Despite the scholarly acceptance of the role of academic registers in the language of instruction as a solid indicator for the educational success of adolescents with an immigrant background (Bourne 2013; Cummins 2000), the role heritage languages play in its development is yet contested (Gogolin & Neumann 2009). On the one hand, large-scale monitoring studies highlight the relationship between heritage language use at home and lower school performance (Klieme, Artelt et al. 2010). On the other hand, research found that transfer of higher order literacy skills is more adequately explained by means of Cummins’ interdependence hypothesis (Dressler and Kamil 2006). Building on these findings, the paper pinpoints the role of written skills in the heritage languages in explaining variation in academic language performance in the majority language in adolescent students. Further, it discusses the role of other common predictors for academic language performance of multilingual speakers. Based on bilingual written samples from 94 Turkish- (n=41) and Russian-German (n=53) 15-year-olds and on data from a background questionnaire, OLS regression analyses were conducted to find adequate predictors for variation in academic language performance. Language proficiency in German, Turkish and Russian was measured with the validated instrument “Fast Catch Bumerang” (Lengyel, Reich et al. 2009), which elicits written texts based on a series of nine pictures. In addition, working memory skills were measured with a sub-test of the “Hamburg-Wechsler-Intelligenztests für Kinder“ (HAWIK-IV) (Petermann & Petermann 2010). Other background variables were collected based on the data of the questionnaire on language use, socio-economic background and cultural capital. Missing data was replaced by plausible values using multiple imputation (Rubin 1987). The results confirm the relevance of written skills in the family languages in predicting academic language performance. In particular, adjectival use significantly contributes to explain the most variation in academic language performance in German. In contrast, SES, cultural capital and working memory are less strong predictors. In addition, the school type (academic or lower tracks) seems to be a better proxy for social-economic background in adolescents than common measures, such as the ISEI. The final model explains 31% of variation in the performance. The study provides solid arguments towards explicitly fostering writing skills in the heritage languages in order to achieve high levels of biliteracy. Further research with more representative samples and including other background variables is needed to assess the effects of heritage language proficiency on school success.

Bourne, J (2013): ‘I know he can do better than that’: strategies for teaching and learning in successful multi-ethnic schools. In: Gogolin, I., Lange, I., Michel, U. & Reich, H.-H. (eds.): Herausforderung Bildungssprache – und wie man sie meistert. Münster: Waxmann, 42-54. Cummins, J. (2000): Language, Power and Pedgogy: Bilingual Children in the Crossfire. Clevedon: Multilingual Matters. Dressler, C., & Kamil, M.L. (2006). First-and second-language literacy. In D. August, & T. Shanahan, (Eds.): Developing literacy in second language learners: Report of the national literacy panel on language-minority children and youth. Mahwah, NJ: Erlbaum, 197-238. Gogolin, I. & Neumann, U. (2009): Streitfall Zweisprachigkeit - The Bilingualism Controversy. Münster: Waxmann. Klieme, E. Artelt, C. Hartig, J., Jude, N., Köller, O, Prenzel, M., Schneider, W., Stanat, P. (eds.) (2010): PISA 2009. Bilanz nach einem Jahrzehnt. Münster: Waxmann. Lengyel, D., Reich, H.H. et al. (eds). (2009): Von der Sprachdiagnose zur Sprachförderung. FörMig Edition. Münster: Waxmann. Petermann, F. & Petermann, U. (2010): HAWIK-IV. Bern: Huber. Rubin, D. (1987): Multiple Imputation for Nonresponse in Surveys. Hoboken NJ: John Wiley & Sons.