Despite the fact that bi/multilingual (in the following the term ‘multilingual will be used to cover both ‘bilingual’ and ‘multilingual’) education has been researched intensively in the last decades, there seems to be a research gap in one particular but important area. This is a matter of how school administrators and leaders should approach multilingual school programme development? Several authors have touched this matter partly or from a limited angle. Firstly, May (2017) has synthesized principal axes of bilingual/immersion education outlining how bilingual education can be thought of in terms of three levels – philosophy (goals), model (specific aims), and programmes (more concrete categorizations). Secondly, several authors like Baker (2011) or García (2009) have provided typologies of bilingual education programmes to illustrate which approaches are considered to suit which target groups of students together with societal and educational aims of programmes as well as language outcomes aimed at. Thirdly, Mehisto (2012) has provided a detailed instructional overview for school principals on how to implement properly a bilingual education programme. Fourthly, de Jong (2006) has written about an alternative approach to bilingual education (calling it integrated bilingual education) and its implementation results. Furthermore, a lot of country case studies have been described in the field. Busch (2011, p. 548) has called for ‘developing tailor-made solutions for particular school environments, flexible enough to accommodate the sociolinguistic macro changes in society and the micro changes within the school community’ but schools or school owners are actually left at their own devices with how to do this. All in all, none of these studies, clearly enable to understand how should a multilingual programme be developed: where to start? which issues should be considered in which order? What steps should be taken? Yet, this is a question more and more leaderships can relate to as diversity and its management has become a reality in Europe and beyond. Moreover, it can be seen from these studies that multilingual education model development remains an implicit process. It is largely a ‘black box’ that cannot be explained or described. Therefore, based on the lack of academic literature on bi/multilingual model development, our article proposes a way of addressing this research gap. By outlining a step-by-step framework of analysis for schools struggling with diversity and coupling this with the analysis of effectiveness factors of multilingual education, we demonstrate how a programme could be reached at. Our article uses an example from Estonia to illustrate the model development in practice. More specifically, our article addresses the following research questions:

-          What does the multilingual school model development involve?

-          What could be one tailor-made multilingual school model development example in Estonia?

Several researcher of multilingual education (e.g. Baker, 1985; Beatens Beardsmore, 2009; Cenoz, 2009; Mehisto, 2015; Spolsky et al., 1976) have tried to summarize the factors or variables shaping the functioning of multilingual education. There have been different types of categorizations and typologies of factors. We are using a synthesized version of these coupled with the Bronfenbrenners’s ecological model (see graph below) that identified the factors at different levels (micro, meso, macro) as well as based on the nature of the variable (input, process, outcome) (Kirss et al., 2020). This approach is detailed enough to offer a close analysis of different types of factors while keeping the number of factors under control and offering a comprehensive overview. The framework is also beneficial in the sense that it allows analyzing consistencies in the system, i.e. if the factors are in line vertically and horizontally.

Since our study focuses on a tailor-made school model development for a multilingual school in Estonia, we adopted a qualitative approach for data collection and analysis. The study was carried out in an Estonian school that is situated in a historically bilingual region where currently about two thirds of the population are Estonian and one third Russian. Based on the conceptual framework described above, participants representing all of the levels of analysis (macro, meso, micro) were included in personal or focus group interviews. In total nine interviews were carried out. The researchers spent two full days in the school and during this time two teachers were open to class observations as well. Hence, data was also obtained from class observations. The interview data was complemented with information from different publicly available documents – both at the school level as well as the municipality level. Interview data was collected through semi-structed face-to-face interviews using an interview schedule. The interview schedule was developed based on the conceptual framework introduced above and focused on the organisation of studies and support provided to students whose home language differs from that of the school (non-Estonian) and/or who represents another culture besides Estonian. Another set of questions address the preparedness of school to educate non-Estonian students. The interviews were audiotaped and later transcribed in verbatim. The class observations were done using the SIOP (Sheltered Instruction Observation Protocol) modified slightly to match the focus of the research. The class observations were carried out as a paper-and-pencil protocols. The History and Estonian as a Second Language classes were observed. Data analysis was carried out using the web-based QCAmap freeware. A deductive approach was chosen for analysis as the conceptual framework provided the main categories to be analysed. Thus, a coding matrix was developed and implemented. The data analysis resulted in a case description with all of the major variables described at different levels. The data form the documents was also mapped based on the factors in the conceptual framework and was used to fill gaps in or to triangulate the interview data. The observations were used also as a triangulation source.

