A multilevel analysis of factors influencing teenagers’ identification with Europe: the effects of migration and learning opportunities

Abstract

The European Union (EU) faces challenges that affect its persistence, including the revival of national populism in many EU member states. Studies have shown that individuals with immigration histories identify less strongly with Europe than individuals without immigration histories. Therefore, fostering a strong identification with Europe is more relevant than ever. This paper will explore the possible historical roots of different levels of identification and examine if differing access to learning opportunities can explain the difference. Drawing on data from the German sample of the International Civic and Citizenship Study 2016, this paper aims to determine the relevance of individual variables and learning opportunities for the development of students’ identification with Europe. Multilevel analyses at individual and classroom level were conducted introducing different independent variables. Results show that having no immigration history from outside the EU, being Christian or atheist, and learning about Europe at school are predictors of a stronger level of identification with Europe. The effect of having an immigration history from outside the EU loses significance when socio-economic status classroom composition is entered into the model. We conclude that differences in identification are not due to different access to learning opportunities, but likely due to personal characteristics.

Introduction

A shared identity is important for communities and political systems. The EU currently faces challenges relating to its collective identity. This can be seen in the Brexit referendum (Ammaturo, 2019) and the revival of national populism and extremism in many EU member states (Ekman & Lane, 2022; Noack & Eckstein, 2023).

Citizens and their identities are vital for European cohesion. Citizens who identify more strongly with Europe can be expected to feel greater solidarity with member states (Verhaegen, 2018) and to be more likely to support European integration. On the other hand, citizens who identify less with the EU are more likely to slow down the process of European integration, for example by voting against European solutions in referendums (van Spanje & Vreese, 2011). Consequently, the question of how a European identity can be fostered and stabilised gains relevance.

Adolescence and early adulthood are critical times for identity formation (Dani, 2015). As the school setting is an important developmental context for young people, the relevance of schools in supporting the development of pupils’ political identities needs to be explored. We draw on data from the German subsample of the International Civic and Citizenship Study 2016 (ICCS 2016), a representative survey that was conducted in 24 different education systems in Asia, Europe, and Latin America with students in their eighth year of schooling (Schulz et al., 2018). This paper aims to increase the understanding of the importance of two sets of factors for the development of identification with Europe: first, individual student background characteristics; and second, schooling and classroom contextual factors. To this end, multilevel analyses were conducted. Previous studies have concluded that a higher socioeconomic status (Roose, 2011), being male (Höllinger & Hadler, 2012; Losito et al., 2018), and not having an immigration history (Agirdag et al., 2012; Losito et al., 2018) are characteristics which correlate with a higher identification with Europe at individual level. Due to the diverse data available from the ICCS questionnaire, it will be possible to include these and additional variables which have rarely been considered in multilevel analyses. The next sections will provide an overview of perspectives on identity in general and the concept and relevance of European identity specifically. Finally, we will discuss identification with Europe in the context of migration and schooling.

Theoretical background

The theory section is organised into four parts. First, we will define our working concept of identity as social identity, explore ways to measure and foster it and present its importance for political systems. Second, we will explore the historical background of European identity, considering its origins along with its current challenges. The third part presents critique on the concept of identity and explores the concept of identification. The fourth part will present results of previous empirical studies that have dealt with similar questions and connect these results to our hypotheses. In the fourth part, we will also consider the relevance of schooling for identification with Europe.

Social identity and its relationship with political systems

Social identity theory tries to explain why people see themselves as members of groups and how this identity influences their attitudes and behaviours. Tajfel (1972) defined social identity as ‘the individual’s knowledge that he belongs to certain social groups together with some emotional and value significance to him of this group membership’. ‘Group’ is defined here as a cognitive entity that is meaningful to an individual at a particular point in time (Tajfel, 1974). Individuals can be members of numerous groups, and these memberships contribute to the image they have of themselves (Tajfel, 1974). Leach et al. (2008) identified two main components for measuring individuals' identification with groups: self-definition and self-investment. Self-definition involves individual self-stereotyping, where individuals perceive themselves as members of the group, and in-group homogeneity, where individuals perceive themselves as similar to a typical group member. Self-investment comprises satisfaction, reflecting positive feelings about group membership, solidarity, indicating a psychological bond with fellow group members, and centrality, where the group becomes a central aspect of an individual's self-concept.

When studying how people come to identify with collective identities, there are two theoretical perspectives to consider: the culturalist and structuralist models. The culturalist model suggests that the formation of a collective identity is influenced by exposure to specific symbols and messages. The exposure to European symbols starts early in life and can occur in various settings, including schools (Recchi, 2014). In our study, we recognize the significance of education in the culturalist model by examining students' reports on how much they learn about Europe in school.

