Gergely Dávid: The Role of Language in Level Descriptors in the Hungarian-, English- and German-Language Editions of the Common European Framework of Reference This study deals with the validity of the Hungarian translation of the Common European Framework of Reference (CEFR) with a special focus on the Hungarian version of the illustrative scales. The scales have special relevance in a validity study as they operationalise the concepts of language proficiency for a wider audience and a variety of uses. Based on the English-language version, the Hungarian translation was the starting point for a number of CEFR linking projects in 2007, following the Council of Europe initiative and subsequent Hungarian legislation to establish a link between language proficiency tests and the CEFR. The linking scheme was pushed through, even though the Hungarian translation only involved a handful of professionals and was not validated at all. Further, the linking projects operated on the basis of different language versions of the CEFR. In this project, thus, quantitative data were collected in three languages (English, Hungarian and German) from 154 CEFR linking project judges by three language test producers based in Hungary. The research questions concerned the extent to which the validity of the Hungarian scales could be supported with appropriate validity evidence, whether the Hungarian scales could be linked to the CEFR scales in English and German and whether rival hypotheses (alternative explanations) could be rejected. These rival hypotheses included the effect of different languages (translation), the raters’ severity and the conditions of filling in the questionnaire. The data consisted of ratings of the CEFR level (A1, A2, etc.) of 829 descriptors from the illustrative scales. These data were drawn from questionnaires originally meant as a survey of the judges’ competence in the CEFR. On the basis of the ratings, the descriptors, including the Hungarian translations, were scaled and classified, and the results were compared to the original CEFR calibrations-based level classifications. The results confirmed the presence of a language-of-the-scales facet. Overall, they show a good degree of fit between the way Hungarian judges rated the descriptors in the three languages and the official CEFR level categories. In comparison with CEFR classifications, Hungarian judges tended to be lenient concerning requirements at levels A1 to B2 and strict concerning levels C1 to C2. MAGYAR PEDAGÓGIA 112. Number 1. 19-39. (2012) Levelezési cím / Address for correspondence: ELTE, Angol-Amerikai Intézet, Angol Nyelvpedagógia Tanszék, H–1088 Budapest, Rákóczi út 5. |
|