渡部 良典

The present paper is an attempt to place a research study on language testing into a larger framework, in the expectation that it helps enhance 'qualities of life' in language classrooms (Allwright, 2005). In so doing, a variety of issues relating to language assessment in general are addressed, including teacher-student perception gap in the purpose of classroom assessment, a motivational function of language assessment, the role of language assessment in deepening the understanding of students proficiency profile, and the method of using assessment for research. These issues are presented and discussed with concrete examples from the empirical research studies that have been conducted to date. The paper concludes by suggesting that the present attempt will also help prevent researchers from falling into an extreme case of specialization.

The present paper portrays three theories of motivation in the expectation that it will help to understand the washback effect of language tests on learning English as a Foreign Language (EFL). The three theories that are identified involve attribution theories of motivation, flow, and functional theories of motivation. The characteristics of these theories are described in a way in which they may help understand the meaning of the recent attempt by the Japanese Ministry of Education to innovate in EFL practices by means of a Criterion-Referenced Assessment System. The paper concludes with several predictions regarding the effectiveness of the project that should be examined in future research.

In the present paper, the author reports that, contrary to the common assumption, interference errors are not due to the lack of knowledge about the structure of the target language being learned alone, but also to the learner's personality factors. It is hypothesized that if Ss are on the same level of proficiency in the target language, more introversive learners or those with a tendency to be self-conscious tend to make more interference errors than more extroversive ( or outgoing ) learners. The result did not show a decisive answer, but the tendency was found, showing that more introversive learners tended to make more errors caused by the interference of Ss' first language.