The structuralist model highlights the impact of social interactions on the formation of collective identities. These interactions, occurring in various contexts, play a crucial role in shaping people’s identification with, in this case, Europe. Indicators of the structuralist model include both physical (e.g. long and short term stays abroad for work or study) and virtual mobility (e.g. having friends or family in a different country) (Recchi, 2014). This aspect is particularly relevant to our research, as we differentiate between individuals with immigrant histories from EU and non-EU countries. Understanding the influence of having family in other European countries can provide insights into the dynamics of European identity formation and how it relates to the experiences of different immigrant groups. These interactions contribute to a broader understanding of Europe and foster a sense of identification with it. By considering both the culturalist and structuralist perspectives, we aim to provide a more comprehensive understanding of how identification with Europe is shaped. In a next step the relevance of identifications will be explored.

According to Norris (2017), political identity is an aspect of diffuse political support. Political support describes the attitudes and behaviours that political systems rely upon to persist (Easton, 1957). Specific aspects of political support, such as trust in political institutions and officeholders, may vary according to citizens’ judgements and may lead to reforms of the system. Diffuse support is a prerequisite for the stability of a political system; it represents abstract feelings and lasting bonds with states, such as national pride and identity (Norris, 2011). While theories of political support have been developed with reference to nation states, more recently the concept has also been applied to the supranational organizations (Abs, 2021; Wiesner, 2019). According to Duina (2008) regional trade areas such as the EU and Mercosur have played a significant role in fostering the formation of supranational identities. These identities encompass a shared sense among citizens of member states that they hold fundamental viewpoints and beliefs about the world. Contrary to viewing regional trade areas solely as economic endeavours, integration processes within these organizations involve various supporting mechanisms and organized activities that contribute to the development of shared beliefs and perspectives. These activities can include the standardization of laws and the organization of summits, as exemplified in both the European Union and Mercosur (Duina, 2008). Conversely, the existence and success of regional trade areas like the EU and Mercosur rely on a foundation of common identity among member states. Fostering a European identity may be one way of promoting the persistence of the EU.

Next, we will explain how a sense of European identity has been developed throughout Europe’s history and what its current challenges are. Understanding European identity and its history is important for explaining why specific groups identify less strongly with Europe. Along with results of previous empirical studies, this understanding influenced the development of our research question and hypotheses.

The historical and political relevance of European identities

When explaining the creation of a sense of European identity, it is important to understand what aspects determined the social identities in the geographical area of Europe before the supranational European identity began to develop. As Triandafyllidou (1998) argued, the process of social identity formation is closely linked to differentiating and distinguishing from others. Up to the seventeenth century, religion—specifically Christianity—was the defining factor of political identity in Europe (Rich, 1999) as well as the key differentiating factor which motivated many wars such as the Crusades (Stråth, 2002).

However, the Thirty Years War revealed that Christianity (alone) couldn't unify Europe, as it pitted Catholics against Protestants (Stråth, 2002). Consequently, new political structures emerged, and the idea of Europe started to crystallize in the late seventeenth century, accompanied by Enlightenment values of science, reason, progress, and democracy (Stråth, 2002).These ideas remain important to the formation of European identity to this day and are now considered as the main identifiers of a European identity.

Official, institutionalized efforts to foster a European identity began in 1973 at the Copenhagen Summit, emphasizing the common heritage of European Community member states (Heinemann et al., 2020). The Maastricht Treaty in 1992 gave birth to the "European citizen" and the EU, ensuring equal social and employment rights for all EU citizens and transferring competencies to the European level (European Union, 1992; Barth & Bijsmans, 2018). These milestones marked the establishment of a comprehensive political system at the European level (Bergbauer, 2017).

The concept of European citizenship is important to our methodology. We acknowledge that possessing European citizenship does not automatically correlate with not having an immigration history, as individuals with immigration history can still hold European citizenship and enjoy the privileges it entails, such as the social and employment rights. Nonetheless, our approach aims to capture a distinction between students with different immigration histories rather than directly assessing their citizenship status. This approach is relevant because it recognizes that even individuals with European citizenship may experience feelings of being an outsider and face discrimination. Findings from the EU-MIDIS II survey conducted in Germany revealed a pattern of racial discrimination in job searches, targeting individuals from sub-Saharan African backgrounds based on their skin colour and accent, and individuals from Turkish backgrounds based on their names and country of birth. These findings emphasize that discriminatory practices are often rooted in factors beyond an individual’s citizenship, such as skin colour and family names (Franke & Schlenzka, 2019).

