English essay writing is a key competence in a modern world where English is the lingua franca of science, business and higher education. MEWS will be the first large-scale study of English essay writing in Germany or Switzerland at upper secondary level (10th grade) and has three research questions: (1) How proficient are learners in essay writing in English two years before their baccalaureate? (2) What is the effect of individual factors (e.g. motivation, intelligence), family background and extracurricular activities on English essay writing competences? (3) What is the effect of school and classroom factors (e.g. classroom instruction, school types, ‘baccalaureate rate’) on English essay writing competences? Learners’ ability to write English essays will be measured in two populations, the first one from Switzerland (AG, BS, BL, SO, ZH, N = 1500), the second one from Germany (Schleswig-Holstein, N = 1500) at two time points, gathering a total of four responses from each learner (12.000 essays in total). We expect individual factors (level 1), classroom factors (level 2) and system factors (level 3) to be substantial predictors of writing skills. Based on our literature review, we expect previous writing skills as well as receptive skills in English to be the strongest level 1 predictors of subsequent writing skills, followed by cognitive abilities and motivational characteristics, i.e., academic self-concept and interest. Concerning level 2 effects, we expect learners’ familiarity with the genre of the English argument essay to be the strongest predictor, either because they studied such texts as models and/or actively practiced writing them. The project has a strong pedagogic focus and will investigate to what degree classroom factors such as peerfeedback, criteria-based assessment or teaching relevant lexico-grammatical features predict writing skills. Concerning systemic factors (level 3), we expect to find the relationship ‘lower baccalaureate rate = better student performances’ to be mirrored in our data. The analysis of these groups of variables will be conducted in the framework of multilevel analysis (MLA). Beyond these substantial research questions, the project strives to apply and evaluate modern artificial intelligence (AI)-based techniques of evaluating texts. We will use rigorously trained human raters as well as the ETS-owned and proprietary automated essay scoring engine e-rater® to measure key aspects of learners’ writing skills: The ability to compose grammatical, well-punctuated sentences, to choose appropriate genrespecific vocabulary, and to structure their compositions coherently by use of connectors and linking words. E-rater is the scoring engine also employed in the Test of English as Foreign Language (TOEFL) and will help to overcome high costs associated with human scoring and avoid biased artefacts due to harshness and leniency effects. The project is done in a binational setting in order to exploit the complementary expertise of the researchers and to compare the educational systems of the two countries involved. It will lay the groundwork for an exciting new paradigm for English writing research and provide empirical evidence on how teachers can create a ‘good writing classroom’ at upper secondary level.
