The Educational Migration of Sub-state Ethnic Minorities on the Outskirts of the EU: A Case Study of Serbia

The goal of this paper is to present an analysis of the status and career choices of Hungarian, Slovak, and Romanian ethnic minority high school graduates in the Autonomous Province of Vojvodina, the northern province of Serbia. We implemented a purpose-built, paper-based-questionnaire process of data collection that involved 2,192 ethnic minority high school students who were finishing high school in their mother tongues in 16 municipalities in Vojvodina. The results of the analysis showed that almost 40 per cent of the sample of ethnic minority high school graduates planned to leave Serbia to study in their mother tongues in nearby European Union (EU) countries. While this brain drain is not a new trend, our research shows that there are new and different reasons for it and it was caused by the insufficiency of Serbian language skills of the ethnic minority studnets. The results of the research show that the main reason for this educational migration is thus the aspiration to obtain a diploma from an EU-based institution, which (1) grants immediate access to the EU’s integrated labor market, and (2) is perceived to improve social status compared to that of remaining in Serbia after graduating. Despite this new motive for educational migration, language barriers are still relevant determinants of the latter. A gender-sensitive analysis of the responses was undertaken that showed that there are gender-based differences in terms of which students continue their studies after high school, and in the reasons for staying in Serbia.


Introduction 1
Vojvodina, the northern autonomous province of Serbia, is known for its stormy past and changing borders due to historical events. 2 Vojvodina is not a territorial autonomy on ethnic grounds, but a multi-ethnic region with a clear (and growing) Serbian majority (Székely and Horváth, 2014: 434). Its autonomy is largely geographical and socio-political. It is home to sizeable Hungarian (13 per cent, ca. 250,000), Romanian (1.3 per cent, ca. 25,000) and Slovak (2.6 per cent, ca. 50,000) ethnic minority communities (Statistical Office of the Republic of Serbia, 2012). 3 When it comes to defining national structure, following Kymlicka's (2007) classification of three general types of minorities in Western democracies (indigenous, sub-state, and immigrants), we consider the above-mentioned ethnic minorities as sub-state 'national groups'. These national groups or national minorities are caught between two mutually antagonistic nationalisms -those of the nationalizing states in which they live, and those of the external national homelands to which they belong by ethnonational affinity, though not by legal citizenship (Brubaker, 1996: 5). In this triadic nexus (Brubaker, 1996), members of these Hungarian, Slovak, and Romanian ethnic minorities are in a challenging position: the main bodies of their nations live in nation states inside the European Union (EU), while they are 'locked out' in Serbia, a non-EU state. The educational system of Serbia allows the young members of these ethnic minorities to finish their primary and secondary education in their mother tongues (Hungarian, Romanian, and Slovak). After high school, most of them need to decide between continuing their studies in the predominantly Serbian-language higher education system in Serbia, or leaving Serbia and continuing their studies in their mother tongues in neighboring/nearby EU states. If they choose to study in Serbian, they need to master the Serbian language, as without proficiency in Serbian they often face a language barrier that makes their participation in classes and examination difficult (Lendák-Kabók and Lendák, 2017). In a recent study that was conducted with the aim of improving multilingual education in Vojvodina, it was shown that ethnic minorities consider knowledge of their mother tongue to be most important, subsequently followed by knowledge of the majority language. The former assessed their knowledge of their mother tongues to be better than their knowledge of the majority Serbian language (ibid.). In their struggle for better employment opportunities, a higher salary, and improved social status, about onethird of high school students from the analysed minorities decide to leave Serbia and study abroad, predominantly in Hungary, Romania, and Slovakia (Takács, 2013;Puja and Badesku, 2009;Šimonji and Černak, 2006).
In relation to the Hungarian ethnic minority in Vojvodina, their countinous educational migration represents a brain drain that has been ongoing for several decades (Gábrity-Molnár, 2009). The arrival into Hungary of qualified professionals with a Hungarian mother tongue and culture from across the border is for Hungary an often unacknowledged benefit, but it represents a loss for the source Hungarian communities (Váradi, 2013: 97). Previous research has identified that Hungarian national policy with its associated instruments is unable to fulfil its most important purpose, which is to safeguard the existence of the intellectual elite within the Hungarian minority community, and help them sustain an adequate livelihood in their country of birth (Erőss et at, 2011: 4).
The situation is similar for the Romanian and Slovak ethnic minority in Serbia as well, who also engage in significant outward educational (and other) migration (Puja andBadescu, 2009, Zlatanović andMarušiak, 2018) In nationalising states, minorities may feel that the state they are living in is structured and governed in favour of the majority, and that their identity is under threat (Kemp, 2006: 111). Research into the Romanian minority community in Vojvodina has shown that members of minority communities often feel that despite their (legal) equality with the majority and their freedom to use their language in schools, in media such as the press, as well as in public life, they still feel a sense of inferiority that is based on the idea of belonging to another nation (Maran, 2013: 232). In this sense, members of the minority community feel the need to turn to their kin-state. The act of educational migration to external national homelands or kin-states (Kemp, 2006) causes the gradual disappearance of ethnic minorities in their homelands. The goal of this paper is to describe the educational migration of the three above-mentioned ethnic minorities to the EU. The study also explains the causes of educational migration and the low majority language skills of ethnic minorities. Additionally, the study puts specific emphasis on the aspect of gender and identifies the key gender differences behind the career decisions of ethnic minority women and men. The study is important from a Central and Eastern European (CEE) perspective as it gives insight into how the continuous brain drain of ethnic minority students inevitably leads to the weakening and disappearance of the above-mentioned ethnic minority communities.