El-Tayeb (2012) further argues that Europe's historical narrative is shaped by an ongoing purging of its internal racial Others, including black, Muslim, Jewish, and Roma populations. This narrative presents Europe as untouched by any form of hybridity, erasing the contributions of these marginalized groups. Consequently, immigrants from non-EU countries, particularly those who belong to ethnic and religious minorities, may face exclusion and a sense of not fully belonging to Europe. Therefore, it is necessary to delve deeper into the complex dynamics of identity formation in Europe and explore how these historical and discursive factors shape the identification processes of Muslim citizens and citizens of other minorities. Religion has thus not been discarded completely as an identity forming aspect in Europe. Christianity and secularism are part of the European history, and a German study found that they are both regarded more positively than other belief systems (Yendell, 2013). Therefore, we assume that Christians and people without religion might be more likely to identify with Europe. Drawing on El Tayeb's insights, which highlight the impact of dominant narratives on identification processes, it is crucial to consider how the persistent portrayal of Europe as a white and Christian space in public and political discourse affects individuals from diverse religious and ethnic backgrounds. The lower sense of identification with Europe among people of non-Christian religious backgrounds may be attributed to these exclusionary practices associated with dominant narratives.

It is important to recognize that the stability of European identity has been further tested in recent years. Factors such as the populist reactions to refugee movements and Brexit have contributed to a sense of uncertainty and diminished Europeanness (Ammaturo, 2019; Fligstein et al., 2012). These challenges to European identity pose additional complexities in the process of identification for individuals from diverse backgrounds. National populism plays a significant role in shaping the deconstruction of European identity, particularly through the creation of binary divisions (Noury & Rolan, 2020). Populist parties argue that cultural diversity undermines the integrity of national and European identity and poses a threat to traditional values and customs. This rejection is particularly apparent in the context of Islam, which is portrayed as an "other" against which the populist right constructs its notion of Western civilization and values (Betz, 2003). This underscores the significance of incorporating religion as a predictor in our methods to examine the relationship between religious affiliation and identification with Europe. Furthermore, Noack and Eckstein (2023) point out the limited empirical understanding of how schools influence youth's populist beliefs. This aspect further emphasizes the potential of our study, as it explores the impact of formal opportunities to learn about Europe on students' identification with the continent.

This section has shown that different ideas and values have been connected with the idea of European identity. The literature review raised the question if students with immigration histories (especially those with non-European immigration histories) and non-Christian students continue to feel less European, or if more recent concepts of Europe have become more inclusive.

The subsequent chapter will delve into the examination of the critical perspectives on the notion of identity. The discussion will shed light on the scepticism surrounding the concept of 'European identity' and its relevance to individuals, as highlighted by scholarly discourse. In addition, we will provide an explanation for our choice to adopt the term ‘identification with Europe’ instead of ‘European identity’.

Identification with Europe and criticism on the idea of identity

The concept of 'European identity' has sparked scepticism within scholarly discourse, prompting a critical re-examination of its usage. Brubaker and Cooper (2000) criticize the term 'identity' itself, emphasizing its ambiguity. They argue that employing 'identity' as a category of analysis is limiting and fails to differentiate diverse affinities, forms of belonging, and self-understandings. Instead, they advocate for the use of 'identification' as a more active term due its derivation from a verb, which avoids reification and facilitates a detailed analysis of the agents involved in the identification process. 'Identification' does not presuppose internal sameness or bounded groupness, acknowledging the situational and contextual nature of self- and other-identification. Brubaker and Cooper (2000) emphasize that identification is intrinsic to social life, occurring in various contexts and often influenced by discourses and public narratives that permeate our ways of thinking and sense-making. By adopting the framework of 'identification with Europe,' we embrace a more dynamic and context-sensitive approach that recognizes the diverse nature of individuals' relationships with Europe. We seek to enhance the precision of our data analysis by utilizing the term "identification" as a more fitting descriptor for the data collected through the scale we employ, It is worth noting that, despite the conceptual distinction between 'identification' and 'identity,' existing research often uses these terms interchangeably.. Diez-Medrano (2010) further contribute to the discussion by drawing attention to the assumption that the category identity holds meaningful relevance to individuals, highlighting how many respondents are surprised when asked about their European identity, as they have never contemplated such a self-perception. This highlights the need for caution when relying on data from surveys on European identity.

