Germany faces continuous in-migration since World War II. Nevertheless, young migrants still do not have equal chances of participation, especially if they live in urban areas. International large scale assessment studies still show significant differences between educational attainment of migrant and autochthonous students. Migration is not the only factor that triggers educational disadvantage. Indeed, a variety of factors deriving from migration are usually confounded with other factors, namely the socio-economic background of families and their cultural capital. With respect to access to vocational education or to work, mechanisms of active discrimination of migrants were identified in respective studies.The research project "Mehrsprachigkeitsentwicklung im Zeitverlauf (MEZ)" (Multilingual Development -a Longitudinal Perspective) takes these observations as its starting point. By answers to its research questions it will not only contribute to the advancement of scientific knowledge, but also uncover basic information for the development of alternative educational practices which can contribute to the mitigation of educational disadvantages of young migrants. We assume that strengthening their learning potential and resources can add to this aim. And we consider that their language competences, namely the bi-or multilingual competences deriving from the heritage languages they use (not only) in their families, are a relevant part of their potential and resources. In German we use the term "lebensweltliche Mehrsprachigkeit/ multilingualism" in order to call attention to the specific linguistic constellation which the young migrants experience: they acquire and make use of more than one language in their everyday life.In the MEZ project we ask how language competences develop over secondary schooling. We apply a two-cohort sequential design. Data collection started in January 2016 with two parallel cohorts (grade 7 and grade 9) and will be carried out in four waves over three years (until grade 9 and 11, respectively). The sample is composed of autochthonous students who grow up and live in German only on the one hand, and students with a "lebensweltlich multilingual" background on the other. The latter speak German-Turkish or German-Russian at home. We test language proficiency in German (all participants), English (1 st foreign language; all participants), Turkish and Russian as heritage languages and French or Russian as 2 nd foreign languages taught in German schools. The tests measure different dimensions of proficiency: global language competencies in the foreign languages English, French and Russian; receptive language skills in the heritage languages German, Russian and Turkish by means of a standardized reading test; productive writing skills in the foreign languages English, French, and Russian as well as in the heritage languages German, Turkish and Russian. Furthermore, we collect data on individual (e.g. motivation, cognitive competence), contextual (e.g. SES, cultural capital, educational biograph...