Continuing education and the educational migration of ethnic minority students
Earlier studies have suggested that non-academic factors, particularly socioeconomic background, affect graduates' post-secondary destinations. (Foley, 2001) For example, students from lower-income families were found to be particularly likely to attend less selective institutions, regardless of their level of academic ability, achievement, and aspirations. (Muskens et al, 2019) Educational migration can be a form of transmigration that takes place when young people move from their home country to another country for the sake of better education (Badikyan, 2011). Such migrants aspire to eventually obtain better career prospects, a higher salary, new professional experiences, and contacts, motivating and interesting jobs, and/or better training facilities (Stukalina et al., 2018). Skill acquisition plays an important role at many stages of an individual's migration (Dustman and Glitz, 2011). Researchers argue that educational migration is unfavorable for the country of origin, as it produces brain drain, which has an economic and social impact (Tremblay, 2005;Lien and Wang, 2005;Teferra, 2005;Váradi, 2013). Others claim the opposite, highlighting the positive effects of a 'brain circulation' that creates knowledge networks, enriching all parties that are involved (Özdenand Schiff, 2006;Walker, 2010;Wooley et al., 2008).
Educational migration from Serbia to Hungary, with a focus on higher education, has become one of the most significant types of migration and is determined by both ethnic and economic factors (Takács et al., 2013: 78). Research conducted by Takács et al. indicated that this kind of educational migration is not transnational (i.e. involving resettlement, return, move to a third country, circulation), but that, in the majority of cases, migration for education constitutes the first step towards permanently leaving the country of origin (ibid.). In Vojvodina, about one-third of ethnic minority Hungarian, Slovak, and Romanian high school students leave Serbia and study abroad, predominantly in their motherland (Takács, 2013;Puja and Badesku, 2009;Šimonji and Černak, 2006), causing a severe brain drain from Serbia, which has been ongoing for several decades (Gábrity-Molnár, 2009). Regarding the effect this has on the kin-statetaking Hungary, for example -, Váradi (2013: 97) states that the former profits since it benefits from educational migrants who speak the Hungarian languagelanguage proficiency being the most commonly required skill of migrants (Raghuram, 2013: 140). The privilege of having such language skills eases social and cultural integration (ibid.). Accordingly, migration for emigration is considered as primarily a form of elite migration, confined for the most part to the upper and middle classes (Brubaker, 1996: 157). When it comes to the Romanian ethnic minority in Vojvodina, most parents enroll their children into the Serbian school system on the assumption that better social integration and obtaining Serbian language skills will make it easier for their children to continue their education (high school and university) (Puja and Badescu, 2009). Such decisions surely have their advantages, especially with regard to learning Serbian terminology in subjects such as physics, chemistry, mathematics, etc. On the other hand, in such settings the native (Romanian) language is not taught and cultivated further; consequently, the Romanian language skills of ethnic minority Romanians remains at the level of the preschool period (Puja and Badescu, 2009). Paradoxically, when Romanian high school graduates continue their higher education in Romania (which has also been a trend in recent years), the enrolment of these children in Serbian-language classes in Vojvodina is by no means an advantage (Puja and Badescu, 2009).
Vojvodinian Slovaks study in Slovakia as foreign students with special benefits (Zlatanović and Marušiak, 2018). Many of them receive scholarships from the Slovak government that cover tuition costs that would potentially not be covered by Serbia. The interviewees who participated in research undertaken by Zlatanović and Marušiak (2018) generally gave pragmatic reasons for their decision to study in Slovakia -for example, after graduation they would get a job there, and with a diploma earned in the EU they could continue their education or seek employment in a more developed country. Some of them were attracted to specific study programs that universities in Serbia did not offer (Zlatanović and Marušiak, 2018).