This article set out to unpack the black box of multilingual education model development by providing an Estonian example and demonstrating how through concrete analytical steps model options could be arrived at. In the previous sections, we have analyzed the different factors influencing the functioning of a multilingual Estonian school and made implications on what this analysis has on possible model development directions. We have presented how, by careful consideration of different factors at different levels, model choices start to emerge. Based on the historical background and the nature of the local community, a two-way language immersion class would be a logical development at the school. This idea has also been reflected in the development documents of the school. However, it seems that several factors, e.g. low demand for Russian proficiency among Estonian families and youth, lack of active support among political and school leadership, lack of expertise and visionary leadership as well as back-up from kindergartens, point to the direction where a different models need to be searched for. Based on the analysis of the factors presented above, we see a way forward towards three possible model options: a. Strengthening of the immersion programme into a classical maintenance /heritage programme (Baker, 2011) where actual full bilingualism and literacy would be achieved. b. Giving up the current immersion that in practice is a transitional subtractive (assimilationist) programme (Baker, 2011) and mixing all of the students regardless of their mother tongue. c. A third option would be to adjust the current immersion programme at least to an extent where contacts with Estonian students are significantly increased (joint integrated classes in a bigger number of subjects) and where the Russian competence of immersion students would be better supported (adding subjects in Russian).

Baker, C. (1985). Aspects of Bilingualism in Wales. Multilingual Matters. Baker, C. (2011). Foundations of Bilingual Education and Bilingualism (5th ed.). Multilingual Matters. https://www.kriso.ee/foundations-bilingual-education-bilingualism-db-97818476950866e.html Beatens Beardsmore, H. (2009). Bilingual Education: Factors and Variables. In O. García (Ed.), Bilingual education in the 21st century: A global perspective. Wiley-Blackwell. Brisk, M. E. (2006). Bilingual Education: From Compensatory to Quality Schooling (2nd edition). Lawrence Erlbaum Associates. Busch, B. (2011). Trends and innovative practices in multilingual education in Europe: An overview. International Review of Education / Internationale Zeitschrift Für Erziehungswissenschaft / Revue Internationale de l’Education, 57(5/6), 541–549. Cenoz, J. (2009). Chapter 2. Towards a Typology of Multilingual Education. In Towards Multilingual Education: Basque Educational Research from an International Perspective (pp. 22–56). Multilingual Matters. Creswell, J. W. (2007). Qualitative Inquiry and Research Design: Schoosing among Five Approaches (2nd ed.). Sage Publications. de Jong, E. (2006). Integrated Bilingual Education: An Alternative Approach. Bilingual Research Journal, 30(1), 23–44. https://doi.org/10.1080/15235882.2006.10162864 García, O. (Ed.). (2009). Bilingual education in the 21st century: A global perspective. Wiley-Blackwell. Kirss, L., Säälik, Ü., Pedaste, M., & Leijen, Ä. (2020). A Synthesis of Theoretical Frameworks on Multilingual Education for School Leaders. TRAMES, 24(1), 27–51. https://doi.org/https//doi.org/10.3176/tr.2020.1.02 May, S. (2017). Bilingual Education: What the Research Tells Us. In O. García, A. M. Y. Lin, & S. May (Eds.), Bilingual and Multilingual Education (3rd edition, pp. 81–100). Springer International Publishing. Mehisto, P. (2012). Excellence in Bilingual Education: A Guide for School Principals. Cambridge University Press. Mehisto, P. (2015). Introduction: Forces, Mechanisms and Counterweights. In P. Mehisto & F. Genesee (Eds.), Building Bilingual Education Systems: Forces, Mechanisms and Counterweights (pp. xv–xxvii). Cambridge University Press. Spolsky, B., Green, J. B., & Read, J. (1976). A Model for the Description, Analysis and Perhaps Evaluation of Bilingual Education. In A. Verdoodt & R. Kjolseth (Eds.), Language in Sociology (pp. 233–263). Institut de Linguistique de Louvain : dépositaire, Editions Peeters. Thomas, W. P., & Collier, V. (1997). School Effectiveness for Language Minority Students (No. 9). National Clearinghouse for Bilingual Education. http://www.thomasandcollier.com/assets/1997_thomas-collier97-1.pdf