Gauging ESL learners' CEFR ratings on oral proficiency in rater training

Abstract

Rater training is fundamental in reducing rater variability in self- and peer assessments practice within the paradigm of assessment as learning (AaL). Since Malaysian education system is examination-oriented in which assessment of learning (summative) and assessment for learning (formative) dominate, ESL learners are rarely asked to rate themselves or their peers as the system is still sceptical in entrusting learners with the role of assessors. Learners are normally perceived as unable to (1) assess accurately, (2) assess consistently and (3) discriminate oral proficiency components in their performance. Therefore, this study attempts to gauge learners' rating skills on these three assumptions through the use of Common European Framework of Reference (CEFR) oral assessment criteria. Quantitative analysis was conducted using the Rasch model while a short semi-structured interview was used to support the quantitative results obtained. Findings from this study suggest that ESL learners were generally able to rate accurately and consistently after rater training. The rater training had also somewhat sensitized these learners to CEFR oral proficiency components but they were still grappling in confidently rating range and accuracy. Based on these findings, it indicates that ESL learners' are ready for AaL paradigm shift and this will possibly launch a platform for self- and peer assessments practice in ESL classrooms. Such assessments will promote active learning and effective learner-centred classroom.