Gender differences in Continuing Education and Education Migration
In the past continuing education has been reserved for boys, who usually received more resources from their parents to pursue their education compared to girls, either because parents expected higher economic returns from their sons' education, or because of traditional gender roles, according to which women do not need, or even should not have, a formal, advanced education (Schneebaum et al., 2015: 240). in even very patriarchal societies such as the Nepali one, parents started supporting grils' education. The latest trends show that in private institutions by investing significant resources into their education (Adhikari, 2013), but typically only in traditionally 'female professions' such as nursing programs.
In today's Europe, there are considerably more girls studying at both the bachelor's and master's level than boys (Eurostat, 2016). Courses in which female students traditionally dominate are increasingly being changed from college courses to three-year undergraduate courses (Dingwall, 2016). Recent years have also seen a focus on campaigns that encourage female students to study science, technology, engineering, and mathematics (STEM) courses (Dingwall, 2016).
Studying in a foreign country might be appealing to women as they may see it as a way of exercising freedom and independence, as opposed to being under surveillance by their parents and other members of society at home (Dhungel, 2019: 362). Opportunities for student mobility are especially important for women (Juhász et al., 2005: 177), although the level of participation of female staff and faculty mobility at university decreases after graduation, usually due to family and care responsibilities (Juhász et al., 2005). The Erasmus program student has significantly enhanced mobility and opened up opportunities for educational migration in Europe. In 2000-2001 per cent of all Erasmus students were female (Juhász et al., 2005: 177). The decision to go abroad may be affected by language skills, which play an important part in the considerations that precede a decision to go abroad (Juhász et al., 2005).
Gender differences in educational migration in the Carpathian basin among ethnic minority students have previously received little scholarly attention. The only gender differences that have been recorded among ethnic minority students is that girls are reported to have a slightly greater desire to continue their studies than boys (Papp, 2003).

