In response to the multidimensional challenges of cultural and linguistic diversity that are posed to education systems in Europe and worldwide, the ProfaLe Teacher Education innovation project at the University of Hamburg, Germany, aims to equip all future subject teachers with relevant skills for dealing with multilingual classroom realities. While the necessity to provide language-sensitive subject teaching to all learners is prevalent in all primary and secondary subjects, there is particular relevance for Sciences and Mathematics. Student Assessment Studies such as PISA point to different results between students with and without a migrant background. Among other things, the PISA report shows that within the German education system, the worst performers in the field of sciences are fourth grade pupils with a migration background (OECD 2011). As research in this field shows this is not only due to the consistent finding of the strong dependency of the educational success from social status but also insufficient support for learners of diverse linguistic backgrounds. Moreover, it has been established that academic language learning and content learning are intertwined, and that developing proficient academic language has a significant impact on general educational achievement (Gibbons 2009).

Against this backdrop, the project work presented in this talk aims at improving teacher-training structures that allow for space in the curriculum in which pre-service teachers can gain relevant skills for dealing with learners who experience linguistic difficulties in subject learning; and this especially for maths-and science teachers as assisting the development of academic language is the responsibility of all subject teachers. In the pilot phase of the project, all future teachers of mathematics, physics, biology and primary school sciences attend a newly created seminar structure in the M.Ed. phase of their training. This structure contains a mandatory blended-learning seminar on the foundations of multilingualism and linguistics, an internship that is framed by a didactics seminar in mathematics or science, in which issues of language across the curriculum and continuous language education (also known as “durchgängige Sprachbildung” established by Gogolin et. al. 2011) are discussed, and further voluntary seminars in intercultural and multilingual education. One of the key innovation aspects in the development and establishment of new structures is the close interdisciplinary cooperation between the departments of German linguistics, intercultural education and four different subject-specific didactics departments. This provided us with the opportunity to create coherent syllabuses, where lecturers were able to draw on contents that had been addressed in other classes.

In the process of implementing this model, our research focusses on its effectiveness and on the skill development of our students. Our dominant research questions evolved around a) uncovering what constitutes good competence development in the relevant areas, b) observing improvements in the knowledge and competence of our students, and c) finding significant predictive explanations on competence improvements.

To gain a better understanding on the competency and skill development in the area of professional competence for inclusive academic language education, all students are tested in a pre-post design before and after the new training structure. The test instrument (DaZKom Test) measures the participants’ skills in dealing with GSL (German as a second language) and multilingual learners (see Köker et al. 2015, Hammer et al. 2015).

In the summary of the paper presented here, we argue that, on the one hand, it should be self-evident that teacher-training programmes in many European contexts would benefit from improvement with regard to linguistic diversity. On the other hand, our results show that careful consideration needs to be given to the design of training modules in order to create equal opportunities for all learners and in respect to the most prevalent sub-themes.

In order to research the skill and competence development of pre-service teachers in the area of inclusive academic language education, we employ the aforementioned GSL-Competency Test, which is a thoroughly validated instrument that provides weighted likelihood estimate (WLE) measures as outcomes based on an Item Response Theory validation. It consists of 47 items of which 12 were open, 11 were semi-open and 24 were closed items. Tested content areas were subject-specific registers, multilingualism and didactics. The testing procedure takes 60 minutes and is carried out in the beginning of a semester and then again six months later after the semester and training internship. Within the test, further data is collected. In one part, participants are asked how they perceive the opportunities to engage with relevant topics across their teacher-training programme. In 16 sub-categories, they can provide information on how often they have had relevant learning opportunities in the area of linguistic diversity and GSL issues. Moreover, participants are asked for their background, personal information, educational career, language knowledge as well as their orientations on multilingualism (beliefs). Based on this information, a number of bivariate and multivariate analyses have been conducted. Our two cohorts (n=380) consist pre-service mathematics and sciences teachers. In a first step, we employed bivariate statistics to explore the datasets and to find relevant significant correlations between various variables and the participants’ WLEs. Among other things, we found significant relationships between the final grade of their upper secondary education, the sum of learned languages and the study of an additional language subject and the test outcomes. In a second step, different multivariate regression models were designed and tested. For the purposes of this presentation, we address an eight-step regression model that best predicts the participants’ performance in the post-test. This model includes information on the pre-test score, information on the study subjects of the participants, gender, secondary education attainment, the perceived learning opportunities in the teacher-training programme, home language usage, students’ motivation, participants’ parents’ education as well as relevant experience. With an adjusted R² of .430, this model provides solid predictions of the test score across all our cohorts.

Within our pre-post design, we observe significant improvements in the test scores across our cohorts. The WLE scores are translated into a model that differentiates between three stages of competence. This three-dimensional model is based on the first three of the five stages in the Dreyfus Model of Skill Acquisition (Dreyfus & Dreyfus 1986). It can be seen that a majority of our students (87%) are on the lowest (novice) stage in the pre-test. While scores in the post-test are significantly better (with still 62% remaining in this category), there is still considerable room for improvement. Among the most interesting aspects, we observe a particularly noteworthy development when analysing the perceived opportunities to engage with relevant issues in the students’ study programmes. In 16 different sub-fields of learning about language issues and multilingualism, only one field correlates significantly with the pre-test score. In the post –test, however, 11 out of the 16 fields increase significantly and 12 out 16 correlate significantly with the post-test score. This shows that in the post-test, there is a far more systematic relationship between recognising and benefitting from the aforementioned learning opportunities. First regression models allow us to name significant predictors for high competence improvement in the post-test. In the aforementioned eight-step regression model, we find statistic evidence can explain an adjusted R² of 44% in the post-test results. Within these models, we find that besides the pre-test results (as an independent variable), the relevant learning experiences form the strongest predictive value for the post-test performance. In the conclusion of our talk, we will provide further insights into these regression models, and discuss them with regard to future questions on how teacher education can be improved concerning competence development in the area of multilingualism in the classroom.

Dreyfus, H.L., Dreyfus, S.E., (1986). Mind over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. New York: The Free Press. Gibbons, P., (2009). English Learners, Academic Literacy, and Thinking. Portsmouth, NH: Heinemann. Gogolin, I., Dirim, I., Klinger, T., Lange, I., Lengyel, D. et al. (Hrsg.) (2011). Förderung von Kindern und Jugendlichen mit Migrationshintergrund FörMig. Bilanz und Perspektiven eines Modellprogramms. Münster u.a.: Waxmann. Hammer, S., Carlson, S., Ehmke, T., Koch-Priewe, B., Köker, A., Ohm, U., Rosenbrock, S., Schulze, N., (2015). Kompetenz von Lehramtsstudierenden in Deutsch als Zweitsprache: Validierung des GSL-Testinstruments. Zeitschrift für Pädagogik 61, pp. 32-54. Köker, A., Rosenbrock-Agyei, S., Ohm, U., Carlson, S. A., Ehmke, T., Hammer, S., Koch-Priewe, B. & Schulze, N., (2015): DaZKom - Ein Modell von Lehrerkompetenz im Bereich Deutsch als Zweitsprache. In: Koch-Priewe, B., Köker, A., Seifried, J. & Wuttke, E. (Eds). Kompetenzerwerb an Hochschulen: Modellierung und Messung: Zur Professionalisierung angehender Lehrerinnen und Lehrer sowie frühpädagogischer Fachkräfte. Bad Heilbrunn: Klinkhardt, pp. 177-205. OECD (Ed), (2011). Divided we stand. Why inequality keeps rising. Paris: OECD.