outlier’ cities, which also explains the low variance between schools in this study. Here we argued that a European identity is more inclusive, because it is less ethno-exclusively defined than (sub)national identities. However, it is likely that the importance of a European identity as a collective identity is more salient in contexts with a higher level of diversity, where ethnic minorities feel that they are excluded from national imaginaries (Faas 2017). Future research should focus on multiple contexts with different demographics, to examine whether differences exist between schools in more or less diverse contexts.

Furthermore, we asked students about their perception of the integration of European and multicultural dimensions into their school curriculum. We are aware that it is a common practice within curriculum studies to analyse textbooks and end requirements. However, we believe that the perceptions of the students – the ‘experienced curriculum’ – are equally important when considering the effect on their identity constructs. Students in secondary education have different teachers for the different subjects that they are taking. Since teachers are given the freedom in Belgium to include cultural diversity approaches in their teaching (Civitillo et al. 2017; Plaut et al.2018), inclusion of these approaches often depends on the goodwill and personal interest of the individual teacher (Agirdag et al. 2016), and therefore cannot be understood as the common practice of the whole school. Moreover, research indeed shows the importance of student perceptions, as young people often develop their actions based on such perceptions (e.g. Verhaegen et al.2013). Therefore, it is interesting to note that the groups we studied are similar with respect to the European dimension of the curriculum but differ significantly with respect to their perception of a multicultural dimension in the curriculum. Currently, the Flemish education system – like many other systems across Europe – is in the process of implementing a stricter curriculum approach emphasising a ‘homogeneous’ Flemish history, culture and language. Based upon our findings, we can expect not only that students’ perceptions of this new curriculum will differ according to their broader position in society and socio-demographic background, but also that this will impact their identification processes. In light of our findings we would argue that – given the low attractiveness of Flemish identity to minorities and the importance of a multicultural curriculum for identification processes – Flemish policy makers will not reach their goal of making Flemish identity more attractive to ethnic minorities.

Moreover, it is important to understand that identification processes are not simply a top-down phenomenon. Although the last decennium has seen a proliferation of new nationalist curriculum content (Vincent 2019), young urban students have a multitude of repertoires they can build on and identify with; micro-level processes of young people seeking new ways to build collective identities must therefore be acknowledged and incorporated into educational processes. Attempts to include students from diverse ethno-cultural backgrounds in a shared collective identity might be hampered when school curricula focus on a more homogeneous definition of these collective identities, as is done in the UK with the explanation of ‘British values’ or in France with the so-called neutral narrative of French republicanism. Minority youth might be more inclined to identify with a collective identity that explicitly acknowledges diversity and heterogeneity, rather than ethno-cultural and national homogeneity. Prioritising inclusive collective identities will become even more important for future European societies, which will continue to encounter more diversification processes as a result of globalisation and migration. Accordingly, policy makers and researchers should take into account that solely teaching about civic (European) knowledge and focusing on strengthening (sub)national identitites will not bind ethnic minority and majority youngsters to the same ‘imagined community’. Such a gap in shared feelings of belonging may affect social cohesion in superdiverse contexts.

Lastly, although European identity might be a vessel to develop such a collective identity, there is also evidence that it can be imagined in ethno-exclusive ways (Jenkins 2008). The juxtaposition between ‘Europeans’ and ‘others’ still implies that there is ‘us’ and there is ‘the rest of the world’. Depending on the interpretation given to European identity in politics and media, this identity construct may still not be suitable to unite youngsters from different backgrounds. To conclude, we do not yet understand all the complexities around these new phenomena and processes. This underlines the need to further untangle ‘the complexities of how identities are formed and performed’ (Gholami 2017, 798), in order to better assess the impact of education on shared feelings of belonging.

No potential conflict of interest was reported by the author(s).

Gert Verschraegen (BA, MA, University of Leuven, MA, University of Hull, PhD, University of Leuven) is Associate Professor of Sociology at the University of Antwerp (Belgium). His research interests are in the areas of social theory, sociology of science, cultural sociology and ethnic relations. His current research focuses on cultural and religious diversity in European cities and on the socio–cultural consequences of the refugee crisis.

Ariadne Driezen is a doctoral researcher at the University of Antwerp. Her research interests are in–group and out–group boundary relations, the intersection between gender, ethnicity and religion, identity constructions and the impact of stigmatisation for ethnic minorities. She studies these topics in different settings, in secondary and tertiary education and on the labour market.

Noel Clycq is a research professor in education sciences at the Department of Training and Education Sciences at the University of Antwerp. His main research topics are diversity, identity, globalisation, and the governance of learning.

Emma Carey Brummer is a doctoral researcher at the University of Antwerp. Her research focuses on belonging, citizenship, and emotions of young people attending secondary education. Other research interests are (in)equality, diversity, intersectionality, sociological theory, and identification processes.

Multicultural dimension in the curriculum

Q1 I feel that enough attention is paid to themes such as cultural diversity in courses in school

Q2 The school has a clear policy on multiculturalism

Q3 I feel that the school organises sufficient events on themes such as cultural and religious diversity

Q4 I think that my school does not take into account the holidays of multiple religious groups that are present in the school

Q5 The school provides spaces in which to offer or perform other religious activities

Q6 I can choose a lesson for the religion classes that matches my religious belief or philosophy

Q7 I feel that we can talk about different cultural backgrounds in school

Q8 My school offers students from different ethnic or religious groups opportunities to understand each other

Q9 In my school not all students are allowed to talk about their own (religious and cultural) views

Q10 In my school, all students can learn about the views of other people and other religious and cultural groups

European dimension in the curriculum

Q1 My school gives me opportunities to learn about European history

Q2 My school gives me opportunities to learn about European institutions such as the European Parliament or the European Commission

Q3 Classes regularly discuss the future of Europe

Q4 My school gives me opportunities to learn about multiculturalism and diversity in Europe

Q5 My school gives me opportunities to learn about solidarity between people in Europe

Q6 My school gives me opportunities to learn about human rights such as equality and freedom

Q7 My school gives me opportunities to learn about European culture and values

Q8 My school offers enough opportunities (eg travel, projects, etc.) to get in touch with other Europeans

Bivariate correlations of main variables for the Total Sample and the subsets with Western European students and Moroccan and Turkish students.