Minority language education and the language barrier
Although linguistic diversity remains entrenched as one of the key values of the European project, the EU has relatively limited competence to manage this issue (Kraus and Kazlauskaite-Gürbüz, 2014: 524). In this sense, every country autonomously deals with the issue of minority languages.
Language may be important to a group at a symbolic level (Jaspal, 2009: 18) and is strongly connected to one's ethnic identity. If language is separated from cultural context, it does not remain a community language; only functioning as such can its existence be assured (Dołowy-Rybińska, 2015: 260). Language is an important component of maintaining a separate ethnic identity. Attempting to establish a cohesive national identity may be detrimental to minority groups' identity, since an important component of self-identity, namely language, is often at stake (De Vos, 1995). The official language is bound to the state and 'integration into a single "linguistic community," which is a product of the political domination that is endlessly reproduced by institutions capable of imposing universal recognition of the dominant language' (Bourdieu, 1991: 46). When it comes to minority groups, education in one's mother tongue is crucial for the preservation of national identity. The importance of education in one's mother tongue is increasingly emphasized in multilingual regions, as teaching and learning in one's mother tongue improves learning outcomes in primary schools (Seid, 2016). Mother tongue education plays an important role in preserving the mother tongue, as in addition to the linguistic dimension, it has an ethnic dimension. The 'imagined community future' also requires the use of the mother tongue (Papp, 2015: 51). National minorities tend to seek some form of institutionalization of mother tongue education, and use their own language at school as a means of preventing or reducing language assimilation (ibid.).
With regard to the different ethnic minorities in Europe, Swedish-speaking Finns have no clear incentive for a 'mother country'-type of affiliation to Sweden have typically struggled after having 700 fewer hours of Lithuanian-language lessons than their counterparts from Lithuanian schools (Barwiński and Leśniewska, 2014). The Slovak community in Hungary has 'Slovak national education'; the term refers to a network of schools that educate members of the Slovak community in Hungary and aim to preserve and develop the Slovak language (Šimonji and Černak, according to Ondrejovič, 2006). However, this has failed to avert progressive linguistic assimilation in the family environment, weak national awareness, deficiencies in the system of learning the Slovak language at an earlier period, and the dominance of the Hungarian language in the modern school system, but has increased interest in learning foreign languages (Šimonji and Černak, according to Homišinová, 2006). Minority language education is sometimes associated with later risks for minority students. Filipović, Vučo and Đurić assesses this issue in Serbia, stating that the teaching of all classes in primary and secondary education in a minority language, with only a few hours per week of Serbian as a majority language, has created generations of unbalanced bilinguals, characterized by a very low level of Serbian proficiency (Filipović et al., 2007). In the first and second year of elementary school there are 72 classes of Serbian for minority students, and 180 classes of Serbian for majority students, while in the third and fourth year of elementary school there are 108 Serbian language classes per year for ethnic minority students. 4 In addition to the low number of majority classes in school, young people living in the north of Bačka, where the majority of Hungarians live in Serbia, have typically not wanted to learn the state language since the Yugoslav wars of the 1990s (Nađ, 2006: 448). 5 Consequently, they face difficulties during their post-secondary education in Serbia, and/or struggle to find satisfactory employment, which by default requires knowledge of Serbian (Filipović et al., 2007). These ethnic minority students first need to overcome the language barrier (i.e., to learn Serbian) when starting their studies, which requires time, effort, and sacrifice (Lendák-Kabók, 2014). Minority students around Europe also experience different challenges after finishing high school (Khattab, 2018); the above-discussed minorities are not exceptional in this sense. In Britain, members of (immigrant) ethnic minority communities demonstrate high levels of motivation and enrollment in higher and further education (Lessard-Phillips, 2018). The situation is quite different in Italy, where only a minority of non-Italian secondary school graduates continue their studies at the university level compared to their Italian peers, (Mantovani et al., 2018). A study of Turkish and Roma ethnic minorities living in Bulgaria showed that they had lower education levels and lower employment rates than their Bulgarian majority peers (Trentini, 2014 Minority language education can be analyzed at either the macro (government) or micro level (family). Although bilingual education might be the government approach to preventing/lessening the language barrier, there are a wide variety of often conflicting ideologies, theories, policies, and practices surrounding bilingual and multilingual education throughout the world (Wright et al, 2015). In some cases, bilingual education may even be misused to limit the access and opportunities of linguistic minority students (Wright et al, 2015) -this has been the case in Ukraine, for example, affecting the Hungarian minority population (Papp, 2017). Some parents enroll their children in majority-language schools, as they feel that a Hungarian-language education does not provide as many opportunities as Ukrainian-language education (which does allow for optional Hungarian lessons). In this way, children 'learn to read and write in Hungarian, too,' but their competences develop in the official (majority) language as well, which is important if they want to 'exist,' prosper, and build a career at home (Papp, 2017: 97). According to data from a survey published recently, the majority of those participating in state-language education are born in interethnic (mixed) marriages (Dombos, 2011). Intermarriages, the level of education of parents, and socio-economic status affect majority-language school choice (Papp, 2017: 92).
Based on the above-mentioned theoretical background, the authors identified the following hypotheses: H1: There are gender differences in the career choices of ethnic minority high school graduates (i.e. in whether they aspire to complete higher education).
H2: There are gender differences with respect to the country in which the ethnic minority students under examination plan to continue their education (i.e. in Serbia or in a nearby EU Member State).
H3: The language barrier has a significant impact on the choice of country of further education (i.e. in Serbia or in a nearby EU Member State).
These hypotheses are relevant, as the shift in motivation for the educational migration of ethnic minority students finishing their secondary education in Vojvodina and its gender aspects has not been previously researched. Additionally, no detailed research has investigated the interplay between the language barrier (i.e. a lack of majority [Serbian] language skills) and the above-described educational migration to nearby EU countries, nor the gender aspects of that trend. This research focuses on three sub-state national groups: Hungarian, Slovak, and Romanian. This choice was made because, on the one hand, those groups are quite numerous (Hungarian is the most populous minority group, Slovak is the second, and Romanian is the fifth), 6 and on the other hand, those national groups can access elementary and secondary education in their mother tongues in Serbia.