The next chapter will present the results of previous empirical studies that explored factors which influence European identity. The discussion will shed light on two aspects: first, which groups of people might be more likely to identify more or less with Europe; and second, which aspects of schooling might be relevant for fostering European identity.

State of research

Multiple empirical studies conducted in various settings within EU countries have shown that several factors and personal background characteristics can influence individuals’ sense of identification with Europe. This section focuses on empirical studies analysing how various predictors influence identification with Europe at individual level. We will discuss the relevance of individual characteristics such as gender, socioeconomic status, and immigration histories, and the pertinent aspects of schooling.

One of the factors that might influence individuals’ sense of identification with Europe is gender. Three studies of children and teenagers showed that girls exhibit significantly lower levels of European identification than boys (Agirdag et al., 2012; Jugert et al., 2019; Losito et al., 2018). However, other recent studies conducted with adults found that women have higher levels of identification with Europe than men (Aiello et al., 2019; Verhaegen et al., 2017). The researchers did not explain this shift; while we will not offer a hypothesis for this connection, we will explore the possible relevance of gender.

Socioeconomic status (SES) is an important aspect of an individual’s life in society. A person’s understanding of their SES can help to establish a sense of belonging and group membership (Destin et al., 2017). Verhaegen and Hooghe (2015) concluded that personal economic benefits due to the country of residence’s EU membership is an important determinant for developing an identification with Europe. According to the utilitarian model of public support, the reason for this might be that wealthy citizens experience more benefits from European integration, such as investment opportunities, whereas citizens with low incomes could see a reduction in their welfare because of the European single market (Gabel, 1998). Quantitative analyses have shown a weak connection of SES with identification with Europe in adults (Aiello et al., 2019; Fligstein, 2009; Verhaegen et al., 2017) and children/teenagers (Agirdag et al., 2012; Jugert et al., 2019). However, this correlation was observed in only two out of the fourteen European countries participating in ICCS 2016 (Ziemes et al., 2019). This indicates that the connection can be context dependent and is not universal. Therefore, we do not expect to see a correlation between SES and identification with Europe in our analysis.

Researchers have often investigated and analysed the relevance of immigration histories. The expression ‘people with immigration history’ refers to individuals who were born outside their current country of residence or have parents who migrated. Children with an immigration history are socialised in the culture of the host society and experience some socialisation from their culture of origin. They often internalise more than one culture and identify with more than one ethnic group (Benet-Martínez & Hong, 2014). A negative connection between identifying with Europe and having an immigration history has been observed in multiple studies (Agirdag et al., 2012; Fligstein, 2009; Jugert et al., 2019; Ziemes et al., 2019). However, few authors distinguish between immigration history from within and outside of the EU. In light of our preceding discussion, this differentiation is important. Teney et al. explored this topic in 2016 with an adult sample and found that immigrants from the so-called EU15 countries (the fifteen countries which were part of the EU prior to the 2004 enlargement) tended to identify significantly more strongly with Europe than immigrants from the newer EU countries and non-EU immigrants. However, non-EU immigrants also considered themselves to be European. It is possible to postulate that ‘access to advantages provided by the EU to its citizens is positively associated with the strength of European identification’ (Teney et al., 2016, p. 2198). Identifying as European may be a way for EU immigrants to reconcile the identities associated with their country of origin and their receiving country (Teney et al., 2016). Recent research suggested that large groups of second- and third-generation individuals with a non-European immigration history do not identify with the national imagery of their country of residence but instead identify to a greater extent as European (Clycq, 2021). Findings showed that students of Turkish or Moroccan origin born in Belgium identify more as European than as Belgian or Flemish (Agirdag et al., 2016; Clycq et al., 2021). This might be due to a sense of not fitting in their community of origin or residence, as highlighted by Diez-Medrano (2010). Fligstein's work (2008) supports this notion by emphasizing the significance of foreign language proficiency and experiences of travel or living abroad in fostering disengagement from one's “national identity” in favour of a “supranational identity”. These findings align with our earlier discussion on the identification patterns of individuals with immigrant backgrounds. It suggests that the process of globalization and exposure to diverse cultural experiences can influence the formation of supranational identities, such as European identification, even among those who may initially prioritize their local identities.

