Acceso usuarios
Paula Elosua Oliden
Among the specificities associated with the accreditation of linguistic competence in minoritised languages, one of the most distinctive is the characterisation of the target population. In terms of diversity, test-takers come from the same geographical region, and most of them live in contact with the test language; in terms of quantity, there are fewer candidates than those assessed in L2 language around the world. These particularities have some important implications in any score-equating programme, and so on validity. In this paper, which focuses on Basque, we explore the conditions to define an equating design. We start by analysing the hypothesis of group equivalence among the editions of the HABE C1 test, and then compare some equating procedures. The research is carried out in two steps. The first (N = 14,981) analyses the sample compositions among test editions with relation to gender, age, place of birth, L1 and language of instruction; the second compares some equating procedures applied to the reading and listening test (N = 4765). The results support the hypothesis of group equivalence, and the equating procedures show total concordance around the cut scores. Some implications of the results are provided.