Method
In this section an overview of the data collection and analytical methodology, as well as the ethical aspects of the research, will be presented.

Sample, procedure and ethics
The total sample was determined using the quota sampling method (Biljan-August et al., 2009) and consisted of 2,192 high school students: 1,119 females (51 per cent) and 1,073 males (49 per cent). The average age of the respondents was 18 years. Ethnic membership structure was the following: Hungarian 1,951 (89 per cent), 175 Slovak (8 per cent), and 66 Romanian (3 per cent). These students were in their final year at high school, a critical period for making decisions about their future, career choices, and further education.
The pupils were finishing high school in their mother tongues. The entire population of the ethnic minority Hungarian, Slovak, and Romanian high school graduates in the 2013/14 and 2014/15 school years numbered 3,311 8 students in Vojvodina; therefore, the sample covered 66.20 per cent of all students. The research was conducted in two consecutive school years; specifically, in May 2014 and in April 2015, as these were the last months of the students' secondary education before their graduation in June. It should be noted that the research focused on students who studied in ethnic minority languages, while those who were enrolled in majority language schools were not included.
After obtaining the principals' approval, the researcher attended one of the classes during which students filled in the paper-based questionnaires. All three types of high schools that operate in the languages of the three ethnic minorities were included in the research -namely, three-year vocational, four-year vocational, and four-year grammar schools.

Instruments
For the purpose of the study, a questionnaire was built, consisting of 20 questions that can be grouped into the following themes of interest: Demographic information consisting of gender, ethnicity, school, and school year (i.e. third or fourth grade); mother tongue of respondents (i.e. Hungarian, Slovak, or Romanian); self-assessed Serbian (majority) language skill level; plans for continued (i.e. higher) education; location of chosen higher education institution (Hungary, Slovakia, Romania, or Serbia), as well as thoughts about the career choices women have in pedagogy and/or in technical fields.
Most questions were closed-ended, multiple choice format, but eight openended choices and one open-ended question were included. Respondents were asked to fill in by hand the name of the chosen higher education institution. Selfassessed knowledge of the Serbian language was measured using a seven-point Likert scale that ranged between 'not at all' and 'very good,' and the frequency of use of the Serbian language with a five-point Likert scale ranging between 'less than once a week' and 'several times a day.'

Data analysis
The data analysis was performed using IBM's Statistical Package for the Social Sciences (SPSS) version 16 for Windows. Given the fact that the majority of variables were defined as categorical data, nominal, cross-tabulation analysis, and Pearson's Chi-Square test were mainly utilized. In the case of continuous variables (e.g. assessment of knowledge and language-use frequency), one-way ANOVA was used to analyze the differences among group means.

Ethics
Use of the paper-based questionnaires was authorized by the principals at each secondary school involved in this research. Students were informed about the goals of the research, and given the option to opt out from filling in the questionnaire. Questionnaires were anonymous, i.e. no names, addresses, or other personally identifiable information of students were recorded. Only the location, name, and type of secondary school were recorded, as this information was relevant in the context of the research.

Results
In the next sections, the authors analyze the education migration patterns of Vojvodinian ethnic minority students. We present the results of our gendersensitive analysis of (1) students' aspirations to continue (not continue) education, (2) the extent, destination, and motivation behind educational migration to nearby EU countries, as well as (3) the interaction between the language barrier and educational migration.