Classroom composition, i.e. the fact that the clustering of students with different attributes in a given classroom can influence the effectiveness and perception of teaching, also affects the identity formation of students with immigration histories. Based on Deutsch's theory of international integration (1953) Agirdag (2012) postulated that SES-composition and ethnic school diversity are relevant to European identity formation because they reflect who students interact with when they are at school. Regarding socio-economic background, findings indicate that in classes that have higher proportions of low-SES pupils, the pupils' level of identification with Europe is lower (Agirdag et al., 2012; Jugert et al., 2019). Similar results could be observed regarding the proportion of minority students in a class. A multilevel study of German adolescents showed that the mainstream identity (in this case, German) is greater in classrooms with a lower proportion of students with immigration histories, although the diversity of immigration histories did not matter in the final model (Edele et al., 2020). This finding could indicate that the proportion of students with immigration histories in classrooms might decrease levels of identification with Europe. Agirdag et al. (2012) and Jugert et al. (2019) came to similar results in their multilevel analysis, however in the former study, it was discovered that ethnic school diversity was only marginally associated with students' identification with Europe.

Religion can influence individuals’ identities if they are significantly committed to their religion (Oppong, 2013). If members of a certain community are mostly religious, religion can play a role in identity formation since it is assumed that most people will adhere to the norms of their community (Oppong, 2013). As stated above, Christianity is historically linked to the idea of Europe itself. Yet few studies have examined the links between religion and identification with Europe. Saroglou and Mathijsen (2007) observed a negative connection between religiosity and attachment to Europe for secondary school students with immigration histories in Belgium. They observed that a high critical regard for religion (encompassing the questioning of faith and critical thinking towards religion) correlated with an endorsement of identification with Europe (Saroglou & Mathijsen, 2007). These results were observed primarily in students from countries of Muslim and Christian traditions (which is how the researchers opted to operationalise students’ religions), showing that students of different faiths who can approach their own religion critically also engage with the EU (Saroglou & Mathijsen, 2007; Teney et al., 2016).

Formal learning opportunities about Europe at school are also directly linked to a feeling of identifying with Europe, predicting the identification with Europe of students from all EU countries that participated in ICCS 2016 (Verhaegen et al., 2013; Ziemes et al., 2019). Recent research conducted in Belgium with students with European and non-European immigration histories presented similar results, concluding that a European and a multicultural dimension in the school curriculum predicted a stronger identification with Europe for both groups of students (Brummer et al., 2022). This can be explained by the fact that information gathering is necessary for resolving identity crises (Waterman, 1989) and developing identity (Marcia, 1989). Furthermore, learning opportunities about Europe have also proved to be strongly related to students’ attitudes towards supranational European cooperation (Hahn-Laudenberg & Abs, 2020; Verhaegen et al., 2013). It is expected that the efficiency of civic education at school and of learning opportunities about Europe depends not only on the background of individual students but also on classroom composition (Deimel et al., 2020); we will examine this relationship in our multilevel analysis model.

Verhaegen et al. (2013) found cognitive learning about Europe to be an important predictor for identification with Europe; however, they did not investigate if those learning opportunities differed for marginalised groups. Brummer et al. (2022) investigated this question with a sample of 636 students in Belgium and did not find a significant difference regarding the perception of opportunities to learn about Europe at school between students with and without a non-European immigration history. We will use a larger sample from another country and use mediation analysis to investigate if we can find similar results. This analysis will allow us to determine if the differences in identification with the EU stem from differences in learning opportunities or from classroom composition.

Research questions

Based on the theoretical background, we will investigate multiple hypotheses and research questions regarding the following variables that may impact identification with Europe: students’ religion; students’ immigration history; and learning about Europe at school. Our core research question is: ‘What factors—either personal characteristics or contextual aspects—can predict individuals’ identification with Europe?’.

Our study goes beyond previous research by using more precise indicators of migration history, by simultaneously regarding learning opportunities, and by statistically accounting for characteristics at both classroom and individual level. Furthermore, our research also considers the possibility of unequal access to education about Europe. Our hypotheses are as follows.

• H1.1: Students with a non-European immigration history have a weaker identification with Europe.

• H1.2: Students who are members of a religion other than Christianity have a weaker identification with Europe.

• H1.3: Students with a lower SES have a weaker identification with Europe..

• H2.1: Learning about Europe at school at individual and at classroom level is connected to a stronger identification with Europe.

• H2.2: The proportion of students with a non-European immigration history in class negatively affects identification with Europe.

• H2.3: The proportion of students with a low SES negatively affects identification with Europe.

• H3.1: Students with a non-European immigration history experience fewer learning opportunities at individual and classroom level.

• H3.2: Students with a lower SES experience fewer learning opportunities at individual and classroom level.

• H3.3: Differences in learning opportunities as per H3.1 and H3.2 explain differences in identification with Europe.

