Abstract
Recently, language testing research has thrown more and more light on the nature and function of performance tests, compared with traditional paper-and-pencil tests. This new focus is closely related to a world-wide trend in education, namely, the introduction of criterion-referenced assessment and the harmonizing of this with norm-referenced assessment. In fact, in many countries or regional organizations (such as the US, the UK, the Council of Europe etc.) scales of language proficiency have been created to provide a more reliable foundation for subjective assessment in oral interview tests. Of course, first and foremost, these scales need to be valid (in accordance with assessment purposes), reliable (regardless of the assessors, the 'assessees' and the testing occasions), and practical (for actual use). On the basis of this background, the present study attempts to develop an English oral proficiency scale by means of Many-facet Rasch Measurement (MFRM) analyses (Linacre, 2004), in order to make it possible for Japanese teachers of English to make reasonable judgements about their students' performance, even in classroom settings.