To study or not to study?
In Section 2, the first hypothesis was formulated (H1) around the notion that there are gender differences between the career choices made by ethnic minority students finishing secondary education in Vojvodina, the northern province of the Republic of Serbia. As elsewhere in Europe (Eurostat, 2016), the sample showed that more girls who are finishing secondary education (83.6 per cent) aspire to continue their studies compared to boys (73.1 per cent), a finding which is in line with the results of previous research done in Vojvodina among ethnic minority Hungarian students (Papp, 2003). We can mention here that a significant percentage of boys and girls do not want to continue their education at all, which fact may be connected to their parents' education level, which has been hypothesized to influence young people's choice of whether to pursue postsecondary education (Foley, 2001).
The reasons given by students who did not wish to continue their studies after secondary education were examined as well, and we found that there were significant gender-related differences in these (ᵡ²(6)=29.70; N=673; p<0.001). Most students who did not plan to continue their education wanted to find a job. As shown in Table 1, male students more often agreed with the statement 'a high school degree is sufficient.' This outcome is markedly different to the situation with other ethnic minorities in the European geographic context, especially firstor second-generation immigrants, who usually understand the need for and aspire to complete higher education (Lessard-Phillips et al., 2018).
Very few students stated that they did not want to enroll in a higher education institution because of the language barrier, which would have indicated that they do not have adequate Serbian language skills. There were no significant gender differences between boys' and girls' answers regarding language skills.
Almost two-thirds (63.6 per cent -rows 1 and 2 in Table 1) of the girls who were not planning to continue their higher education claimed that they wanted to get a job or had financial difficulties. Although in absolute terms the number of girls and boys who referred to financial difficulties was quite similar, more than one in five girls who did not aspire to complete higher education claimed that their lack of motivation was caused by some form of financial barrier. The relatively high number of ethnic minority students stating that they wanted to get a job or that financial difficulties barred them from higher education points to class-related issues, considering the fact that both state-funded higher education and subsidized public transport are quite affordable in Serbia (based on the authors' own experience with commuting and obtaining higher education degrees without the need to pay any tuition fees).

Where to study?
In order to verify our second hypothesis (H2), we investigated the gender aspect of the long-standing educational migration of ethnic minority students who are finishing their secondary education in Vojvodina. Our goal was to determine the extent and causes of the educational migration of ethnic minority students to nearby EU countries. Therefore, we first analyzed where the respondents planned to continue their higher education. Findings are shown in Table 2. 684 students (out of 2192) intended to study abroad only, which is a staggering 39.6 per cent (see row 2 in Table 2 for details). The finding that almost 40 per cent of ethnic minority pupils who were finishing high school in Vojvodina planned to study abroad is higher than that identified in previous studies (Takács, 2013). Migration for education would not be an issue if the students returned to Serbia after completing their studies, but multiple decade-long trends show that most of them will not . Additionally, a smaller number of respondents (from rows 3 to 5 in Table 2) also plan to leave Serbia if they gain entry to a higher education institution abroad. Statistical analysis of the data showed significant gender differences in the career aims of respondents (ᵡ²(3)=9.18; N=1728; p<0.05). We found that female students are more determined about their career choices, as the total number of undecided girls answering 'Do not know yet' or not answering at all was slightly lower than that of boys. We found that more girls desired to continue their studies in Hungary, Slovakia, or Romania (see row 2, Table 2) than boys. If we add to these numbers those students who will try to gain entry to an institution of higher education in both Serbia and another country but will continue in the country of their ethnic origin if accepted, the proportion is even larger. It is important to note that there were no gender differences in favor of boys as regards pursuing education abroad, which may represent a greater financial burden on parents. Therefore, we conclude that parents were not expecting a higher economic return on their sons' education (Schneebaum et al., 2015).
Although the questionnaires were filled in only one or two months before the deadline to submit applications to a higher educational institution either in Serbia or abroad, a greater number of boys (16.2 per cent) were undecided. This might have a detrimental effect on the future prospects of these students and their social groups as well, as they will either not continue their education, thereby putting themselves in an inferior position in the Serbian labor market, or make uninformed and rushed decisions at the very last moment. Fortunately for the male students, even if they do make a decision at the very last minute, the chances are high that they will choose a more 'masculine' track (technical college or university) and thereby be in a better starting position after graduating. The undecided girls (12.2 per cent; see row 4 in Table 2 above) might not fare as well. Due to their general inclination towards the social sciences and humanities (SSH), in combination with hastily made decisions, they may end up in areas with a lower number of vacancies and lower average salaries.
Next, we analyzed why students chose to stay and study in Serbia, and here we also found gender differences (ᵡ²(5)=16.93; N=1015; p<0.005). Our findings (Table 3)  in their opinion, the Serbian higher education system will give them a high level of knowledge. Male students opt to study in Serbia because of financial problems more often than female ones, as well as because of ties with friends. On the other hand, female students opt for Serbia because they think they will succeed more easily with a Serbian diploma at home. We also examined why students choose to study outside Serbia. Results are shown in Table 4. Somewhat surprisingly, we did not find significant gender differences between the answers (x²=2.06(4); N= 887; p=0.724). The main reason for leaving Serbia to study in a nearby EU Member State was, for both male and female students, the desire to obtain a university degree from a country within the EU. Earlier studies typically identified language barriers as the primary source of motivation for migration. This reason seems to have been superseded since Hungary, Slovakia, and Romania joined the EU. However, our analysis showed that language barriers were still a major source of concern, as this factor was quoted as the second most frequent reason for leaving.