Method

Data set and sample

We used a subsample of the ICCS 2016 study. The data set contains data from 24 education systems and multiple pre-calculated scale values based on Rasch analyses that are documented in the technical report (Schulz et al., 2018). The full sample included 1451 students from 59 classes in schools located in the federal state of North Rhine-Westphalia, Germany. The sample of classes was stratified for school tracks and number of students with immigration histories. All students attended the 8th grade. The mean age was 14.31 (SD = 0.58), and 52% of the students in the sample were girls. Further information on the sample can be found in the technical report of ICCS 2016 (Schulz et al., 2018).

Measures

Religion

Of the students in the sample, 973 (67%) indicated being Christian, 195 (13%) were Muslim, and 182 (13%) indicated not belonging to any religion. The remaining 56 students (4%) reported belonging to other religions. For our analyses, we created a dummy variable that differentiated between a group of students who reported being Christian or who did not report a religious belonging (82%), as opposed to a group of students who belong to religions other than Christianity (marginalised religions) (18%). Most of the religiously marginalised students were Muslim. Combining Christian students and students who did not indicate any religion follows previous theoretical backgrounds which say that Christianity and secularism are a part of European history and are regarded more positively than other belief systems in Germany (Yendell, 2013) and avoids missing values.

Socioeconomic status (SES)

The SES indicator is part of the ICCS 2016 dataset (Köhler et al., 2018). The indicator combined information on the number of books at home, parental education, and parental occupation. It was calibrated to have a weighted national mean of zero and a standard deviation of one.

Non-European immigration history

Students’ immigration history was assessed using a question that required students to indicate the place of birth of each of their parents. Three options were available: Germany, another country in the EU, and another country outside of the EU. Students who had at least one parent who had been born in another EU country were placed in the group ‘EU immigration history’ irrespective of whether the other parent was German or from a country outside of the EU since both immigration history and a connection to the EU were present. Students who had both parents from a country outside of the EU or one parent from a country outside of EU and one parent from Germany were placed in the group ‘non-EU immigration history’ because only an immigration history from outside the EU was present. Students who marked Germany as the birthplace of both parents were placed in the group without immigration history. Out of the 1,313 students who provided information on this item, 27% (n = 330) belonged to the group ‘non-EU immigration history’ and 15%

(n = 226) were part of the group ‘EU immigration history’. Consequently, 57% (n = 829) of the students did not have an immigration history according to our definition. Missing values accounted for 5% of the sample.

Instruments

Identification with Europe

Though the ICCS data set included a scale on students’ identification with Europe, we opted to construct a scale that more closely aligned with our theoretical understanding of identification with Europe. The mean scale we constructed was based on five statements. The scale contained multiple aspects of identification with Europe, including self-stereotyping (‘I see myself as European’, ‘I feel part of Europe’ and ‘I feel part of the European Union’) and affective aspects of identity falling under the category of self-investment (‘I am proud that my country is a member of the European Union’ and ‘I am proud to live in Europe’). One item from the ICCS data set, which suggested a preference or ranking in identification aspects ('I see myself first as a citizen of Europe and then as a citizen of the world'), was omitted from our scale due to our focus on capturing factors such as affectivity and sense of belonging rather than personal preferences. Items could be answered on a four point Likert- scale from strongly disagree to strongly agree. The scale has a mean of 3.18 and a standard deviation of 0.61. Scale reliability was assessed as good with α = 0.80. Though the scale was moderately skewed towards higher values (− 0.76), this should not significantly impact the reliability of parametric statistics as the sample size was very large (Piovesana & Senior, 2016).

Learning opportunities

We used the ICCS 2016 scale on learning opportunities about Europe. Learning opportunities were assessed with one scale comprised of four items concerning students’ perceptions of opportunities to learn about Europe. Students were asked to indicate on a four-point Likert scale to what extent they learned something about the political and economic system, the history, political and social issues, and the political and economic integration of Europe. According to Recchi (2014) the learning opportunities can be regarded as cultural because the items refer to an exposure to Europe-related content. The scale was created based on Rasch analyses by the IEA (for further information, see Schulz et. al, 2018). For the German data set, the scale had a mean of 49.37 and a standard deviation of 9.75.

Procedure

In a multilevel regression model completed with Mplus 8 (Muthén & Muthén, 1998–2018), the variables gender, socioeconomic background, religion, European background, and students’ reports about learning about Europe at school were introduced as independent variables. All independent variables were centred at the group mean. The group was defined by the students’ classes. This procedure served to separate fully the variance at individual and classroom level (Enders & Tofighi, 2007).

