Despite the increasing popularity of task-based language teaching (TBLT) among researchers and teachers, further classroom-based empirical evidence is needed to support its use over more traditional teaching approaches such as presentation-practice-production (PPP). The argument that TBLT is ‘legislation by hypothesis’ (Swan, 2005), is yet to be completely refuted (Ellis, 2018a). Four intact classes of Japanese university students taking a one-year oral English course were randomly assigned to either a TBLT or PPP teaching condition. Speaking proficiency and speaking self-efficacy were measured three times during the year to discover the differential impact of teaching approach on growth. Student interviews provided insights into their views on the different approaches. Results showed that students in TBLT and PPP classes experienced growth in both proficiency and self-efficacy, and there was no difference in final outcomes for the two conditions. However, patterns of growth for the groups were different.

As task-based language teaching (TBLT) has grown in popularity in many Asian EFL countries, it has sometimes been the target of criticism from locally situated researchers and teachers, who question its appropriateness for such contexts (Burrows, 2008), often in contrast to more traditional methods, such as the "presentation, practice, production" (PPP) approach (Sato 2009). Partially in response to this criticism, the Task Based Language Learning Special Interest Group (TBLSIG) of the Japanese Association for Language Teachers was formed in 2010 in order to contribute to research into the implementation of TBLT in Japanese and Asian EFL classrooms. As the majority of TBLSIG members are teachers, presently interested in using TBLT in classrooms, they provide an ideal source of insight into how the approach can be implemented effectively in Japan, and by extension, in other similar EFL countries in Asia. This study investigated these teachers' beliefs about TBLT in Japanese classrooms through an online survey. The aim of the study was to add to existing research on teacher beliefs by not only investigating teachers' understanding of the principles of TBLT, but also by enquiring into what special considerations practicing teachers have made when implementing TBLT in Japanese classrooms.

This paper outlines a qualitative study of 10 in-service teachers in Japan who use a task-based language teaching (TBLT) approach in their language classrooms. The study investigates the reasons why these teachers began using TBLT, their beliefs about the advantages and disadvantages of implementing TBLT, and whether or not they agree with common criticisms levelled at TBLT for use in foreign language classrooms in Japan and Asia in general. Critics argue that TBLT is not suitable for many language learning situations, explaining that students are used to more ‘traditional’ teacher-centered language classes and that they prefer more ‘passive’ approaches. Such criticisms also hold that TBLT is not able to adequately prepare students for the high-stakes tests that are used for secondary school and university entrance purposes. The results from interviews with these 10 teachers suggest that they reject many of these criticisms, and are in fact successfully using TBLT to develop both language ability and motivation in their students. At the same time, the teachers noted that there may be a need to use a ‘weaker’ form of TBLT with beginner students, or with those used to more passive styles of learning. They also highlighted the importance of heavily scaffolding tasks in such cases. Finally, a number of the teachers discussed the need for the development of more TBLT related materials, both for students and teachers.

This article outlines the development of a 16-item instrument for measuring language learner’s foreign language self-efficacy (SE) concerning their speaking and listening skills through repeated administrations to groups of Japanese tertiary students. Responses were analysed through the Rasch model, which allows researchers to investigate unidimensionality of each proposed component, as well as the functioning of individual items. Logit scores from Rasch were used for a regression analysis of listening SE with listening proficiency test data, and a correlation study with speaking SE and speaking proficiency scores of a subset of participants. The results suggest that the two proposed components in the instrument are unidimensional and are able to differentiate between different levels of SE among participants. Also, there are low to moderate correlations between learners’ aural/oral SE and their language proficiency.

No abstract