Language barriers
Students from ethnic minority communities have limited higher education options in their mother tongue in Serbia (namely, the departments for Hungarian, Slovak, and Romanian language and literature at both the University of Novi Sad [UNS] and at the University of Belgrade, the state-funded Hungarian Language Teachers Training Faculty in Subotica [part of the UNS] and a number of courses taught in Hungarian at the Faculty of Economics, also in Subotica). Partial or complete Hungarian-language tuition is offered outside the University of Novi Sad in Vojvodina at the Subotica Tech College, and at the privately-owned Educons University. Slovak-and Romanian-language tuition is also offered at the Faculty of Philosophy in the native language departments at the Teacher Training Faculty in Sombor. There was also once a Study Program in the Slovak language in Bački Petrovac, but this has since closed due to an insufficiency of staff. In Bački Petrovac, the St. Elisabeth University of Health and Social Work Bratislava, Slovakia, EU, is operated by the Slovak government. At the University of Belgrade, there is a study program called Professor of Classroom Teaching that is taught in Romanian in Vršac. Such a limited choice of courses might explain why respondents stated that their insufficient knowledge of the Serbian language was the second most frequent reason for leaving Serbia to study abroad (see row 2 in Table 4). That lack of Serbian language proficiency might be partly explained by the Yugoslav civil wars, which led to increased nationalism and segregation, even in Vojvodina, which was spared armed conflict apart from the NATO bombings in March to June 1999. Because of the above-mentioned segregation, children from ethnic minorities do not often need to master the Serbian language before finishing their secondary education, especially in municipalities in which they are the majority ethnic group. We decided to investigate the causes and effects of this language barrier with our questionnaire. Ninety-seven per cent of respondents indicated that their ethnic minority language was their mother tongue. A relatively small number of students explained that they had two mother tongues, typically the ethnic minority language and Serbian. Most of these exceptions involved intermarried families.
We asked students how often they speak Serbian each week. More than 30 per cent (rows 2 and 3 in Table 5) of our respondents used Serbian no more than once a week. That means that some students managed to get by almost without ever using the majority language. 9 There was no significant gender-related difference in the frequency of majority language use. In the questionnaire, students were asked to self-assess their skills in both their mother tongues and in Serbian. Results are shown in Table 6 (below). A majority of the ethnic minority secondary school graduates spoke good, very good, or excellent Serbian (56.9 per cent of girls, and 52.9 per cent of boys); a little more than half of all respondents. The language barrier is greater for male students, meaning they are more likely not to opt for higher education at all (as shown in Table 1, where male students more often chose the option 'a high school degree is sufficient' compared to female ones). The self-assessed Serbian language skills of the rest of the respondents (around 45 per cent) were inadequate to permit them to enroll in a higher education study program in Serbian. The level of Serbian language proficiency appeared to influence where respondents were planning to continue their education. In order to examine this proposition, a one-way ANOVA with a Bonferroni post-hoc test was used (the dependent variable was the frequency of use of the Serbian language -see the Y axis in Figure 1, and the grouping variable was the location of the chosen higher education institution [Serbia, EU Member State, etc.]).
Figure 1 -Frequency of use of Serbian language and country of further education Significant correlation was identified between the frequency of Serbian language use and the intended country of further education (F(4.2190)=8.94; p<0.01). Participants who used the Serbian language less frequently were more liable to want to study outside Serbia. This confirms our third hypothesis (H3); namely, that a lack of Serbian language skills significantly impacts the career choices of ethnic minority secondary school graduates, and is an important driving force behind the education migration faced by their communities.