A multilevel approach was used in Mplus. This had two major advantages: First, multilevel approaches account for the clustered nature of the sample. Students within classes are more similar to each other than a random sample of students. Second, multilevel analyses allow researchers to investigate relationships between variables separately at individual and classroom levels. They are therefore valuable for educational research that aims to understand when individual or classroom characteristics are relevant in educational processes (Goldstein, 2011).

Results

Descriptive analyses

Table 1 presents the descriptive statistics of our analyses. It shows that the sample mean of Identification with Europe of 3.18 was higher than the theoretical mean of 2.5, indicating that students generally identified with the EU. Identification with Europe correlated with learning opportunities about Europe and with the students’ SES. However, the latter correlation was rather low. The intraclass correlations (ICC) of SES, religion, and immigration history were over 0.10, indicating that more than 10% of the variance of these variables can be located at classroom level. For identification with Europe, around 4% of the variance lay at classroom level.

Table 1 Descriptive statistics of variables

These findings indicate three things. First, identification with Europe varies much more at individual level than at classroom level. Second, the same is true for learning opportunities about Europe. This is surprising as students ostensibly experienced the same lessons; yet they had very little shared perception of being taught about Europe. Third, students are clustered in certain classrooms according to their socioeconomic status and their immigration history. This might be connected to catchment areas and to the tracked nature of the German secondary education system (lower academic, i.e. vocational, track versus higher academic track). Students with immigration histories are more likely to attend schools within the lower academic tracks. In North Rhine-Westphalia, students with an immigration history are overrepresented in the lower academic tracks and underrepresented in the higher academic track (Kemper & Weishaupt, 2016).

Multilevel analyses

The results of the two-level analyses are listed in Table 2. The first model included the predictor variables at student level (Level 1). Gender and SES were not significantly related to identification with Europe. Students with non-European immigration histories and students who followed a marginalised religion reported a weaker identification with the EU. Learning about Europe predicted a stronger identification with Europe. The proposed model explained about 8% of the variance of the variable identification with Europe. Overall, our analyses support H1.1 and H1.2; but not H1.3. Where one or both parents of students were born in the EU, students identified more strongly with the EU. Further, students who perceived more learning opportunities about Europe at school also reported a higher identification with Europe.

Table 2 Multilevel Regression Analyses of Identification with Europe

The second model predicted identification with Europe at the individual and at the classroom level and used opportunities to learn, proportion of students with non-European immigration histories in a class and the socioeconomic status composition in a class as predictors.

Model 2 in Table 2 reveals that the shared perception of learning opportunities was a strong significant predictor of identification with Europe at classroom level. Conversely, a high proportion of students with low SES predicted lower identification with Europe. This model explained about 44% of the variance at classroom level, i.e. only about 1.5% of the overall variance. The classroom SES composition as well as learning opportunities were relevant to explain the limited classroom-level variance of identification with Europe. Therefore, the analyses support our hypotheses H2.1 and H2.3. The proportion of students with a non-European immigration history in a class was only a significant predictor for classroom level variance of identification with Europe when SES was left out of the analysis (Model 2). As soon as SES is added into the second model (Model 3), proportion of minority students stops being a significant predictor of identification with Europe. This shows that in a class, SES plays a greater role to explain the variance in identification with Europe than immigration history. On the other hand, it is important to note that at the individual level, this was not the case and SES was not significant, whereas Non-European immigration history was. Additional analyses showed that the predictors did not correlate at classroom level.

Model 4 in Table 2 introduced a mediation analysis at both levels. Learning opportunities were introduced as mediators between immigration history status and identification with Europe. Immigration history did not predict learning about Europe at either level. Consequently, there was also no indirect effect of immigration history over learning on identity. We therefore do not find support for H3.1. Students with non-European immigration histories experienced the same amount of learning opportunities as their peers. On the other hand, SES was able to predict learning about Europe on individual level, but not at classroom level (Model 5). Therefore, there is a partial support for H3.2. It means, pupils who have a lower SES compared to their classmates report fewer EU learning opportunities and, as a result, convey a lower identification with Europe. We find little support for the hypothesis that biographical or classroom composition variables impact students’ access to learning (H3.3).

