Assessing C1 level what construct ? A comparative study between CLES C1 and ACLES C1.

Abstract

This paper introduces a research effort undertaken to further cooperation and recognition amongst two Higher Education certification systems, CLES (Certificat de compétences en langues de l'enseignement supérieur) in France and CertAcles (Model for language certification from the Association of Language Centres in Higher Education) in Spain. Our research undertakes the comparative analysis of C1 writing assessment in both certification systems, from the point of view of specifications, rating scales and linking to the CEFR. Furthermore, comparability claims will be made based on C1 CEFR descriptors illustrated in performances obtained from both certification systems. Our goal is to find a common path towards the certification of language competence in a university context, based on the descriptors of the CEFR -and thus comparable and internationally transferable-, but sustained on different context-based assessment models.
The challenges of assessing writing skills are inherent to the difficulty of examining the construct of writing, a process that needs to be learnt, not only in L2 but also in L1. Writing is deeply connected to thinking, a reflexive process that requires a high degree of language competence including stages of planning, drafting, revision, re-organisation, etc. (Heaton, 1975). The conflict of obtaining psychometrically reliable results when assessing such a complex construct was resolved by dissociating form and content, which resulted in indirect forms of testing writing, particularly in standardised testing. However, models of communicative competence (Hymes, 1972; Canale & Swain, 1980; Bachman, 1990; Celce-Murcia, Dornyei & Thunrrel, 1995; Bachman & Palmer, 1996) tilted the balance towards the validity of assessment and thus towards using direct testing methods while implementing mechanisms that ensured reliability of results (design of specifications and scales, rater training, double marking, etc.). The most critical aspect of validity in a writing test is that of construct validity, that is, whether the test designed is successful in measuring the element required for successful writing in English (Kroll, 1998). In the case of university certification systems, the element required is the ability to go beyond the knowledge-telling model characteristic of immature writers towards the knowledge-transforming model of a mature writer (Bereiter & Scardamalia, 2003).
The Network of University Language Testers in Europe (NULTE) was established under the umbrella of CERCLES (European Confederation of Language Centres in Higher Education) to encourage co-operation amongst university-specific certification systems with the goal of contributing to their shared goals of quality improvement. The co-operation agreement is based on a set of quality assessment principles, amongst which the importance of productive skills and context sensitive tasks were emphasised. CLES and CertAcles, as members of NULTE, have vowed to work together towards the mutual recognition of their language certificates.