Conclusion
We have presented research into the status and career choices of Hungarian, Slovak, and Romanian ethnic minority high school graduates in Vojvodina. We were particularly interested in the higher education aspirations of the target group in terms of the desired destination (Serbia or neighboring/nearby EU Member States, which represent their ethnicity-based homeland) and the effect of the language barrier on outward educational migration towards nearby EU states. We found that almost 40 per cent of ethnic minority high school graduates intend to leave Serbia to study abroad. If we add to this figure those who will try to pass an entrance exam in both countries (for example, in both Serbia and Slovakia) and whose aspiration is to move to the neighboring country if successful, the number is even higher. As most emigrants will not return , this represents a significant brain drain that will affect the ethnic minority communities and Serbia as well. We have shown that language barriers may no longer be the most important driving force behind the multiple-decades-long educational migration of ethnic minority students living in Vojvodina. According to our findings, the most important reason for ethnic minority students to leave Serbia is the pull of a higher education degree from an EU-based institution, which grants them immediate access to the EU's integrated labor market, and the potential for increased social status. The language barrier is however relativized, as there is strong overlap between the economic and linguistic elements of the phenomenon under analysis. The second reason for students leaving their country of origin is the language barrier; i.e. a lack of Serbian language proficiency that hinders entry into the Serbian higher education system, in which most study programs are available in the Serbian language only. Reasons for leaving the country students were born in are also likely to be linked to social status and empowerment. Respondents felt that obtaining a higher education degree in an EU-based institution might allow them to earn more and obtain higher social status than their peers who stay at home. Additionally, if the former continued their studies in Serbia, the language barrier would place them in an inferior position compared to their Serbian peers. This inferiority might even extend to their entry into the labor market in Serbia in which certain professions are reserved for well-integrated individuals with excellent Serbian language skills.
We identified significant gender differences in educational aspirations and migration patterns. First, our results confirmed previous research (Papp, 2013) that found that more girls who finish secondary education aspire to continue their studies compared to boys. In relation to not continuing their post-secondary studies, male students more frequently claimed that a high school degree was sufficient, while girls not planning to continue their education claimed that they wanted to get a job or had financial difficulties. We also found that almost every third girl wanted to continue her studies in a nearby EU country, whereas this proportion was significantly lower for boys. Boys faced greater language barriers (i.e. weaker Serbian language skills), which meant that they might not opt for INTERSECTIONS. EAST EUROPEAN JOURNAL OF SOCIETY AND POLITICS, 6(2): 130-153.
higher education at all. Additionally, we found that female students are more determined about their career choices: the total number of undecided girls (i.e. those not knowing where they would continue their studies) was significantly lower than that of boys. This lack of timely decision making might have a detrimental effect on the future prospects of both genders and their social groups. The latter will either not continue their education, thereby putting themselves in an inferior position on the Serbian labor market, or will need to make uninformed and rushed decisions at the very last moment.
As far as future research is concerned, the authors have identified two ways to potentially extend this research. First, they intend to 'close the circle' by extending this questionnaire-based study to the entire Carpathian basin and analyzing the status of different minorities in parts of Serbia, Hungary, Romania, Ukraine, Slovakia, and Croatia. Second, they intend to work towards implementing some related measures (e.g. career orientation for ethnic minority high school graduates) and subsequently repeat the questionnaire-based research to measure the impact of these measures.