Discussion

This study aimed to investigate which factors influence the identification with Europe of 8th-grade students in Germany at individual and at classroom level. Our results show that having at least one parent born in the EU and learning about Europe at school positively influenced the feeling of identification with Europe. In contrast with previous research, neither gender nor socioeconomic background significantly predicted students’ identification with Europe. Further, we were unable to find evidence of a mediational effect for learning about Europe on the classroom level. Access to learning opportunities about Europe did not seem to be unequally distributed among students with and without non-European immigration histories. These findings are in line with similar research conducted in Belgium using items inspired by ICCS 2016 (Brummer et al., 2022). Therefore, we assume that our results are not a result of specific circumstances of the education system in North Rhine-Westphalia. We find a weak indirect effect of SES on identity via learning on the individual level, but not the classroom level. On the classroom level we found a strong direct effect of classroom SES on identification with Europe. Similar research has shown that SES classroom composition plays a role in access to participation at school and a higher the average SES in a class, the more likely that students participate in civic activities (Deimel et al., 2020). Therefore, these findings indicate not only possible differences in access to education and participation in German schools according to pupils’ socioeconomic background but also a clustering phenomenon in German schools. As a solution for this, Agirdag et al. (2012) suggests the implementation of a “desegregation” school policy, to achieve more SES and ethnic diversity in classrooms. It is not a surprise that students who have immigration histories from outside Europe have a weaker identification with Europe since the negative correlation between immigration history and identification with Europe was shown by previous research (Agirdag et al., 2012; Fligstein, 2009; Jugert et al., 2019; Ziemes et al., 2019).

Some limitations to our findings must be noted. As previously mentioned, our study lacks the data regarding the citizenship status of the students in relation to EU countries. We only have data about where their parents were born. One important aspect emphasized by Diez-Medrano (2010) is that the perception of something as "real" is closely linked to an individual's identification with it. In the context of our study, the lack of data on the students' citizenship status in EU countries hinders our understanding of how their identification with Europe is influenced by the tangible aspects of EU membership. Citizenship confers a sense of reality to the European Union, as it entails rights, privileges, and a legal status within the supranational framework. However, given this limitation, we have employed an alternative measurement approach of considering students' immigration history. By incorporating immigration history from EU or non-EU countries as a measurement variable, we are able to draw conclusions that contribute to a more comprehensive understanding of the influence of individuals' parental backgrounds on their identification with Europe. Another point is that we present cross-sectional analyses. Learning opportunities and identity were measured at the same time; for this reason, we cannot empirically show that learning about Europe influences identity formation. That said, this assumption is plausible based on the theory and analyses that we reviewed in the theory section. Further, we focused our analyses on one grade in one education system. Further studies need to investigate the generalisability of our results to other age ranges and samples. Moreover, we did not examine the relationship of identification with Europe with national identity; this can be relevant, especially for students with immigration histories. Young people with immigration histories can be pushed away from national identities and pushed towards collective, supranational identities (Erisen, 2017), and this question should be addressed in further research. All that said, one of the strengths of our analyses is the multilevel approach that allowed us to estimate the relative proportions of variance at individual and classroom levels. We were able to explain a relevant portion of the admittedly limited classroom-level variance of identification with Europe. The fact that students with non-European backgrounds in the sample were clustered in certain classrooms could be interpreted either as a problem or as an opportunity; an opportunity, in that the situation might enable schools to reach those students more directly with special classroom projects and activities designed to foster a identification with Europe.

It is worth noting that our results show that opportunities to learn about Europe help to strengthen students’ identification with Europe, supporting the culturalist model of identity formation which suggests that exposure to cultural elements associated with Europe, such as education about its history, values, and traditions, plays a crucial role in shaping individuals' sense of European identity (Recchi, 2014). Our results are in alignment with the findings of previous research using data of ICCS 2009, which concluded that cognitive learning opportunities about Europe are strongly related with identification with Europe (Verhaegen et al., 2013). Consequently, those who have a better understanding of the EU are more likely to hold a “European identity” and to be predisposed favourably towards it (Ceka & Sojka, 2016). Schools could implement syllabi and actions to show students who are affiliated with marginalised religions that they, too, can be and are a part of Europe. A multicultural dimension in the curriculum that encompasses topics such as cultural and religious diversity can have a stronger impact on the European identification of students with non-European backgrounds than a purely European dimension (Brummer et al., 2022). These findings have broader implications for other supranational entities. Results suggest that implementing educational initiatives that focus on relevant topics related to regional or supranational organizations could potentially enhance individuals' identification with them. Classroom learning about Europe is a strong predictor of identity; however, students within the same classroom only agreed vaguely with each other as to whether they learned something about Europe. One solution would be for teachers to emphasise clearly when and how a lesson relates to Europe and the EU. This way, students might absorb the lesson’s topic more readily and might be more likely to use the information to explore their own identification with Europe. Overall, our results show that there is limited variance at classroom level, but that identification with Europe is still something individual for students in Germany. The results underline the relevance of learning opportunities.